phase: 163_pdf_strip_additive_selfheal

All gates verified. Final report:

**Phase 163 — final verification pass (all 5 tasks already complete; verification-only run)**

- Verified D1–D4 in code: `page_strip_mode` column + migration 0027, `DocContent` absent-never-null rule keyed off `page_count`; `_persist_pdf_pages` render → atomic tmp+`os.replace` writes → stale-tail-last + `strip_class` stamp; `_regate_pdf_strip` on both unchanged paths (mirror gate, heal, repair, fail-soft, one INFO line only when acting); viewer mode branch + Text/Pages toggle (Text default, `aria-pressed`, no innerHTML, lazy imgs).
- `uv run pytest --cov=app --cov-report=term-missing` → **3796 tests, exit 0, coverage 99%** (>90%)
- `uv run pytest tests/e2e/test_pdf_strip_additive_selfheal.py -v --no-cov` → **5 passed** in isolation
- `test_pdf_pages_in_sources + test_pictorial_pdf + test_garbage_pdf` → 10 passed; `test_document_viewer(_context) + test_image_documents` → 20 passed
- `uv run ruff check .` clean; `uv run pyright` → 0 errors; alembic **0027 downgrade→upgrade CLI round-trip clean** (+ integration pins)
- Completion criteria: additive view w/ toggle ✅ · primary DOM byte-identity ✅ · write-first order/mid-write-failure/atomic pins ✅ · heal + repair E2E ✅ · zero vision/LLM on fast path (mock-captured) ✅ · non-qualifying payload byte-identity ✅ · docs pass ✅
- **Documentation**: README.md + AGENTS.md reviewed in full — the phase's updates are accurate; all config-table defaults cross-checked against `app/config.py` (no drift), every page route, script, and E2E command referenced exists, every `.env.example` var documented — **verified consistent, no changes needed**.
- Notable: live `Letumdei` repair is owner-runtime post-deploy per D5 (next production Sync self-heals; no live changes made); the modified `.agents/screenshots/doc_modal_scrollbar_1280.png` is an E2E-regenerated artifact left for the pipeline commit.
- **Next pending phase:** `164_chat_error_honesty`

.agents/phases/complete/163_pdf_strip_additive_selfheal/00_phase.md

@@ -0,0 +1,73 @@
+# Phase 163 — PDF page strip: additive text+pages view, write-first atomic copies, deterministic fast-path re-gate (self-heal)
+
+**Source:** owner live report, 2026-10-07 — brain.reeseapps.com, the document modal for `Letumdei_ Heirs to the Imperium/Letumdei_ Heirs to the Imperium.pdf` (doc id `b72f5515-dc40-4167-98ff-179a6503e6e5`). Every page rendered "Page N unavailable." and the document was unreadable in the viewer. **Pre-fix evidence (measured live, 2026-10-07):** the payload carries 36 `page_urls` (so `page_count = 36`); `GET /api/documents/<id>/pages/1…36` **all 404**; the content is genuine extracted text (114 389 chars, 36 `---PAGE-N---` markers, exactly ONE `[Page 36 illustration: …]` block — a map page under `pdf_pictorial_min_chars`); the same server serves the 10/3-indexed IKEA manual's pages 200 (the pages dir was NOT wiped — only this doc's copies are gone). Two stacked defects:
+
+1. **Eligibility consequence (the "shouldn't be a pictorial pdf" complaint).** The phase-140 gate is per-page (`len(text.strip()) < pdf_pictorial_min_chars`, default 100); ONE sparse page in a text-heavy 36-page book set `pictorial_described = True` → the doc joined the strip class → the phase-148 viewer renders the strip **INSTEAD OF** the text content and the Summary panel. The owner's call (2026-10-07): **additive strip** — a usable-layer (pictorial-class) doc shows its TEXT (plus the Summary panel, parity with any text doc) with a **Text/Pages toggle** for the strip; the text is never hidden. The fallback class (unusable text layer — the phase-133 screenshots ARE the content) keeps today's strip-only view, DOM byte-identical.
+2. **Lost copies, stale `page_count`, no self-heal.** `_persist_pdf_pages` **deletes the old copies from disk first**, and the filesystem work is NOT transactional with the DB session: any later rollback of the file's transaction (the per-file `EmbeddingError` handler, a killed sync) — or an external wipe of `pdf_page_dir` — leaves the row claiming `page_count = 36` over ZERO files. The phase-147 fast path ("the copies persist") never re-renders, so the strip 404s forever. The owner's call: **write-before-delete + fast-path re-gate** — the changed path writes the new copies (atomically, per file) BEFORE any deletion; the unchanged paths deterministically re-gate the strip (no LLM/vision call) — re-stamp the class, re-render missing copies, or clear a strip that no longer qualifies.
+
+**Story:** n/a (owner-directed bug fix; extends `148_pdf_pages_in_sources`, `147_sync_changed_only`, `140_pictorial_pdf_pages`, `133_pdf_screenshot_fallback`).
+
+## Objective
+A pictorial-class PDF is a **readable text document with the pages one toggle away** (Text default; Pages shows the phase-148 strip); a fallback-class PDF keeps the strip-only view byte-identically; and a strip never dies silently — the next sync **re-gates and re-renders** missing copies (or clears a strip that no longer qualifies) with zero LLM calls. The live `Letumdei` doc is repaired by the next production Sync (no manual script): additive mode stamped, 36 copies re-rendered, the book's text back in the viewer.
+
+## Dependencies
+- `148_pdf_pages_in_sources` (complete) — `page_count`, `pdf_page_dir`, `<doc-id>-p{N}.png`, the `/pages/{n}` route, the viewer's `docPagesStrip`, and the persist/regate seams this phase changes.
+- `147_sync_changed_only` (complete) — the `file_hash` fast path + shared `_unchanged_doc` (the re-gate hook site) whose "the copies persist" assumption this phase replaces.
+- `140_pictorial_pdf_pages` (complete) — the deterministic `pictorial_pages` gate + the kill switch `settings.pdf_pictorial` (the re-gate mirrors this gate exactly).
+- `133_pdf_screenshot_fallback` (complete) — `is_usable_pdf_text` (the fallback-class signal the re-gate mirrors).
+
+## Design (shared by all tasks — the executor reads this, not the chat)
+
+**D1 — the strip-mode wire signal (task 01, LOCKED A2/A3).** New nullable `documents.page_strip_mode` VARCHAR(16) (migration **0027**, tested downgrade, A13): `"primary"` = fallback class (strip replaces the text — today's behavior), `"additive"` = pictorial class (text + toggle). Set exactly when `page_count` is set (cleared with it). `DocContent` gains `page_strip_mode: str | None = None` with the `image_url` omission rule: **ABSENT (never null) when `page_count` is NULL** — every non-qualifying doc's payload stays byte-identical to pre-phase. For a qualifying doc the key is present; a **NULL-mode pre-phase row serializes ABSENT** and the viewer treats absent as `"primary"` (existing rows keep today's DOM until the re-gate stamps them). The `/pages/{n}` route is untouched.
+
+**D2 — write-first atomic persist (task 02, LOCKED A4).** `_persist_pdf_pages` reorders: **render → write every page (atomic per file: write to `<name>.png.tmp`, `os.replace`) → only then delete stale copies** outside the written set (the shrunk-PDF tail; the current blanket `_delete_page_copies` BEFORE the render is removed — a rollback or crash can no longer leave the row over zero files). The partial-write failure path unlinks BOTH the new final files and the `.tmp` files before re-raising (no orphans; the existing fail-soft contract — caller sets `page_count` NULL + warning — unchanged). The function also stamps `doc.page_strip_mode`: `"primary"` when `fallback_fired`, `"additive"` when `pictorial_described` (the caller passes the CLASS, not just the boolean — `eligible=fallback_fired or pictorial_described` is replaced by `strip_class="fallback"|"pictorial"|None`); the ineligible arm clears mode with `page_count`. **The eligibility RULE itself is unchanged** (the phase-148 lock `fallback_fired or pictorial_described` stands — only its VIEW consequence changes, via D1/D4).
+
+**D3 — the deterministic fast-path re-gate (task 03, LOCKED A5).** A new `_regate_pdf_strip(session, doc, *, full_path, llm, summary)` runs on BOTH unchanged paths — inside `_unchanged_doc` (an optional `full_path` kwarg, defaulting to `Path(doc.full_path)`), so it fires on the `file_hash` fast path AND the content-digest-hit path — guarded by `doc.page_count is not None` (only PDF rows ever carry it; every other file's fast path keeps its zero-extra-read contract). It mirrors the changed-path gate **deterministically from the current file bytes + current settings — ZERO vision/LLM calls** (`render_pdf_pages` is CPU-only, A21-safe):
+- `not is_usable_pdf_text(extract_pdf_text(full_path))` → class `fallback` (mode `primary`) — same doc as today;
+- elif `settings.pdf_pictorial` and `pictorial_pages(extract_pdf_page_texts(full_path), min_chars=settings.pdf_pictorial_min_chars)` is non-empty → class `pictorial` (mode `additive`) — **candidates are exactly the existing gate's** (text-carrying pages under `min_chars`; no-text pages are NOT candidates, mirroring `pictorial_pages` — the re-gate adds no new semantics);
+- else → **repair arm**: the strip no longer qualifies (e.g. the owner lowered `pdf_pictorial_min_chars`) → `page_count = None`, `page_strip_mode = None`, `_delete_page_copies(doc, page_dir)`.
+
+For a re-gated-eligible doc, **verify the copies**: any of `page_dir/<doc.id>-p1..p{page_count}.png` missing → re-render ALL pages (`render_pdf_pages`), atomic write-first (D2's writer), and set `page_count` to the fresh render's length (a drift names the doc in a warning; bytes unchanged makes drift pathological). A fresh render producing NO pages (unopenable despite the just-computed hash — pathological) keeps the row as-is with a warning (fail-soft, never clears on a transient). One greppable INFO line whenever re-gate ACTS (never per-unchanged-doc): `import: PDF strip re-gate source=… path=… — healed N pages | cleared stale strip | mode <old>-><new>`. The re-gate mutates the row INSIDE the unchanged path's transaction (the commit already owned by `_unchanged_doc`'s caller flow); a copy-verify/render exception is caught, warned, and never fails the sync.
+
+**D4 — the additive viewer (task 04, LOCKED A1/A6).** `renderDocument` (the ONE shared core — modal AND `/document.html`): the strip branch now keys on the mode —
+- `page_strip_mode` absent or `"primary"` (with `page_urls`): the EXACT current DOM (`.doc-pages` strip, no toggle, no Summary panel) — byte-identical, pinned;
+- `"additive"` (with `page_urls`): the text doc renders **exactly as a non-qualifying text doc** (Summary panel + content, the pre-148 branch reused), PLUS a `Text`/`Pages` toggle above the content: `<div class="doc-view-toggle" role="group" aria-label="Document view">` with two `<button type="button" aria-pressed>`; **default `Text`** (LOCKED A6 — no sparse-ratio default flip); switching to `Pages` appends/reveals the `docPagesStrip` (imgs `loading="lazy"` — hidden-until-shown never fetches) and hides the text block; switching back reverses. State is per-render (no persistence). createElement + properties/textContent only (the never-build-HTML rule); styles.css gains `.doc-view-toggle` per the house button idiom (focus-visible ring, ≥4.5:1, pressed state is text + style, never color alone — B5).
+- no `page_urls`: byte-identical today's text doc (untouched).
+
+**D5 — the live repair (task 05, records; no code).** With D3 shipped, the next production Sync takes the `Letumdei` doc's unchanged fast path → re-gate: usable text layer + ≥1 sparse candidate (page 36) → mode `additive` stamped → all 36 copies missing → re-rendered → the viewer shows the book's text with the Pages toggle. The IKEA-class doc (bytes unchanged, copies present, mode NULL pre-phase) gets mode `additive` stamped with zero render work. The E2E fixture reproduces the exact live shape (one illustration block on the last page, copies deleted out from under a stamped `page_count`).
+
+**Invariants (all tasks).** The changed-path eligibility rule (`fallback_fired or pictorial_described`), every vision pass, the `/pages/{n}` route, the fallback-class strip DOM, the non-qualifying payloads, all SSE frame shapes, the mock-LLM markers, and the prune (`_delete_page_copies` at prune time) are byte-identical. `PLAN.md` untouched (the §7 viewer line describing the strip is folded at the next redo; AGENTS.md's viewer bullets get the additive note in the final pass).
+
+## Tasks
+1. `01_schema_wire_strip_mode.md` — the `page_strip_mode` column + migration 0027 + the `DocContent` field + the payload builder + unit pins (D1).
+2. `02_persist_write_first.md` — `_persist_pdf_pages` write-before-delete + atomic replace + the `strip_class` stamp (D2).
+3. `03_fast_path_regate.md` — `_regate_pdf_strip` on the unchanged paths: mirror gate, heal, repair, log line (D3).
+4. `04_viewer_additive_toggle.md` — `renderDocument` mode branch + Text/Pages toggle + styles + frontend unit pins (D4).
+5. `05_e2e_selfheal_pins.md` — the dedicated Playwright suite (additive view, primary byte-identity, heal, repair arm), the final gates, the docs pass (D5 + gates).
+
+## Testing & Quality
+- Unit: `tests/unit/test_pdf_page_copies.py` (D2: write-first order, atomic replace, partial-write cleanup, the class stamp; D1: column defaults), `tests/unit/test_importer_unchanged_fast_path.py` (D3: every re-gate arm incl. kill-switch, heal, repair, fail-softs, the zero-extra-read guard for non-strip docs), `tests/unit/test_docs_api.py` or the phase-148 payload-pin file (D1: the key present/absent rule, the byte-identity of non-qualifying payloads), a frontend source-contract file per the house style (D4: the mode branch, the toggle semantics, the never-build-HTML rule, the primary-DOM byte-identity).
+- E2E: `tests/e2e/test_pdf_strip_additive_selfheal.py`, **in isolation** (`uv run pytest tests/e2e/test_pdf_strip_additive_selfheal.py -v --no-cov`); `test_pdf_pages_in_sources.py` / `test_pictorial_pdf.py` / `test_garbage_pdf.py` stay green (the re-cut pins included).
+- Coverage: **>90%** on `app/`; `uv run ruff check . && uv run pyright` clean.
+
+## Completion Criteria
+- [ ] A pictorial-class PDF opens in the viewer showing its text + Summary panel with a Text/Pages toggle (Text pressed by default); Pages shows the strip and every page image loads (200); a fallback-class PDF's strip-only DOM is byte-identical to pre-phase (no toggle).
+- [ ] The changed path writes copies before deleting any (measured by the unit order pin); a mid-write failure leaves the prior copies intact and no orphans.
+- [ ] Deleting a strip's copies on disk + a re-sync (unchanged bytes) restores them and stamps the mode (the E2E heal test); a stale `page_count` on a now-ineligible doc is cleared on re-sync with the copies removed (the E2E repair test); zero vision/LLM calls on the fast path (asserted via the mock).
+- [ ] Non-qualifying-doc payloads and the primary-class DOM are byte-identical; the full `uv run pytest` suite is green; coverage >90%; `uv run ruff check . && uv run pyright` clean.
+- [ ] `README.md` + `AGENTS.md` reviewed in full in the final pass (the viewer description gains the additive-mode sentence; `PLAN.md` untouched per the invariants).
+- [ ] One `--no-gpg-sign` commit; the phase dir moves to `complete/` by the pipeline gate. Post-deploy (owner runtime, recorded here): the next production Sync self-heals `Letumdei` per D5 — no manual script.
+
+## Locked decisions
+- **A1 — additive strip**: usable-layer (pictorial-class) PDFs show text + Summary + a Text/Pages toggle; the text is NEVER hidden for them; fallback-class keeps the strip-only view byte-identical (owner-confirmed 2026-10-07 — supersedes phase-148's "strip instead of the text content" for the pictorial class only).
+- **A2 — the wire signal**: `page_strip_mode: "primary"|"additive"` ABSENT (never null) when `page_count` is NULL; absent ⇒ viewer treats `"primary"`; non-qualifying payloads byte-identical (owner-accepted default, 2026-10-07).
+- **A3 — the column**: `documents.page_strip_mode` VARCHAR(16) nullable, migration 0027 with a tested downgrade (A13); stamped/cleared exactly with `page_count`.
+- **A4 — write-first atomic copies**: render → atomic per-file write (tmp + `os.replace`) → stale-tail delete; a rollback can never leave the row over zero files; the fail-soft contract unchanged (owner-confirmed 2026-10-07).
+- **A5 — the deterministic re-gate**: unchanged paths mirror the changed-path gate exactly (`is_usable_pdf_text` / `pictorial_pages` + the `pdf_pictorial` kill switch, current settings), re-render missing copies (CPU-only, no vision), or clear a stale strip; only rows with `page_count` pay the extra PDF parse; fail-soft on any re-gate exception (owner-confirmed 2026-10-07).
+- **A6 — default view = Text** for additive docs (no sparse-ratio default flip; the illustration descriptions already live in the text via phase 140) (owner-accepted default, 2026-10-07).
+- **A7 — the changed-path eligibility RULE is unchanged** (`fallback_fired or pictorial_described`, phase-148 lock intact); this phase changes only the persist ordering, the recovery, and the view consequence.
+
+## Commit
+```bash
+git add app/ frontend/ tests/ alembic/ .agents/phases/ && git commit --no-gpg-sign -m "fix(pdf): additive text+pages viewer for pictorial PDFs, write-first atomic page copies, and a deterministic fast-path re-gate that self-heals lost strip images"
+```

.agents/phases/complete/163_pdf_strip_additive_selfheal/01_schema_wire_strip_mode.md

@@ -0,0 +1,26 @@
+# Task 01 — The `page_strip_mode` column + wire signal (phase-163 D1, LOCKED A2/A3)
+
+**Phase:** `163_pdf_strip_additive_selfheal` · **Source:** phase overview D1 — the live modal showed the strip replacing the text of a genuinely textual book; the viewer needs an explicit class signal (fallback class = strip primary; pictorial class = additive) that never disturbs non-qualifying payloads.
+**Story:** n/a
+
+## Objective
+`documents.page_strip_mode` (VARCHAR(16), nullable, migration 0027 with a tested downgrade) exists and `DocContent` carries `page_strip_mode` with the `image_url` omission rule — ABSENT (never null) when `page_count` is NULL, present for qualifying docs — so the frontend (task 04) can branch on the class, and every non-qualifying doc's payload stays byte-identical to pre-phase.
+
+## Work
+1. `app/models.py` — `Document.page_strip_mode: Mapped[str | None]` (VARCHAR(16), nullable, default NULL), docstring right after `page_count` (phase 148's column): the strip CLASS — `"primary"` = the phase-133 fallback class (the strip IS the view — the screenshots are the content), `"additive"` = the phase-140 pictorial class (the text view + the strip behind a toggle); NULL = no strip. Stamped/cleared exactly with `page_count` (tasks 02/03 own the writes; this task only adds the column).
+2. `alembic/versions/0027_documents_page_strip_mode.py` — add the column; downgrade drops it. Follow `0026_documents_page_count.py` exactly (revision naming, `nullable=True`, no index). `uv run alembic upgrade head && uv run alembic downgrade -1 && uv run alembic upgrade head` round-trips.
+3. `app/schemas.py` — `DocContent` gains `page_strip_mode: str | None = None` beside `page_urls`, docstring: the phase-163 strip class; ABSENT (never null — the `image_url` omission rule, `exclude_none`-style serialization already in place for `image_url` — verify how `DocContent` drops `image_url=None` and apply the SAME mechanism) when there is no strip; `"primary"`/`"additive"` otherwise. A pre-phase row (column NULL while `page_count` is set) serializes ABSENT — the viewer treats absent as `"primary"` (task 04).
+4. `app/api/docs.py` — the content payload builder (the `page_urls=(…)` block, ~line 250): pass `page_strip_mode=doc.page_strip_mode` (the None-drops-the-key serialization handles the absent rule). The `/documents/{doc_id}/pages/{n}` route is NOT touched.
+5. Unit tests — extend the phase-148 payload-pin file (find where `page_urls` serialization is pinned — `tests/unit/test_docs_api.py` or `test_pdf_page_copies.py`):
+   - a qualifying doc with mode set: the serialized payload carries `"page_strip_mode": "primary"` / `"additive"` next to `page_urls`;
+   - a doc with `page_count` set and mode NULL (pre-phase row): the key is ABSENT (not null) and `page_urls` still present — the exact live `Letumdei` state;
+   - a non-qualifying doc (`page_count` NULL): NO `page_strip_mode` and NO `page_urls` keys — pin the payload byte-identity against the current expected dict;
+   - `page_strip_mode` never appears alone: with `page_count` NULL it is absent even if the column somehow holds a value (the builder keys the whole strip block off `page_count` — if the current builder does NOT, make it so).
+
+## Testing & Quality
+- Unit: the cases above; the existing `page_urls` pins stay green unmodified except where a test asserts an exact payload dict (then re-cut in this task). Coverage **>90%** on `app/`.
+
+## Completion Criteria
+- [ ] `uv run alembic upgrade head` applies 0027; downgrade removes the column; upgrade again is clean.
+- [ ] The three payload shapes (mode set / mode NULL with strip / no strip) serialize exactly per D1, pinned by unit tests.
+- [ ] Full `uv run pytest` green; `uv run ruff check . && uv run pyright` clean.

.agents/phases/complete/163_pdf_strip_additive_selfheal/02_persist_write_first.md

@@ -0,0 +1,32 @@
+# Task 02 — Write-first atomic page-copy persist + the strip-class stamp (phase-163 D2, LOCKED A4/A7)
+
+**Phase:** `163_pdf_strip_additive_selfheal` · **Source:** phase overview D2 — the live loss: the changed path deletes the old copies BEFORE rendering, outside the DB transaction, so any later per-file rollback (the `EmbeddingError` handler in the import loop) restores the stale `page_count` over ZERO files on disk (filesystem ops are not transactional). External wipes (`pdf_page_dir` cleaned) land the same state; the heal is task 03.
+**Story:** n/a
+
+## Objective
+`_persist_pdf_pages` never removes a live copy before the replacement bytes are durably written: render → atomic per-file write (tmp + `os.replace`) → stale-tail delete, and the function stamps `page_strip_mode` from a `strip_class` keyword instead of the bare `eligible` boolean — the eligibility RULE (`fallback_fired or pictorial_described`, phase-148 lock) is unchanged at the call site.
+
+## Work
+1. `app/rag/importer.py` — `_persist_pdf_pages(doc, *, full_path, page_dir, strip_class, source, rel)` (replaces the `eligible: bool` keyword):
+   - `strip_class` is `"fallback" | "pictorial" | None` (the caller passes the CLASS — see 2).
+   - `strip_class is None`: clear BOTH `doc.page_count = None` and `doc.page_strip_mode = None`, THEN `_delete_page_copies(doc, page_dir)` (the deletion-first order is correct HERE — there is no replacement set to protect; note the helper reads `doc.page_count`, so call it BEFORE nulling or pass the old count — check the current order and keep the deletion actually deleting: today the helper runs FIRST while the count is still set; preserve that behavior).
+   - eligible: `render_pdf_pages(full_path)` → the no-pages guard unchanged (`page_count`/mode NULL, warning) → `page_dir.mkdir` → **write loop FIRST**: every non-empty page to `<doc-id>-p{n}.png.tmp` then `os.replace` onto the final name (atomic per file — a reader never sees a torn PNG); collect the written names → **then** delete the stale set: `for n in range(1, doc.page_count + 1)` beyond/present-but-not-written → unlink (the shrunk-PDF tail; reuse `_delete_page_copies` semantics by deleting the old extent files that are NOT in the written set — a file about to be overwritten is not deleted at all).
+   - the failure path: on any exception in the render/write, unlink this call's `.tmp` AND final-name files it wrote (track both), then re-raise (the caller's fail-soft handler — `page_count` NULL + warning — is extended to also null `page_strip_mode`). The OLD copies are untouched by a failed call (they were never deleted up-front — the crash/rollback orphan state is structurally impossible now).
+   - success tail: `doc.page_count = len(pages)` AND `doc.page_strip_mode = "primary"` if `strip_class == "fallback"` else `"additive"`.
+   - docstring: the write-first rationale (the 2026-10-07 live loss: delete-first + a later rollback = the row over zero files) + the per-file atomicity note.
+2. `app/rag/importer.py` — the `_index_pdf_file` call site (~line 1671): `strip_class = "fallback" if fallback_fired else ("pictorial" if pictorial_described else None)` passed as the keyword; the `eligible=fallback_fired or pictorial_described` expression is GONE (the class carries the same truth table — pin it in the docstring: class non-None ⇔ eligible, the phase-148 rule verbatim). The `except Exception` fail-soft handler nulls mode too.
+3. `tests/unit/test_pdf_page_copies.py` — extend (the file's existing fixture pattern):
+   - **the order pin**: patch `_delete_page_copies`-class file ops / use a real tmp dir — after a successful re-render of a CHANGED doc, the new files exist and the stale tail (old count > new count) is gone; the overwritten pages keep the NEW bytes;
+   - **the failure pin**: a render that raises mid-write (the 2nd of 3 pages) → the FIRST render's old copies still on disk (the rollback-orphan state is unrepresentable), no `.tmp`/partial orphans from this call, the caller's fail-soft fires (`page_count`/mode NULL + the warning);
+   - **the atomic pin**: the writer uses tmp + `os.replace` (spy or a failing-replace test — a reader opening the final path never sees truncated bytes; assert the tmp name pattern is used and `os.replace` is called);
+   - **the class stamp**: fallback → `page_strip_mode == "primary"`; pictorial → `"additive"`; ineligible → both NULL and old copies deleted;
+   - the existing phase-148 pins (gap = per-page `b""` leaves no file and counts in `page_count`; the shrunk tail) stay green, re-cut only where the deletion order changes expectations.
+
+## Testing & Quality
+- Unit: the pins above; coverage **>90%** on `app/`.
+
+## Completion Criteria
+- [ ] A successful changed-path render: new copies durable BEFORE any old copy is removed (measured by the order pin); the shrunk tail still cannot leak.
+- [ ] A mid-write failure leaves the prior copies intact and zero orphans (files on disk exactly the prior set), `page_count`/`page_strip_mode` NULL, sync continues (the fail-soft unchanged).
+- [ ] `strip_class` carries the exact phase-148 eligibility truth table; no call site passes a bare `eligible` anymore.
+- [ ] `uv run pytest tests/unit/test_pdf_page_copies.py -v` green; full suite green; `uv run ruff check . && uv run pyright` clean.

.agents/phases/complete/163_pdf_strip_additive_selfheal/03_fast_path_regate.md

@@ -0,0 +1,37 @@
+# Task 03 — The deterministic fast-path re-gate: heal & repair (phase-163 D3, LOCKED A5)
+
+**Phase:** `163_pdf_strip_additive_selfheal` · **Source:** phase overview D3 — with the copies lost (external wipe, or the pre-phase-163 delete-first window), the phase-147 fast path assumes "the copies persist" and NEVER re-renders: the live `Letumdei` row (36 `page_urls`, zero files, unchanged bytes on every sync since) is the proof. The unchanged paths must re-gate the strip deterministically — no LLM/vision calls — heal missing copies, and clear strips that no longer qualify.
+**Story:** n/a
+
+## Objective
+`_regate_pdf_strip(session, doc, *, full_path, llm, summary)` runs on BOTH unchanged paths (the `file_hash` fast path and the content-digest-hit path), guarded by `doc.page_count is not None`: it mirrors the changed-path gate from the current bytes + settings, re-stamps `page_strip_mode`, re-renders when any copy is missing, or clears a stale strip — one greppable INFO line only when it ACTS, fail-soft on every exception.
+
+## Work
+1. `app/rag/importer.py` — new `_regate_pdf_strip(session, doc, *, full_path, llm, summary) -> None`:
+   - **mirror the changed-path gate exactly** (the truth table in the phase overview D3):
+     `text = extract_pdf_text(full_path)`; `is_usable = is_usable_pdf_text(text)`;
+     - `not is_usable` → class `fallback` (mode `"primary"`) — the phase-133 signal verbatim;
+     - elif `llm.settings.pdf_pictorial` and `pictorial_pages(extract_pdf_page_texts(full_path), min_chars=llm.settings.pdf_pictorial_min_chars)` non-empty → class `pictorial` (mode `"additive"`) — **the existing gate verbatim** (text-carrying pages under `min_chars` only; no-text pages are NOT candidates — `pictorial_pages`'s own contract; the re-gate adds no new eligibility semantics);
+     - else → **repair arm**: `_delete_page_copies(doc, page_dir)` (while `page_count` is still set — the helper reads it), then `doc.page_count = None`, `doc.page_strip_mode = None`; INFO `import: PDF strip re-gate source=… path=… — cleared stale strip`.
+   - **mode re-stamp** (eligible arms): a mode change (NULL→stamped, or the owner's kill-switch/class flip) writes `doc.page_strip_mode` and, when it changes an already-set mode, INFO `… mode <old>-><new>`; a NULL→stamp on existing rows is silent-beneficial (fold into the same line only when work follows).
+   - **copy verify + heal** (eligible arms): `page_dir = Path(llm.settings.pdf_page_dir).expanduser()`; missing = any of `page_dir/f"{doc.id}-p{n}.png"` for `n in 1..doc.page_count` absent. Any missing → `render_pdf_pages(full_path)`: the no-pages result (unopenable — pathological right after the fast path hashed the file) keeps the row untouched + WARNING `import: PDF strip re-gate source=… path=… — re-render produced no pages (state kept)`; else WRITE EVERY page via task-02's atomic writer (tmp + `os.replace`) — reuse it (extract the shared writer helper from task 02 rather than duplicating), then delete the stale tail beyond the new count; `doc.page_count = len(pages)` — a drift from the old count adds a WARNING naming both counts; INFO `… healed N pages`.
+   - **fail-soft**: the whole body in `try/except Exception` → `logger.warning` (the re-gate never fails a sync; the row keeps its pre-re-gate state — but note the repair arm's DB writes and the heal's row writes are in the caller's transaction, so an exception AFTER them rolls back naturally with the file anyway).
+   - docstring: the 2026-10-07 live state as the rationale; the zero-LLM-call property (render is CPU — A21-safe); the guard means every non-strip file's fast path keeps its zero-extra-read contract.
+2. `app/rag/importer.py` — `_unchanged_doc(...)`: gains the optional keyword `full_path: Path | None = None` (defaults to `Path(doc.full_path)`); at its tail, `if doc.page_count is not None and str(doc.path).lower().endswith(".pdf"): _regate_pdf_strip(session, doc, full_path=…, llm=llm, summary=summary)`. Call sites: the `_index_file` fast path passes the live `full_path`; the branches' content-digest-hit calls pass their own `full_path` (every call site already has it). Docstring: the phase-147 "the copies persist" paragraph is replaced — "unchanged files re-gate the PDF strip deterministically (phase 163)".
+3. `tests/unit/test_importer_unchanged_fast_path.py` — extend (the file's fast-path fixture pattern; the mock embedder asserts ZERO chat/vision calls across every case):
+   - **heal**: a stamped strip doc (`page_count=3`, mode NULL or set), copies deleted out from under it, unchanged bytes → sync → the 3 files exist (PNG bytes), mode stamped `"additive"` (usable layer + sparse candidate), INFO `healed 3 pages`, `page_count` unchanged (no drift);
+   - **fallback heal**: an unusable-layer doc → mode `"primary"`, copies restored;
+   - **no-op**: all copies present, mode set → zero file writes, zero INFO lines (pin "INFO never fires per unchanged doc" — assert the log records contain no re-gate line);
+   - **repair**: a row with `page_count` set (copies present) on a PDF whose gate now says ineligible (clean text, zero sparse pages) → cleared (both columns NULL, copies deleted), INFO `cleared stale strip`;
+   - **kill switch**: `pdf_pictorial=False` + pictorial-class doc → repair arm clears the strip (the changed-path semantics mirrored);
+   - **fail-soft**: `render_pdf_pages` patched to raise → sync completes, row keeps its prior state, WARNING, no copies touched;
+   - **guard**: a non-PDF unchanged file and a PDF with `page_count` NULL → `extract_pdf_text` NEVER called (spy) — the zero-extra-read contract intact.
+
+## Testing & Quality
+- Unit: the cases above; coverage **>90%** on `app/`.
+
+## Completion Criteria
+- [ ] Unchanged re-sync restores deleted copies (files + row + route-serveable) and stamps the mode — the live-loss state cannot persist past one sync (unit-proven; the E2E pins it in task 05).
+- [ ] A stale strip on a now-ineligible doc (incl. via the kill switch) is cleared with its copies; the re-gate fires ZERO LLM/vision calls and no INFO when there is nothing to do.
+- [ ] Non-strip fast paths perform zero PDF parsing (spy-pinned); every re-gate exception is fail-soft.
+- [ ] `uv run pytest tests/unit/test_importer_unchanged_fast_path.py tests/unit/test_pdf_page_copies.py -v` green; full suite green; `uv run ruff check . && uv run pyright` clean.

.agents/phases/complete/163_pdf_strip_additive_selfheal/04_viewer_additive_toggle.md

@@ -0,0 +1,35 @@
+# Task 04 — The additive viewer: text + Summary + the Text/Pages toggle (phase-163 D4, LOCKED A1/A6)
+
+**Phase:** `163_pdf_strip_additive_selfheal` · **Source:** phase overview D4 — the live complaint: "Notice how all the pages are unavailable… it shouldn't be a pictorial pdf." One short map page flipped a 114 KB-text book into the phase-148 strip-only view (the strip renders INSTEAD of the text content AND the Summary panel). The owner's call: additive for the pictorial class — the text is never hidden; the strip is a toggle away.
+**Story:** n/a
+
+## Objective
+`renderDocument` (the ONE shared core — the Sources modal AND `/document.html`) branches on `doc.page_strip_mode`: absent/`"primary"` keeps the current strip-only DOM byte-identically; `"additive"` renders the doc exactly as a text doc (Summary panel + content) plus a `Text`/`Pages` toggle (default `Text`) revealing the existing `docPagesStrip`.
+
+## Work
+1. `frontend/assets/document.js` — the `renderDocument` strip branch (~line 374, `const pageUrls = Array.isArray(doc.page_urls) ? doc.page_urls : null;`):
+   - `pageUrls` truthy AND `doc.page_strip_mode === "additive"`: fall through to the EXISTING text-doc rendering (the pre-148 branch — Summary panel + content, reused unchanged), THEN append the view toggle + the strip container:
+     - `<div class="doc-view-toggle" role="group" aria-label="Document view">` + two `<button type="button" class="doc-view-btn" aria-pressed>` — labels `Text` / `Pages`; initial state `Text` pressed (`"true"`), the strip NOT in the content flow;
+     - `Pages` click: `aria-pressed` flips; the text block + Summary panel hide (a class toggle, e.g. a `.doc-views[data-view="pages"]` wrapper — CSS-driven, never re-render), the `docPagesStrip(doc, pageUrls)` (the EXISTING builder, unmodified — lazy imgs never fetch while hidden) is appended once and shown;
+     - `Text` click reverses (the strip node stays mounted — no re-fetch on flip-flop);
+     - structure: wrap the text view + the strip in one `.doc-views` container appended to `contentEl` (both surfaces get it via the shared core; the wrapper exists ONLY on additive docs — every other doc's DOM is untouched).
+   - `pageUrls` truthy AND mode absent or `"primary"`: the CURRENT branch verbatim (strip only, no Summary panel, no toggle) — byte-identical DOM, pinned;
+   - falsy `pageUrls`: untouched.
+   - the never-build-HTML rule: createElement + properties/textContent only (the labels are fixed strings — no doc-derived text inside the toggle); the `strip_mode` read is `doc.page_strip_mode === "additive"` (anything else ⇒ primary — the tolerant-reading house rule).
+   - the module docstring: the phase-163 paragraph (the additive class contract, LOCKED A1/A6) replacing the phase-148 "INSTEAD of the text content AND the .doc-summary panel" sentence's scope (it now reads: primary class only).
+2. `frontend/assets/styles.css` — `.doc-view-toggle` (house button idiom beside the existing segmented/pill controls — check `.format-badge`/the modal action buttons for the closest idiom): the two buttons adjacent, pressed state via `aria-pressed="true"` styling (weight/background pair — **text + style, never color alone**, B5), `:focus-visible` 3px ring, ≥4.5:1 text contrast on both states, ≥44px touch targets on mobile (§7.1); `.doc-views[data-view="pages"]` hides the text block + Summary panel and shows the strip; default (`data-view="text"`) hides the strip container entirely.
+3. Unit (frontend source-contract — the house pattern; extend the file that already pins the phase-148 strip DOM — locate it via `docPagesStrip` in `tests/unit/`, likely `test_pdf_page_copies.py` / a viewer-contract file):
+   - the additive branch builds the `.doc-views` wrapper + the toggle with the exact labels/roles/`aria-pressed` initial state (source-regex or the file's established DOM-assertion style);
+   - the primary/absent branch contains NO `doc-view-toggle` string and the strip-only structure is byte-identical to the pre-phase pins;
+   - the toggle handler mutates only `aria-pressed` + the `data-view` attribute (no innerHTML anywhere in the new code path — the never-build-HTML pin);
+   - `loading = "lazy"` remains on the strip imgs (the existing pin re-runs green).
+
+## Testing & Quality
+- Unit: the contract pins above; coverage **>90%** on `app/` (no `app/` change in this task — the gate must simply stay green).
+- The phase's E2E (task 05) exercises the real DOM; `test_pdf_pages_in_sources.py` (the phase-148 E2E) is re-cut ONLY where a modal-DOM assertion meets the additive fixture (the fixture's mode stamp decides; prefer fixture docs that keep primary semantics for the existing pins).
+
+## Completion Criteria
+- [ ] An additive doc shows Summary + text with `Text` pressed by default; one click shows the strip (every page img loads); clicking back shows the text with zero refetch (the strip node persists — E2E-pinned in task 05).
+- [ ] A primary/absent-mode doc's viewer DOM is byte-identical to the pre-phase structure (unit pin green); a text doc and every non-PDF are untouched.
+- [ ] The toggle passes the WCAG basics: labeled group, `aria-pressed` state as text + style, focus-visible ring (checked against §7.2; the E2E asserts the attributes).
+- [ ] Full `uv run pytest` green; `uv run ruff check . && uv run pyright` clean.

.agents/phases/complete/163_pdf_strip_additive_selfheal/05_e2e_selfheal_pins.md

@@ -0,0 +1,33 @@
+# Task 05 — The dedicated E2E: additive view, primary byte-identity, heal & repair; the final gates (phase-163 D5)
+
+**Phase:** `163_pdf_strip_additive_selfheal` · **Source:** phase overview D5 + the AGENTS.md per-phase E2E rule (A16).
+**Story:** n/a
+
+**Pre-fix evidence (measured on the live server, 2026-10-07 — recorded here per house practice):** `Letumdei_ Heirs to the Imperium.pdf` — payload carries 36 `page_urls`; `GET /api/documents/b72f5515-dc40-4167-98ff-179a6503e6e5/pages/1,2,18,36` → **404/404/404/404** (page 37 → 404, the out-of-range arm); the content is 114 389 chars of genuine extracted text with 36 `---PAGE-N---` markers and exactly ONE `[Page 36 illustration: …]` block; the same server serves `besta-frame-white__AA-1272068-11-100.pdf` page 1 → **200** (the dir is intact; only the 10/1-indexed doc's copies were lost). The viewer showed "Page N unavailable." on every page with the text hidden.
+
+## Objective
+One Playwright suite proves the whole phase end to end — the additive view with the toggle, the fallback class's byte-identical strip-only DOM, the disk-level heal, and the repair arm — and the phase's final pass runs every gate plus the full documentation consistency review.
+
+## Work
+1. `tests/e2e/test_pdf_strip_additive_selfheal.py` — new suite (run in isolation; read `tests/e2e/test_pdf_pages_in_sources.py` and `tests/e2e/test_pictorial_pdf.py` first and reuse their synthetic-PDF builder + fixture-KB pattern; deterministic mock LLM, the default):
+   - **fixture**: source `strip_fixture` with (a) `manual.pdf` — usable text layer, 3 text pages + 1 sparse `<100-char` map-style page (the pictorial class — the Letumdei shape in miniature); (b) `scanned.pdf` — the garbage/unusable-layer fixture (the phase-133 pattern) → fallback class; (c) `plain.pdf` — clean text, zero sparse pages → no strip.
+   - **assert 1 (additive view)**: index → the modal for `manual.pdf`: mode stamped `additive` (via the content API), the text + Summary panel visible by default, the `Text`/`Pages` toggle present with `Text` `aria-pressed="true"`; click `Pages` → the strip renders, EVERY page `<img>` has `naturalWidth > 0` (routes serve — no "unavailable" note anywhere); click `Text` → text visible, strip hidden, the strip node still mounted (no refetch: assert the img count does not reset).
+   - **assert 2 (primary byte-identity)**: the modal for `scanned.pdf`: the strip-only DOM (no `.doc-view-toggle`, no Summary panel) and the "Page N" figures match the phase-148 structure (the existing suite's DOM assertions adapted); `plain.pdf` renders the plain text doc with no strip artifacts.
+   - **assert 3 (heal)**: for `manual.pdf`, delete ALL `pdf_page_dir/<id>-p*.png` (the fixture's page dir) + confirm a page route 404s → run a sync (unchanged bytes → the fast path) → the files exist, the page route 200s, the payload still carries `page_urls` + `additive`, the viewer's Pages view loads every img (the live-loss state, healed by the next sync — the exact D5 promise).
+   - **assert 4 (repair)**: direct-DB seed — set `page_count` (and stale copies on disk) on `plain.pdf`'s row (the stale-eligible state no current index produces) → sync → `page_count`/`page_strip_mode` NULL (read through the fixture session), the stale copies deleted, the modal shows the plain text (the strip gone).
+   - **assert 5 (no vision)**: the whole suite's syncs ran against the deterministic mock — no chat/vision endpoint hits beyond the initial index's own passes (the heal burns zero LLM calls; assert via the mock's call counter or the suite's request log, the house pattern).
+2. Final pass (the phase pipeline runs this after the tasks): full-suite + gates:
+   - `uv run pytest` green; `uv run pytest --cov=app --cov-report=term-missing` **>90%**;
+   - `uv run pytest tests/e2e/test_pdf_strip_additive_selfheal.py -v --no-cov` green **in isolation**; `tests/e2e/test_pdf_pages_in_sources.py`, `test_pictorial_pdf.py`, `test_garbage_pdf.py`, `test_document_viewer*.py`, `test_image_documents.py` green;
+   - `uv run ruff check . && uv run pyright` clean; the alembic 0027 up/down/up round-trip clean.
+3. Documentation & consistency (mandatory): review `README.md` and `AGENTS.md` IN FULL against the post-phase app — the viewer description gains the additive-mode sentence (text + Text/Pages toggle for pictorial-class PDFs; the fallback class keeps the strip), the importer notes (write-first copies, the fast-path re-gate + its INFO line) — and fix EVERY inconsistency found, including pre-existing ones not caused by this phase. `PLAN.md` is NOT touched (its §7 strip line folds at the next redo — recorded in the phase overview invariants).
+
+## Testing & Quality
+- This task IS the E2E + gates task; its own new-code coverage rides the suite (page route + fixture helpers are test-side; `app/` coverage stays **>90%**).
+
+## Completion Criteria
+- [ ] All five E2E asserts green in isolation; the four existing viewer/PDF suites green unmodified (except explicitly re-cut pins, each named in the commit body).
+- [ ] The heal assert proves: copies deleted externally → one unchanged-byte sync → files restored, routes serve, viewer complete (the live `Letumdei` failure mode is impossible after this phase).
+- [ ] `uv run pytest` fully green; coverage >90%; ruff + pyright clean; migration round-trip clean.
+- [ ] `README.md` + `AGENTS.md` reviewed in full, every inconsistency fixed (the additive-viewer sentence included); the phase dir moves to `complete/` and the single `--no-gpg-sign` commit (see the overview) is made by the pipeline gate.
+- [ ] Owner follow-up recorded (not code): after deploy, the next production Sync self-heals `Letumdei` (D5) — verify once via the Sources modal (text visible, Pages loads all 36) and `grep` the server log for the one `import: PDF strip re-gate … healed 36 pages` line.

.agents/phases/todo/164_chat_error_honesty/00_phase.md

@@ -0,0 +1,66 @@
+# Phase 164 — Chat failure honesty: the shared error-detail serializer, a distinct pre-stream failure line, and caps that accept a full-length answer
+
+**Source:** owner live report, 2026-10-07 — brain.reeseapps.com, right after the live answer cap was raised to `BOR_LLM_MAX_OUTPUT_TOKENS=131072`: when an answer hit the length limit, the next turn's bubble read *"My answer didn't make it — the connection dropped. Use Retry to ask again. Failed [object Object]."* The failure chain (traced in code, 2026-10-07): a long answer (> 32 000 chars) lives in `bor.chat.v1` and rides back to the server as `history` on the next `POST /api/chat`; `HistoryTurn.text` (`max_length=32_000`) rejects it with a **422 whose `detail` is a LIST OF OBJECTS** (`[{loc, msg, type}, …]`); the chat's pre-stream error path (`frontend/assets/app.js` ~3956: `if (body.detail) detail = body.detail`) has **no type guard** — `new Error(<array>)` stringifies to `"[object Object]"`, which the zero-frame settle shows as the fixed `FAILED_TURN_TEXT` ("the connection dropped" — untrue: the server WAS reachable and answered 422) plus the garbage detail. This is the SECOND occurrence of the same class: `HistoryTurn`'s own docstring records the 4 000 → 32 000 raise for an identical 422 (phase-42 E2E); the 32 000 ceiling broke again the moment outputs could exceed it — **at the DEFAULT `max_output_tokens=32_768` (~130k chars), not only at 131072**.
+
+**Story:** n/a (owner-directed bug fix; extends the phase-120 failed-turn contract, the phase-114 hint channel, the phase-83 SEC-05 boundary caps, the phase-74 history budgets).
+
+## Objective
+A failed turn states **what actually happened**: pre-stream HTTP rejections get their own honest line (not "the connection dropped") with the validation message serialized verbatim — never `[object Object]` — and a full-length answer (up to a 131 072-token cap ≈ 524k chars) flows through the whole chain: accepted as `history` on the next turn, and saveable/restorable as part of a saved chat. The anonymous saved-chat surface keeps a bounded worst case (the phase-83 SEC-05 property survives the raise).
+
+## Dependencies
+- `83_chat_save_payload_limits` (complete) — the SEC-05 caps this phase raises TOGETHER (`ChatMessage.text`/`thinking` mirror `HistoryTurn`; the mirror invariant must stay true).
+- `74`-line history budgets (complete; `history_to_messages`, `history_max_turns=40` / `history_max_chars=24_000`) — the LOCKED budget trim (owner 2026-09-08: overflow turns dropped WHOLE, never cut) — **untouched**; it is the real prompt-side gate and stays so.
+- `120_failed_turn_persistence` (complete; the `FAILED_TURN_TEXT` / `finalizeFailedTurn` / `.failed-note` contract) and `114` hint channel (`err.hint` → banner) — the bubble/note mechanics this phase extends.
+
+## Design (shared by all tasks — the executor reads this, not the chat)
+
+**D1 — the caps chain up, together (task 01, LOCKED A1).** `app/schemas.py`:
+- `HistoryTurn.text` / `HistoryTurn.thinking`: `max_length=600_000` (rationale IN the docstring: a full-cap answer at the owner's `BOR_LLM_MAX_OUTPUT_TOKENS` up to 131 072 ≈ 524 300 chars at ~4 chars/token — 600 000 is that with margin; the budget trim (locked A3, 2026-09-08) remains the real prompt gate — a turn over `history_max_chars` is dropped WHOLE, so a long answer trims itself out of later prompts; this phase fixes the 422, NOT the trim (its locked semantics are a separate owner decision, recorded in the overview));
+- `ChatRequest.history` list: `max_length=100` → `500` (the same "DoS sanity ceiling only" docstring, now also noting the client sends the whole `bor.chat.v1` minus the current question);
+- `ChatMessage.text` / `ChatMessage.thinking`: `max_length=600_000` (MIRRORS `HistoryTurn` — the phase-83 invariant "a saved chat can never legitimately carry more than the chat path accepts" must stay TRUE: raise both or neither);
+- `SavedChatCreate.messages` / `SavedChatUpdate.messages`: `max_length=200` → `500` (symmetry with the history list) PLUS a new `model_validator` on both: **total chars across all messages (`len(text) + len(thinking or "")`) ≤ 8_000_000** — this bounds the ANONYMOUS SEC-05 surface at the same order as the audited pre-phase worst case (200 × 64k ≈ 12.8 MB) while fitting ~15 full-cap answers; the validator error surfaces as the standard 422 (`ValueError` message names the budget);
+- `ChatRequest.message`: `max_length=4000` UNCHANGED (the composer textarea carries the cap client-side — pin it).
+
+**D2 — the shared `errorDetailText` serializer (task 02, LOCKED A2).** `frontend/assets/app.js` exports `errorDetailText(value, fallback)` (exported const/function, the `TURN_TIMEOUT_MS` unit-pin idiom): a non-empty **string** → verbatim; an **array** → per-entry `entry.msg` when the entry is an object else `String(entry)`, empties dropped, `"; "`-joined, all-empty → `fallback`; an **object** → `String(entry.msg)` when present, else `fallback`; nullish / empty / anything else → `fallback`. NEVER `JSON.stringify` a raw detail (the FastAPI 422 entries carry the FULL `input` — the whole conversation — which must never reach the DOM; only `msg` and plain strings serialize). Applied at the TWO unguarded chat paths: the pre-stream `!res.ok` handler (~3956 — `detail = errorDetailText(body.detail, \`Brain's API answered with HTTP ${res.status}.\`)`) and the SSE `error`-frame throw (~4353 — `new Error(errorDetailText(ev.detail, "Something went wrong on my side."))`, the `err.hint` line untouched). The seven EXISTING guarded copies (`app.js` tuning form, `tuning.js`, `theme.js`, `document.js`, `git-sources.js`, `doc-edit.js`) are behaviorally correct and stay byte-identical (no opportunistic refactor — the new helper matches their semantics for the shared shapes, generalized to arrays-of-strings and multi-entry joins).
+
+**D3 — the distinct pre-stream failure line (task 03, LOCKED A3/A4).** `app.js`:
+- `const FAILED_REQUEST_TEXT = "I couldn't take that — my API rejected the request. Use Retry to ask again.";` (sibling of `FAILED_TURN_TEXT`, its docstring: the PRE-STREAM HTTP-rejection settle — the server answered, the request did not pass; the real detail rides the `.failed-note` + persisted `error` key exactly as `FAILED_TURN_TEXT` does);
+- the pre-stream `!res.ok` throw carries `err.bubbleText = FAILED_REQUEST_TEXT` and `err.hint = "The app answered — it rejected the request before the model ran. Nothing was lost but this send."` (the phase-114 banner-hint channel);
+- `runTurn`'s catch passes `bubbleText: err?.bubbleText` into `finalizeFailedTurn` (new options key); the zero-frame settle uses `bubbleText || FAILED_TURN_TEXT` — the persisted record text stays "what the user saw is what is stored";
+- UNCHANGED settles: a transport-level failure (the `fetch` itself rejects — `TypeError`) has no `bubbleText` → keeps `FAILED_TURN_TEXT` (that one IS the connection); the SSE `error`-frame zero-frame path (a model-side failure) keeps `FAILED_TURN_TEXT`; the stop / abort paths are untouched. A partial-stream settle (`wrap` + `acc`) never shows a bubble text at all — unchanged.
+
+**D4 — verification (task 04).** Dedicated Playwright suite `tests/e2e/test_chat_error_honesty.py` (isolation; deterministic mock LLM) with (a) a `page.route` stub of `POST /api/chat` returning the EXACT FastAPI 422 shape for a long-history turn → assert the bubble reads `FAILED_REQUEST_TEXT`, the `.failed-note` detail contains the pydantic `msg` verbatim ("String should have at most 600000 characters" — never `[object Object]`, never `input` text), the banner hint shows, Retry + Edit land, composer re-enables (§7.4); (b) a 500 stub with a non-JSON body → the `HTTP 500` default string line (the serializer's fallback path); (c) a real-server acceptance: send a turn whose history carries a ~525k-char prior brain answer → the request is ACCEPTED (no 422) and the turn settles normally with the mock (the budget trim may drop the giant turn WHOLE from the prompt — the locked A3 behavior; the test asserts acceptance + settle, not prompt content); (d) a stubbed SSE `error` frame carrying an ARRAY detail → the note serializes (client tolerance). Server-side acceptance/422 matrix for every raised cap + the total-chars validator rides integration pins in task 01 (the `test_chats_api.py` pattern).
+
+**Invariants (all tasks).** The budget trim (`history_to_messages`, budgets, whole-drop, contiguous window) is byte-identical (locked A3 2026-09-08). `ChatRequest.message` 4000. The `FAILED_TURN_TEXT` wording, the phase-120 persistence keys (`failed`, `error` ≤500), Retry/Edit pairing, the SSE frame shapes, and the seven guarded frontend detail-readers are byte-identical. `PLAN.md` untouched.
+
+## Tasks
+1. `01_caps_long_answer_chain.md` — the `schemas.py` raise + the total-chars validator + docstrings + the unit/integration acceptance-rejection matrix (D1).
+2. `02_error_detail_serializer.md` — the exported `errorDetailText` + the two chat call sites + unit pins (D2).
+3. `03_prestream_failure_copy.md` — `FAILED_REQUEST_TEXT` + the `bubbleText` threading + the banner hint + the failed-turn pins (D3).
+4. `04_e2e_error_honesty.md` — the dedicated E2E (D4) + the final gates + the docs consistency pass (incl. the proxy body-size deploy note).
+
+## Testing & Quality
+- Unit: `tests/unit/test_schemas.py` (or the file pinning `HistoryTurn`/`ChatMessage` bounds — locate via `32_000` in `tests/`): every cap's boundary (600_000 accept / +1 reject; 500 vs 501 turns; the total-8M validator: accept ~15 full-cap messages, reject beyond), the 422 `detail` shapes; `tests/unit/test_frontend_error_detail.py` (new; the `test_frontend_*` house pattern): all serializer shapes incl. the exact live FastAPI 422 body; `tests/unit/test_frontend_failed_turn.py` extended: the `bubbleText` threading (pre-stream vs transport vs SSE-frame settles), the never-`innerHTML` rule.
+- Integration: `tests/integration/test_chats_api.py` extended — a saved chat carrying ONE 525k-char answer saves 201 (the phase-83 surface stays bounded by the total validator), over-total rejects 422.
+- E2E: `tests/e2e/test_chat_error_honesty.py` green **in isolation**; the phase-120 failed-turn E2E stays green (the `FAILED_TURN_TEXT` path unchanged for transport failures).
+- Coverage **>90%** on `app/`; `uv run ruff check . && uv run pyright` clean.
+
+## Completion Criteria
+- [ ] A pre-stream 422 shows the distinct honest line + the validation `msg` verbatim in the note — `[object Object]` is structurally impossible on the chat path (unit + E2E pinned; no raw `input` reaches the DOM).
+- [ ] A history turn with a 525k-char brain answer is ACCEPTED (200 turn settles); >600k rejects 422; the saved-chat chain accepts the same answer inside a 201 and the anonymous surface stays bounded (total-8M validator pinned).
+- [ ] Transport failures and SSE error frames keep `FAILED_TURN_TEXT` verbatim; stop/abort paths byte-identical; §7.4 failure feedback re-enables every control.
+- [ ] The budget trim, `message` 4000, the seven guarded detail-readers, and all SSE shapes are byte-identical; full `uv run pytest` green; coverage >90%; ruff + pyright clean.
+- [ ] `README.md` + `AGENTS.md` reviewed in full in the final pass — includes the deploy note: the proxy `client_max_body_size` must accommodate the now-legal large chat POSTs (recommend ≥ 16 MB) — and every other inconsistency fixed.
+- [ ] One `--no-gpg-sign` commit; the phase dir moves to `complete/` by the pipeline gate.
+
+## Locked decisions
+- **A1 — the caps chain together**: `HistoryTurn.text/thinking` AND `ChatMessage.text/thinking` → 600_000 (≈ 131 072 × 4 chars with margin); history list → 500; saved-messages list → 500 with a total-chars ≤ 8_000_000 model validator (the SEC-05 bounded-anonymous property preserved at the audited order); `message` 4000 unchanged; the phase-83 mirror invariant explicitly honored (owner-confirmed 2026-10-07 — "static generous caps", with the SEC-05 coupling added by this phase's audit of the chain — raise ONE side would break saves).
+- **A2 — `errorDetailText`**: the shared serializer's exact semantics (string verbatim; array entries `msg`-else-String joined `"; "`; object `msg`-else-fallback; never `JSON.stringify`; never the `input`), exported from `app.js`, applied at the two unguarded chat paths ONLY; the seven guarded copies untouched (owner-confirmed 2026-10-07).
+- **A3 — `FAILED_REQUEST_TEXT`**: distinct pre-stream HTTP-rejection bubble text verbatim; the persisted record is what the user saw (the phase-120 contract extended, not changed); transport failures keep `FAILED_TURN_TEXT` (owner-confirmed 2026-10-07 — "distinct pre-stream line").
+- **A4 — the budget trim is NOT touched**: the locked 2026-09-08 whole-drop contiguous-window semantics stand; the consequence — a >24k-char answer is dropped whole from later prompts (follow-ups WORK with a trimmed window, without that answer's context) — is recorded here; making long answers useful in later prompts (bigger `BOR_HISTORY_MAX_CHARS` is pure env, or a smarter cut) is an owner call outside this phase.
+- **A5 — PLAN.md untouched** (the caps rationale lives in the schema docstrings; folded at the next redo).
+
+## Commit
+```bash
+git add app/ frontend/ tests/ .agents/phases/ && git commit --no-gpg-sign -m "fix(chat): honest pre-stream failure line, a shared error-detail serializer, and history caps that accept a full-length answer"
+```

.agents/phases/todo/164_chat_error_honesty/01_caps_long_answer_chain.md

@@ -0,0 +1,30 @@
+# Task 01 — The caps chain up: 600k text, 500 turns, a bounded anonymous total (phase-164 D1, LOCKED A1/A4)
+
+**Phase:** `164_chat_error_honesty` · **Source:** phase overview D1 — the live 422: a >32k-char answer rode back as `history` and rejected the NEXT turn. `HistoryTurn`'s docstring already records this failure class happening once (4 000 → 32 000, phase-42 E2E); the ceiling broke again the moment outputs could exceed 32k — at the DEFAULT `max_output_tokens=32_768` too.
+**Story:** n/a
+
+## Objective
+Every schema cap in the long-answer chain (`HistoryTurn` → `ChatRequest.history` → `ChatMessage` → saved-chat lists) accepts a full-cap answer (~524k chars at `max_output_tokens=131072`), the phase-83 mirror invariant between the chat path and the saved-chat path stays TRUE, and the anonymous saved-chat worst case stays bounded by a new total-chars validator.
+
+## Work
+1. `app/schemas.py` — `HistoryTurn`: `text`/`thinking` `max_length=32_000`/`32000` → **600_000** both; docstring: the cap accepts a full-length answer (`BOR_LLM_MAX_OUTPUT_TOKENS` up to 131 072 ≈ 524 300 chars at ~4 chars/token, plus margin — the 2026-10-07 live 422: the 32 000 ceiling rejected the FOLLOW-UP turn of a long answer, the same class as the recorded 4 000 → 32 000 raise); the real gate stays the locked budget trim (`history_to_messages` — overflow turns dropped WHOLE; a long answer trims itself out of later prompts, never a 422).
+2. `app/schemas.py` — `ChatRequest.history`: `max_length=100` → **500**; the docstring sentence updates: still "a DoS sanity ceiling only — the config budgets do the real trimming (A3)", plus: the client sends the whole `bor.chat.v1` minus the current question, so the ceiling must clear long conversations.
+3. `app/schemas.py` — `ChatMessage.text`/`thinking`: → **600_000** (docstring: the phase-83 SEC-05 mirror RAISED with `HistoryTurn` — the invariant "a saved chat can never legitimately carry more than the chat path accepts" survives the raise; raise ONE side and every done-time auto-save of a long answer 422s into the phase-113 quiet-failure).
+4. `app/schemas.py` — `SavedChatCreate.messages` / `SavedChatUpdate.messages`: `max_length=200` → **500** AND a `model_validator(mode="after")` on each: total chars `sum(len(m.text) + len(m.thinking or "") for m in messages) <= 8_000_000` else `raise ValueError("saved chat exceeds the 8,000,000-character budget")` — docstring: bounds the ANONYMOUS SEC-05 worst case at the audited order (the pre-phase 200×64k ≈ 12.8 MB) while fitting ~15 full-cap answers; the 422 is the framework's standard validation response (the phase-83 A2 boundary-only style — NO route changes).
+5. `ChatRequest.message`: **unchanged** `max_length=4000` (the composer textarea caps client-side — `app.js` ~1122 comment; pin it stays).
+6. Unit — extend the file pinning these bounds (locate via `32_000` under `tests/unit/`):
+   - `HistoryTurn.text` 600_000 accept / 600_001 reject; same for `thinking`;
+   - `history` 500 accept / 501 reject; `message` 4000 boundary re-pinned unchanged;
+   - `ChatMessage.text/thinking` 600_000 accept / +1 reject;
+   - messages 500 accept; the total validator: ~15 × 524k messages REJECT with the budget-naming `ValueError` surfaced in the 422 `detail`, a saved chat with ONE 525k answer + normal turns ACCEPTS;
+   - the 422 `detail` SHAPE of a cap rejection is the pydantic list-of-objects (the serializer's task-02 input — pinned here as the server-side half of the contract).
+7. Integration — `tests/integration/test_chats_api.py`: `POST /api/chats` with a one-525k-answer conversation → 201 (round-trips through JSONB — the DB is JSONB, no column changes needed — verify); over-total → 422; the existing phase-83 boundary matrix green (re-cut only the pins that assert the literal `32_000`/`200` values, naming each in the task's completion).
+
+## Testing & Quality
+- Unit + integration above; coverage **>90%** on `app/`.
+
+## Completion Criteria
+- [ ] The full-cap-answer conversation flows: accepted as `/api/chat` history (integration-level) AND saveable as a saved chat (201), with the 8M total cap the only new anonymous bound.
+- [ ] `ChatMessage` mirrors `HistoryTurn` exactly (the phase-83 invariant re-stated in both docstrings, verified by identical-boundary tests).
+- [ ] The budget-trim suites (`history_to_messages`) are untouched and green — D1 changes what ARRIVES, never what the trim keeps.
+- [ ] `uv run pytest tests/unit tests/integration -v` green; full suite green; `uv run ruff check . && uv run pyright` clean.

.agents/phases/todo/164_chat_error_honesty/02_error_detail_serializer.md

@@ -0,0 +1,35 @@
+# Task 02 — The shared `errorDetailText` serializer for the two chat paths (phase-164 D2, LOCKED A2)
+
+**Phase:** `164_chat_error_honesty` · **Source:** phase overview D2 — the live "[object Object]": `frontend/assets/app.js` ~3956 does `if (body.detail) detail = body.detail` on a FastAPI 422 whose `detail` is a LIST OF OBJECTS; `new Error(<array>)` → message `"[object Object]"`. The SSE `error`-frame path (~4353 `new Error(ev.detail || …)`) is the second unguarded reader. The guarded idiom ALREADY exists in seven copies (`tuning.js`, `theme.js`, `document.js`, `git-sources.js`, `doc-edit.js`, `app.js`'s tuning form) — this task generalizes it ONCE and applies it ONLY where it is missing.
+**Story:** n/a
+
+## Objective
+`app.js` exports `errorDetailText(value, fallback)` with exact, total semantics (every input shape yields a human string or the fallback — `[object Object]` and raw `input` leakage are structurally impossible), wired at the pre-stream HTTP-error handler and the SSE `error`-frame throw; the seven existing guarded copies stay byte-identical.
+
+## Work
+1. `frontend/assets/app.js` — new exported function (the `THINKING_NEAR_BOTTOM_PX` / `TURN_TIMEOUT_MS` export-for-unit-pin idiom; place beside the other error constants):
+   ```
+   errorDetailText(value, fallback) → string
+   ```
+   - non-empty **string** → the string verbatim;
+   - **array** → per entry: `String(entry.msg)` when the entry is an object with a non-empty `msg`, else (non-object entry) `String(entry)`; entries with neither drop; join with `"; "`; empty result → `fallback`;
+   - **object** → `String(value.msg)` when a non-empty `msg` exists, else `fallback`;
+   - nullish / empty string / any other type → `fallback`;
+   - NEVER `JSON.stringify` anything, NEVER touch `input`/`loc`/`ctx` (the FastAPI entries carry the full rejected input — the whole conversation — which must never reach the DOM). Docstring names the live 2026-10-07 `[object Object]` failure and the `input`-leak rule.
+2. `frontend/assets/app.js` — the pre-stream handler (~3956, the `!res.ok || !res.body` block): replace the unguarded assignment with `const text = errorDetailText(body.detail, \`Brain's API answered with HTTP ${res.status}.\`); throw new Error(text);` (the `try`-around-`res.json` and the non-JSON catch keep their current shape; the default per-status string remains the fallback). The `bubbleText`/hint threading is task 03 — do NOT add it here.
+3. `frontend/assets/app.js` — the SSE `error` frame (~4353): `const err = new Error(errorDetailText(ev.detail, "Something went wrong on my side."));` — the `err.hint = ev.hint;` line byte-identical. A string `detail` (every server-built frame, `ChatErrorEvent`) serializes VERBATIM — pin that the user-visible strings for "Question too long", "dropped the connection" etc. are unchanged.
+4. `tests/unit/test_frontend_error_detail.py` (new — read `tests/unit/test_frontend_failed_turn.py` first and reuse its harness/assertion style):
+   - the EXACT live 422 body (`{"detail":[{"type":"string_too_long","loc":["body","history",5,"text"],"msg":"String should have at most 600000 characters","input":"<huge>"}]}`) → the `msg` verbatim, and the result contains NO `input` text and no `"[object Object]"`;
+   - multi-entry array → the `msg`s joined `"; "`; mixed entries (object + plain string) → joined; all-empty array → fallback;
+   - object with `msg` → verbatim; object without → fallback;
+   - string → verbatim (incl. the server's fixed copy — pinned); `""`/null/undefined/number → fallback;
+   - source-contract pins: the two call sites now read through `errorDetailText`; `JSON.stringify` appears NOWHERE in the new helper; the seven guarded copies are byte-identical (hash their regions or assert the untouched lines).
+
+## Testing & Quality
+- Unit: the shape matrix above (the harness the `test_frontend_*` files use); no `app/` change — coverage gate stays green.
+
+## Completion Criteria
+- [ ] Every serialized shape yields a human string or the named fallback — `[object Object]` is unrepresentable at the two chat call sites (unit-proven).
+- [ ] A 422 `detail` never puts request `input` text into the DOM (pinned); server string frames render byte-identical copy.
+- [ ] The seven existing guarded readers are untouched; the `err.hint` channel is untouched.
+- [ ] Full `uv run pytest` green; `uv run ruff check . && uv run pyright` clean.

.agents/phases/todo/164_chat_error_honesty/03_prestream_failure_copy.md

@@ -0,0 +1,28 @@
+# Task 03 — The distinct pre-stream failure line: `FAILED_REQUEST_TEXT` (phase-164 D3, LOCKED A3)
+
+**Phase:** `164_chat_error_honesty` · **Source:** phase overview D3 — the live bubble said *"My answer didn't make it — the connection dropped"* for a **422**: the server was reachable and answered deliberately; the copy lied, and the owner-confirmed remedy is a distinct honest line for pre-stream HTTP rejections.
+**Story:** n/a
+
+## Objective
+A pre-stream HTTP rejection (the `!res.ok` settle) bubbles `FAILED_REQUEST_TEXT` — "I couldn't take that — my API rejected the request. Use Retry to ask again." — with the serialized detail on the `.failed-note` and the phase-114 hint on the banner; transport failures (the `fetch` itself rejects) and SSE `error` frames keep `FAILED_TURN_TEXT` byte-identically; the phase-120 persistence contract ("what the user saw is what is stored") extends to the new text.
+
+## Work
+1. `frontend/assets/app.js` — new constant beside `FAILED_TURN_TEXT`: `const FAILED_REQUEST_TEXT = "I couldn't take that — my API rejected the request. Use Retry to ask again.";` docstring: the PRE-STREAM HTTP-rejection settle (the server ANSWERED with a 4xx/5xx — the request died before the model; the 2026-10-07 live 422-after-a-long-answer is the finding); the real detail rides the in-bubble `.failed-note` + the persisted `error` key exactly as `FAILED_TURN_TEXT`'s does; NOT for transport failures (no answer happened at all — `FAILED_TURN_TEXT` is honest there).
+2. `frontend/assets/app.js` — the pre-stream throw (task 02's site): attach `err.bubbleText = FAILED_REQUEST_TEXT;` and `err.hint = "The app answered — it rejected the request before the model ran. Nothing was lost but this send.";` to the thrown `Error` (the phase-114 `hint` channel already flows from `err.hint` to `showErrorBanner` via the catch's `setUiState` opts — verify, do not duplicate).
+3. `frontend/assets/app.js` — `finalizeFailedTurn(detail, { acc, thinking, tools, segments, wrap, leavePartialIndex, bubbleText })`: the ZERO-frame branch (`else` — no `wrap`) renders `bubbleText || FAILED_TURN_TEXT` (the `addMessage` text AND the `rememberBrainTurn` record text — same value, the stored-what-you-saw contract) and the note/persistence/Retry/Edit machinery is otherwise untouched (partial-stream settles never receive a `bubbleText` by construction — a pre-stream throw has no `wrap`; still, the partial branch must ignore the key: read it ONLY in the zero-frame branch). Docstring update (the `bubbleText` key, its caller-only meaning).
+4. `frontend/assets/app.js` — `runTurn`'s catch: pass `bubbleText: err?.bubbleText` into the `finalizeFailedTurn` options (the ONE line-level change; the detail/hint/banner sequencing after the call is untouched).
+5. `tests/unit/test_frontend_failed_turn.py` (extend; the file's harness):
+   - a thrown `Error` with `bubbleText` → the zero-frame bubble + persisted record text are `FAILED_REQUEST_TEXT`, note = the detail, `failed: true` + the `error` key unchanged in shape (≤500 cap intact);
+   - a thrown `Error` WITHOUT `bubbleText` (the transport-failure shape) → `FAILED_TURN_TEXT` byte-identical (the existing pins re-run green);
+   - the SSE-frame throw (task 02's site) sets NO `bubbleText` → model-side failures keep `FAILED_TURN_TEXT`;
+   - the banner: `err.hint` reaches the opts the same way for both throw kinds (the phase-114 pin re-runs);
+   - `FAILED_REQUEST_TEXT`'s exact string pinned ONCE (the copy is owner-locked); the never-`innerHTML` rule holds.
+
+## Testing & Quality
+- Unit: the pins above; the phase-120 suite green WITHOUT modification (the default path byte-identical); coverage gate unchanged (no `app/` code).
+
+## Completion Criteria
+- [ ] Pre-stream HTTP error → bubble `FAILED_REQUEST_TEXT`, note carries the serialized detail (task 02's serializer), hint on the banner, Retry + Edit land, composer re-enables (§7.4).
+- [ ] Transport failures and SSE error frames are byte-identical to the phase-120 behavior (`FAILED_TURN_TEXT`, same persistence).
+- [ ] The persisted `bor.chat.v1` record for a new-line failure stores the new text + `failed: true` + the capped `error` — a refresh restores it identically (the same property the existing suite pins for `FAILED_TURN_TEXT`).
+- [ ] Full `uv run pytest` green; `uv run ruff check . && uv run pyright` clean.

.agents/phases/todo/164_chat_error_honesty/04_e2e_error_honesty.md

@@ -0,0 +1,30 @@
+# Task 04 — The dedicated E2E: error honesty + long-answer acceptance; the final gates (phase-164 D4)
+
+**Phase:** `164_chat_error_honesty` · **Source:** phase overview D4 + the AGENTS.md per-phase E2E rule (A16).
+**Story:** n/a
+
+**Pre-fix evidence (2026-10-07, recorded per house practice):** owner-live instance after raising `BOR_LLM_MAX_OUTPUT_TOKENS=131072` — the post-long-answer turn's bubble read *"My answer didn't make it — the connection dropped. Use Retry to ask again. Failed [object Object]."*; code trace: the 422's list-of-objects `detail` hit the unguarded `body.detail` assignment (`app.js` ~3956) → `new Error(<array>)` → `"[object Object]"` (the example was not preserved live — the cap raise removed the repro; the 32 000 ceiling repros it at DEFAULT settings with any >32k-char answer). `HistoryTurn`'s own docstring records the identical failure class once already (4 000 ceiling, phase-42 E2E).
+
+## Objective
+One Playwright suite proves the phase end to end against the REAL client code paths — the distinct pre-stream line with the serialized validation message (never `[object Object]`, never `input` leakage), the fallback and tolerance paths, and the real-server acceptance of a full-length-answer history — and the final pass runs every gate plus the full documentation consistency review.
+
+## Work
+1. `tests/e2e/test_chat_error_honesty.py` (new; isolation; deterministic mock LLM; read `tests/e2e/test_frontend_failed_turn*.py`-adjacent suites and any existing `page.route` usage first and reuse the established stubbing approach — if no route-stub precedent exists, intercept `POST /api/chat` with Playwright `page.route` fulfilling the fetch; the assertions must drive the real `app.js` send path):
+   - **assert 1 (the live bug)**: stub `POST /api/chat` → 422 with the EXACT FastAPI shape for a long-history rejection (`{"detail":[{"type":"string_too_long","loc":["body","history",5,"text"],"msg":"String should have at most 600000 characters","input":"x".repeat(1000)}]}`) → send a question: the bubble text is `FAILED_REQUEST_TEXT` verbatim; the `.failed-note` detail contains the `msg` substring; NO `[object Object]` anywhere in the DOM; the rejected `input` text is NOT in the DOM; the banner shows the hint; Retry + Edit are present; the composer is re-enabled (§7.4).
+   - **assert 2 (fallback)**: stub → 500 with a NON-JSON body → the note/line settle on the `HTTP 500` default string, `FAILED_REQUEST_TEXT` bubble (still pre-stream), no crash.
+   - **assert 3 (acceptance — the D1 fix)**: seed `localStorage` `bor.chat.v1` with a conversation whose LAST brain record carries a ~525k-char answer (the seeded-restore pattern the saved-chat suites use), reload, send a short question → the request is NOT 422 (assert the response status 200 via the route/network log) and the turn settles normally with the mock answer (the locked budget trim may drop the giant turn WHOLE from the prompt — asserting settle, never prompt content, per A4).
+   - **assert 4 (SSE-frame tolerance)**: stub → a valid SSE stream whose `error` frame carries an ARRAY `detail` → the note serializes the `msg` (client tolerance; server-built string frames are pinned in task 02 units).
+2. Final pass (the phase pipeline runs this after the tasks): full-suite + gates:
+   - `uv run pytest` green; `uv run pytest --cov=app --cov-report=term-missing` **>90%**;
+   - `uv run pytest tests/e2e/test_chat_error_honesty.py -v --no-cov` green **in isolation**; the phase-120/114 failed-turn + banner suites and `test_chats_api.py` green (re-cut pins named in the commit body);
+   - `uv run ruff check . && uv run pyright` clean.
+3. Documentation & consistency (mandatory): review `README.md` and `AGENTS.md` IN FULL against the post-phase app — add the **deploy note**: *the proxy in front of the app must accept chat POSTs sized to the new caps (recommend `client_max_body_size ≥ 16m` — a follow-up turn after a full-cap answer posts ~0.5 MB+, a saved chat up to the 8 MB budget); a proxy with a smaller limit turns a legal request into a 413* — review the error-handling descriptions for accuracy (the failed-turn copy is documented where phase 120's bubbles are), and fix EVERY inconsistency found, including pre-existing ones not caused by this phase. `PLAN.md` untouched (A5).
+
+## Testing & Quality
+- This task IS the E2E + gates task; its fixtures are test-side; `app/` coverage stays **>90%**.
+
+## Completion Criteria
+- [ ] All four asserts green in isolation; assert 1 proves the exact live message renders honestly end to end (bubble + note + hint + controls).
+- [ ] The seeded 525k-answer conversation sends, is accepted, and settles — the live "every follow-up dies after a long answer" failure mode is impossible after this phase.
+- [ ] `uv run pytest` fully green; coverage >90%; ruff + pyright clean; no existing suite modified except re-cut pins named in the commit body.
+- [ ] `README.md` + `AGENTS.md` reviewed in full, the proxy body-size deploy note added, every inconsistency fixed; the phase dir moves to `complete/` and the single `--no-gpg-sign` commit (overview) is made by the pipeline gate.

.agents/reports/163_pdf_strip_additive_selfheal/163_pdf_strip_additive_selfheal__00_phase.a1.md

@@ -0,0 +1,13 @@
+All gates verified. Final report:
+
+**Phase 163 — final verification pass (all 5 tasks already complete; verification-only run)**
+
+- Verified D1–D4 in code: `page_strip_mode` column + migration 0027, `DocContent` absent-never-null rule keyed off `page_count`; `_persist_pdf_pages` render → atomic tmp+`os.replace` writes → stale-tail-last + `strip_class` stamp; `_regate_pdf_strip` on both unchanged paths (mirror gate, heal, repair, fail-soft, one INFO line only when acting); viewer mode branch + Text/Pages toggle (Text default, `aria-pressed`, no innerHTML, lazy imgs).
+- `uv run pytest --cov=app --cov-report=term-missing` → **3796 tests, exit 0, coverage 99%** (>90%)
+- `uv run pytest tests/e2e/test_pdf_strip_additive_selfheal.py -v --no-cov` → **5 passed** in isolation
+- `test_pdf_pages_in_sources + test_pictorial_pdf + test_garbage_pdf` → 10 passed; `test_document_viewer(_context) + test_image_documents` → 20 passed
+- `uv run ruff check .` clean; `uv run pyright` → 0 errors; alembic **0027 downgrade→upgrade CLI round-trip clean** (+ integration pins)
+- Completion criteria: additive view w/ toggle ✅ · primary DOM byte-identity ✅ · write-first order/mid-write-failure/atomic pins ✅ · heal + repair E2E ✅ · zero vision/LLM on fast path (mock-captured) ✅ · non-qualifying payload byte-identity ✅ · docs pass ✅
+- **Documentation**: README.md + AGENTS.md reviewed in full — the phase's updates are accurate; all config-table defaults cross-checked against `app/config.py` (no drift), every page route, script, and E2E command referenced exists, every `.env.example` var documented — **verified consistent, no changes needed**.
+- Notable: live `Letumdei` repair is owner-runtime post-deploy per D5 (next production Sync self-heals; no live changes made); the modified `.agents/screenshots/doc_modal_scrollbar_1280.png` is an E2E-regenerated artifact left for the pipeline commit.
+- **Next pending phase:** `164_chat_error_honesty`

.agents/reports/163_pdf_strip_additive_selfheal/163_pdf_strip_additive_selfheal__00_phase.a1.md.status

@@ -0,0 +1 @@
+ok

.agents/reports/163_pdf_strip_additive_selfheal/163_pdf_strip_additive_selfheal__00_phase.a1.validate

@@ -0,0 +1,124 @@
+........................................................................ [  1%]
+........................................................................ [  3%]
+........................................................................ [  5%]
+........................................................................ [  7%]
+........................................................................ [  9%]
+........................................................................ [ 11%]
+........................................................................ [ 13%]
+........................................................................ [ 15%]
+........................................................................ [ 17%]
+........................................................................ [ 18%]
+........................................................................ [ 20%]
+........................................................................ [ 22%]
+........................................................................ [ 24%]
+........................................................................ [ 26%]
+........................................................................ [ 28%]
+........................................................................ [ 30%]
+........................................................................ [ 32%]
+........................................................................ [ 34%]
+........................................................................ [ 36%]
+........................................................................ [ 37%]
+........................................................................ [ 39%]
+........................................................................ [ 41%]
+........................................................................ [ 43%]
+........................................................................ [ 45%]
+........................................................................ [ 47%]
+........................................................................ [ 49%]
+........................................................................ [ 51%]
+........................................................................ [ 53%]
+........................................................................ [ 55%]
+........................................................................ [ 56%]
+........................................................................ [ 58%]
+........................................................................ [ 60%]
+........................................................................ [ 62%]
+........................................................................ [ 64%]
+........................................................................ [ 66%]
+........................................................................ [ 68%]
+........................................................................ [ 70%]
+........................................................................ [ 72%]
+........................................................................ [ 73%]
+........................................................................ [ 75%]
+........................................................................ [ 77%]
+........................................................................ [ 79%]
+........................................................................ [ 81%]
+........................................................................ [ 83%]
+........................................................................ [ 85%]
+........................................................................ [ 87%]
+........................................................................ [ 89%]
+........................................................................ [ 91%]
+........................................................................ [ 92%]
+........................................................................ [ 94%]
+........................................................................ [ 96%]
+........................................................................ [ 98%]
+....................................................                     [100%]
+=============================== warnings summary ===============================
+.venv/lib/python3.13/site-packages/fastapi/testclient.py:1
+  /var/home/ducoterra/Projects/Personal/brain_of_reese/.venv/lib/python3.13/site-packages/fastapi/testclient.py:1: StarletteDeprecationWarning: Using `httpx` with `starlette.testclient` is deprecated; install `httpx2` instead.
+    from starlette.testclient import TestClient as TestClient  # noqa
+
+-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
+================================ tests coverage ================================
+_______________ coverage: platform linux, python 3.13.13-final-0 _______________
+
+Name                           Stmts   Miss  Cover
+--------------------------------------------------
+app/__init__.py                    1      0   100%
+app/api/__init__.py                0      0   100%
+app/api/auth.py                   52      0   100%
+app/api/chat.py                  277      1    99%
+app/api/chat_images.py            50      0   100%
+app/api/chats.py                 110      0   100%
+app/api/config.py                 13      0   100%
+app/api/doc_drafts.py             99      0   100%
+app/api/docs.py                  196      1    99%
+app/api/git_sources.py           259      0   100%
+app/api/health.py                 10      0   100%
+app/api/steering.py               42      0   100%
+app/api/suggestions.py            33      0   100%
+app/api/sync.py                  204      0   100%
+app/api/tokens.py                 40      0   100%
+app/api/ui_settings.py            55      0   100%
+app/config.py                    266      0   100%
+app/core/__init__.py               0      0   100%
+app/core/auth.py                  45      0   100%
+app/core/caching.py              124      0   100%
+app/core/debugging.py             29      2    93%
+app/core/docs_push.py             39      0   100%
+app/core/errors.py                 5      0   100%
+app/core/logging.py               13      0   100%
+app/core/rate_limit.py            44      0   100%
+app/core/security_headers.py      20      0   100%
+app/core/theming.py               38      0   100%
+app/core/tokens.py                44      0   100%
+app/db.py                         22      0   100%
+app/main.py                       68      0   100%
+app/models.py                    136      0   100%
+app/rag/__init__.py                0      0   100%
+app/rag/agent.py                 454      1    99%
+app/rag/archive_upload.py        134      0   100%
+app/rag/chunker.py               220      4    98%
+app/rag/doc_dates.py              18      0   100%
+app/rag/folder_summaries.py      134      0   100%
+app/rag/git_sources.py            77      0   100%
+app/rag/importer.py              841      4    99%
+app/rag/llm.py                   246      1    99%
+app/rag/md_images.py              63      0   100%
+app/rag/overview.py               78      0   100%
+app/rag/pdf.py                   192      0   100%
+app/rag/prompts.py               156      0   100%
+app/rag/retriever.py             252      3    99%
+app/rag/scaffolding.py            55      0   100%
+app/rag/source_removal.py         41      0   100%
+app/rag/sources_meta.py           16      0   100%
+app/rag/suggestions.py            27      0   100%
+app/rag/summarizer.py             58      0   100%
+app/schemas.py                   473      0   100%
+--------------------------------------------------
+TOTAL                           5869     17    99%
+coverage gate: app/ 99% (>90%) OK
+All checks passed!
+0 errors, 0 warnings, 0 informations
+WARNING: there is a new pyright version available (v1.1.411 -> v1.1.414).
+Please install the new version or set PYRIGHT_PYTHON_FORCE_VERSION to `latest`
+
+validation OK

.agents/reports/163_pdf_strip_additive_selfheal/163_pdf_strip_additive_selfheal__01_schema_wire_strip_mode.a1.md

@@ -0,0 +1,21 @@
+All green. Final state:
+
+**Implemented (task 01 — phase 163 D1, LOCKED A2/A3)**
+- `app/models.py`: `Document.page_strip_mode` (VARCHAR(16), nullable, no server default, unindexed) right after `page_count`, docstringed as the strip CLASS (`primary` = fallback class, `additive` = pictorial class, NULL = no strip).
+- `alembic/versions/0027_documents_page_strip_mode.py`: additive add/drop migration mirroring 0026; `upgrade head → downgrade -1 → upgrade head` round-trips clean (DB now at `0027 (head)`).
+- `app/schemas.py`: `DocContent.page_strip_mode` + the `image_url`-style omission in the existing `model_serializer` (ABSENT, never `null`).
+- `app/api/docs.py`: content payload passes `page_strip_mode=doc.page_strip_mode if doc.page_count is not None else None` — the whole strip block keys off `page_count`, so a stray mode never travels alone; `/pages/{n}` untouched; endpoint docstring + README "Which kind of strip it is (phase 163)" section updated. No AGENTS.md change (its viewer bullets are the final pass's).
+- Tests: `tests/integration/test_migration_0027.py` (new, house pattern: contract + pre-phase `page_count=36`/mode-NULL backfill + ORM round-trip both classes + tested downgrade + round-trip), `tests/unit/test_models.py` (column contract + ORM defaults), `tests/unit/test_document_viewer.py` (model omission rule; endpoint: stamped primary/additive, pre-phase mode-NULL row → key absent with `page_urls` kept, stray mode + `page_count` NULL → neither key, exact non-qualifying key-set), `tests/unit/test_pdf_page_copies.py` (importer leaves the class unstamped until task 02), `tests/integration/test_docs_api.py` (`_seed_page_doc` gains `page_strip_mode`; real-DB class pins + never-alone + exact key set).
+
+**Gates**
+- `uv run pytest` → **3772 passed**; `uv run pytest --cov=app --cov-report=term-missing` → **3772 passed, app TOTAL 99%** (`app/schemas.py` 100%).
+- `uv run ruff check .` → All checks passed; `uv run pyright` → 0 errors.
+- `uv run alembic upgrade head && downgrade -1 && upgrade head` → clean, `current` = `0027 (head)`.
+- E2E regression (isolation, `--no-cov`): `test_pdf_pages_in_sources.py` 4 passed, `test_pictorial_pdf.py`+`test_garbage_pdf.py` 6 passed, `test_smoke.py` 3 passed.
+
+**Notes / decisions**
+- Non-qualifying payloads stay byte-identical (both strip keys dropped when `page_count` is NULL); pre-phase rows serialize the class ABSENT and the viewer will read that as `primary` (task 04) — pinned, no data backfill.
+- Column carries no CHECK constraint (purely additive per A13); the class values are enforced by the writers (tasks 02/03).
+- No live-infrastructure changes needed for this task.
+
+**Next pending task:** `163_pdf_strip_additive_selfheal/02_persist_write_first.md`.

.agents/reports/163_pdf_strip_additive_selfheal/163_pdf_strip_additive_selfheal__01_schema_wire_strip_mode.a1.md.status

@@ -0,0 +1 @@
+ok

.agents/reports/163_pdf_strip_additive_selfheal/163_pdf_strip_additive_selfheal__01_schema_wire_strip_mode.a1.validate

@@ -0,0 +1,124 @@
+........................................................................ [  1%]
+........................................................................ [  3%]
+........................................................................ [  5%]
+........................................................................ [  7%]
+........................................................................ [  9%]
+........................................................................ [ 11%]
+........................................................................ [ 13%]
+........................................................................ [ 15%]
+........................................................................ [ 17%]
+........................................................................ [ 19%]
+........................................................................ [ 20%]
+........................................................................ [ 22%]
+........................................................................ [ 24%]
+........................................................................ [ 26%]
+........................................................................ [ 28%]
+........................................................................ [ 30%]
+........................................................................ [ 32%]
+........................................................................ [ 34%]
+........................................................................ [ 36%]
+........................................................................ [ 38%]
+........................................................................ [ 40%]
+........................................................................ [ 41%]
+........................................................................ [ 43%]
+........................................................................ [ 45%]
+........................................................................ [ 47%]
+........................................................................ [ 49%]
+........................................................................ [ 51%]
+........................................................................ [ 53%]
+........................................................................ [ 55%]
+........................................................................ [ 57%]
+........................................................................ [ 59%]
+........................................................................ [ 61%]
+........................................................................ [ 62%]
+........................................................................ [ 64%]
+........................................................................ [ 66%]
+........................................................................ [ 68%]
+........................................................................ [ 70%]
+........................................................................ [ 72%]
+........................................................................ [ 74%]
+........................................................................ [ 76%]
+........................................................................ [ 78%]
+........................................................................ [ 80%]
+........................................................................ [ 82%]
+........................................................................ [ 83%]
+........................................................................ [ 85%]
+........................................................................ [ 87%]
+........................................................................ [ 89%]
+........................................................................ [ 91%]
+........................................................................ [ 93%]
+........................................................................ [ 95%]
+........................................................................ [ 97%]
+........................................................................ [ 99%]
+............................                                             [100%]
+=============================== warnings summary ===============================
+.venv/lib/python3.13/site-packages/fastapi/testclient.py:1
+  /var/home/ducoterra/Projects/Personal/brain_of_reese/.venv/lib/python3.13/site-packages/fastapi/testclient.py:1: StarletteDeprecationWarning: Using `httpx` with `starlette.testclient` is deprecated; install `httpx2` instead.
+    from starlette.testclient import TestClient as TestClient  # noqa
+
+-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
+================================ tests coverage ================================
+_______________ coverage: platform linux, python 3.13.13-final-0 _______________
+
+Name                           Stmts   Miss  Cover
+--------------------------------------------------
+app/__init__.py                    1      0   100%
+app/api/__init__.py                0      0   100%
+app/api/auth.py                   52      0   100%
+app/api/chat.py                  277      1    99%
+app/api/chat_images.py            50      0   100%
+app/api/chats.py                 110      0   100%
+app/api/config.py                 13      0   100%
+app/api/doc_drafts.py             99      0   100%
+app/api/docs.py                  196      1    99%
+app/api/git_sources.py           259      0   100%
+app/api/health.py                 10      0   100%
+app/api/steering.py               42      0   100%
+app/api/suggestions.py            33      0   100%
+app/api/sync.py                  204      0   100%
+app/api/tokens.py                 40      0   100%
+app/api/ui_settings.py            55      0   100%
+app/config.py                    266      0   100%
+app/core/__init__.py               0      0   100%
+app/core/auth.py                  45      0   100%
+app/core/caching.py              124      0   100%
+app/core/debugging.py             29      2    93%
+app/core/docs_push.py             39      0   100%
+app/core/errors.py                 5      0   100%
+app/core/logging.py               13      0   100%
+app/core/rate_limit.py            44      0   100%
+app/core/security_headers.py      20      0   100%
+app/core/theming.py               38      0   100%
+app/core/tokens.py                44      0   100%
+app/db.py                         22      0   100%
+app/main.py                       68      0   100%
+app/models.py                    136      0   100%
+app/rag/__init__.py                0      0   100%
+app/rag/agent.py                 454      1    99%
+app/rag/archive_upload.py        134      0   100%
+app/rag/chunker.py               220      4    98%
+app/rag/doc_dates.py              18      0   100%
+app/rag/folder_summaries.py      134      0   100%
+app/rag/git_sources.py            77      0   100%
+app/rag/importer.py              764      4    99%
+app/rag/llm.py                   246      1    99%
+app/rag/md_images.py              63      0   100%
+app/rag/overview.py               78      0   100%
+app/rag/pdf.py                   192      0   100%
+app/rag/prompts.py               156      0   100%
+app/rag/retriever.py             252      3    99%
+app/rag/scaffolding.py            55      0   100%
+app/rag/source_removal.py         41      0   100%
+app/rag/sources_meta.py           16      0   100%
+app/rag/suggestions.py            27      0   100%
+app/rag/summarizer.py             58      0   100%
+app/schemas.py                   473      0   100%
+--------------------------------------------------
+TOTAL                           5792     17    99%
+coverage gate: app/ 99% (>90%) OK
+All checks passed!
+0 errors, 0 warnings, 0 informations
+WARNING: there is a new pyright version available (v1.1.411 -> v1.1.414).
+Please install the new version or set PYRIGHT_PYTHON_FORCE_VERSION to `latest`
+
+validation OK

.agents/reports/163_pdf_strip_additive_selfheal/163_pdf_strip_additive_selfheal__02_persist_write_first.a1.md

@@ -0,0 +1,13 @@
+**Task 02 complete — `_persist_pdf_pages` write-first atomic persist + `strip_class` stamp (D2, A4/A7).**
+
+- `app/rag/importer.py`: `_persist_pdf_pages` rewritten — render → per-file atomic write (`<name>.png.tmp` + `os.replace`, never a direct final-path write) → stale-tail delete LAST (only old-extent files not rewritten; a file about to be overwritten is never deleted); failure path unlinks this call's `.tmp`s and files it **created** (an overwritten copy keeps standing — deleting it would re-open the live-loss hole); `strip_class="fallback"|"pictorial"|None` replaces `eligible`, stamping `page_strip_mode` `primary`/`additive` exactly with `page_count`; the None arm and no-pages guard delete-then-null (helper reads the count); module + function docstrings carry the write-first rationale and the verbatim phase-148 truth table.
+- Call site (`_index_pdf_file`): `strip_class = "fallback" if fallback_fired else ("pictorial" if pictorial_described else None)`; fail-soft handler nulls `page_strip_mode` too; comments updated.
+- `tests/unit/test_pdf_page_copies.py`: re-cut task-01 pin → additive stamp; fallback→`primary`, ineligible/qualifier-loss/no-pages→NULL pins; new pins — order (event log: all `os.replace`s strictly before the first stale `unlink`, new bytes win), mid-write failure (prior 10-file set intact, zero `.tmp` orphans, count/mode NULL, doc indexed), atomic tmp+replace writer (every write to `.png.tmp`, `dst`/`src` pairing), best-effort stale-tail OSError branch; writer-exception test re-cut for tmp semantics.
+
+Gates: `uv run pytest --cov=app --cov-report=term-missing` → **3776 passed, app coverage 99%** (importer 99%; remaining misses pre-existing); `uv run ruff check . && uv run pyright` → clean; E2E `test_pdf_pages_in_sources` (4), `test_pictorial_pdf` (3), `test_garbage_pdf` (3) all pass in isolation.
+
+Docs: no further doc impact — README's phase-163 strip-class section (task 01) already covers the user-visible contract; deletion ordering is internal.
+
+Notable decision: the partial-write cleanup keeps an **overwritten** final (keeps only newly-created files) — strictly satisfies "a failed call never leaves more files than before, and never removes a live copy" (LOCKED A4).
+
+Next pending task: `163_pdf_strip_additive_selfheal/03_fast_path_regate.md`.

.agents/reports/163_pdf_strip_additive_selfheal/163_pdf_strip_additive_selfheal__02_persist_write_first.a1.md.status

@@ -0,0 +1 @@
+ok

.agents/reports/163_pdf_strip_additive_selfheal/163_pdf_strip_additive_selfheal__02_persist_write_first.a1.validate

@@ -0,0 +1,124 @@
+........................................................................ [  1%]
+........................................................................ [  3%]
+........................................................................ [  5%]
+........................................................................ [  7%]
+........................................................................ [  9%]
+........................................................................ [ 11%]
+........................................................................ [ 13%]
+........................................................................ [ 15%]
+........................................................................ [ 17%]
+........................................................................ [ 19%]
+........................................................................ [ 20%]
+........................................................................ [ 22%]
+........................................................................ [ 24%]
+........................................................................ [ 26%]
+........................................................................ [ 28%]
+........................................................................ [ 30%]
+........................................................................ [ 32%]
+........................................................................ [ 34%]
+........................................................................ [ 36%]
+........................................................................ [ 38%]
+........................................................................ [ 40%]
+........................................................................ [ 41%]
+........................................................................ [ 43%]
+........................................................................ [ 45%]
+........................................................................ [ 47%]
+........................................................................ [ 49%]
+........................................................................ [ 51%]
+........................................................................ [ 53%]
+........................................................................ [ 55%]
+........................................................................ [ 57%]
+........................................................................ [ 59%]
+........................................................................ [ 61%]
+........................................................................ [ 62%]
+........................................................................ [ 64%]
+........................................................................ [ 66%]
+........................................................................ [ 68%]
+........................................................................ [ 70%]
+........................................................................ [ 72%]
+........................................................................ [ 74%]
+........................................................................ [ 76%]
+........................................................................ [ 78%]
+........................................................................ [ 80%]
+........................................................................ [ 81%]
+........................................................................ [ 83%]
+........................................................................ [ 85%]
+........................................................................ [ 87%]
+........................................................................ [ 89%]
+........................................................................ [ 91%]
+........................................................................ [ 93%]
+........................................................................ [ 95%]
+........................................................................ [ 97%]
+........................................................................ [ 99%]
+................................                                         [100%]
+=============================== warnings summary ===============================
+.venv/lib/python3.13/site-packages/fastapi/testclient.py:1
+  /var/home/ducoterra/Projects/Personal/brain_of_reese/.venv/lib/python3.13/site-packages/fastapi/testclient.py:1: StarletteDeprecationWarning: Using `httpx` with `starlette.testclient` is deprecated; install `httpx2` instead.
+    from starlette.testclient import TestClient as TestClient  # noqa
+
+-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
+================================ tests coverage ================================
+_______________ coverage: platform linux, python 3.13.13-final-0 _______________
+
+Name                           Stmts   Miss  Cover
+--------------------------------------------------
+app/__init__.py                    1      0   100%
+app/api/__init__.py                0      0   100%
+app/api/auth.py                   52      0   100%
+app/api/chat.py                  277      1    99%
+app/api/chat_images.py            50      0   100%
+app/api/chats.py                 110      0   100%
+app/api/config.py                 13      0   100%
+app/api/doc_drafts.py             99      0   100%
+app/api/docs.py                  196      1    99%
+app/api/git_sources.py           259      0   100%
+app/api/health.py                 10      0   100%
+app/api/steering.py               42      0   100%
+app/api/suggestions.py            33      0   100%
+app/api/sync.py                  204      0   100%
+app/api/tokens.py                 40      0   100%
+app/api/ui_settings.py            55      0   100%
+app/config.py                    266      0   100%
+app/core/__init__.py               0      0   100%
+app/core/auth.py                  45      0   100%
+app/core/caching.py              124      0   100%
+app/core/debugging.py             29      2    93%
+app/core/docs_push.py             39      0   100%
+app/core/errors.py                 5      0   100%
+app/core/logging.py               13      0   100%
+app/core/rate_limit.py            44      0   100%
+app/core/security_headers.py      20      0   100%
+app/core/theming.py               38      0   100%
+app/core/tokens.py                44      0   100%
+app/db.py                         22      0   100%
+app/main.py                       68      0   100%
+app/models.py                    136      0   100%
+app/rag/__init__.py                0      0   100%
+app/rag/agent.py                 454      1    99%
+app/rag/archive_upload.py        134      0   100%
+app/rag/chunker.py               220      4    98%
+app/rag/doc_dates.py              18      0   100%
+app/rag/folder_summaries.py      134      0   100%
+app/rag/git_sources.py            77      0   100%
+app/rag/importer.py              788      4    99%
+app/rag/llm.py                   246      1    99%
+app/rag/md_images.py              63      0   100%
+app/rag/overview.py               78      0   100%
+app/rag/pdf.py                   192      0   100%
+app/rag/prompts.py               156      0   100%
+app/rag/retriever.py             252      3    99%
+app/rag/scaffolding.py            55      0   100%
+app/rag/source_removal.py         41      0   100%
+app/rag/sources_meta.py           16      0   100%
+app/rag/suggestions.py            27      0   100%
+app/rag/summarizer.py             58      0   100%
+app/schemas.py                   473      0   100%
+--------------------------------------------------
+TOTAL                           5816     17    99%
+coverage gate: app/ 99% (>90%) OK
+All checks passed!
+0 errors, 0 warnings, 0 informations
+WARNING: there is a new pyright version available (v1.1.411 -> v1.1.414).
+Please install the new version or set PYRIGHT_PYTHON_FORCE_VERSION to `latest`
+
+validation OK

.agents/reports/163_pdf_strip_additive_selfheal/163_pdf_strip_additive_selfheal__03_fast_path_regate.a1.md

@@ -0,0 +1,10 @@
+**Task 03 complete — all green.**
+
+- **`app/rag/importer.py`**: new `_regate_pdf_strip(session, doc, *, full_path, llm, summary)` — mirrors the changed-path gate verbatim from current bytes + settings (phase-133 `is_usable_pdf_text` → phase-140 `pictorial_pages` + `pdf_pictorial` kill switch), **zero LLM/vision calls**; heals any missing `<id>-p{N}.png` (full re-render), stamps `page_strip_mode` (silent NULL→stamp, `mode old->new` on a set mode), clears a strip that no longer qualifies + deletes its copies, one greppable INFO only when it **acts**, whole body fail-soft (`re-gate failed` / `re-render produced no pages (state kept)` / `page_count drift a->b` warnings).
+- `_unchanged_doc` gained `full_path: Path | None = None`; the **guarded** re-gate (`page_count` set + `.pdf`) runs at **both** exits (incl. the D1 manual-date return) and commits what it wrote; all 4 call sites (fast path, text/PDF/image content-digest hits) pass their `full_path`. Docstrings: the phase-147 "the copies persist" assumption is replaced (module doc + `_unchanged_doc` step 7 + `_index_pdf_file`).
+- De-duplicated task 02's writer: extracted `_write_page_copies` (tmp + `os.replace`, orphan cleanup) and `_delete_stale_page_tail`; `_persist_pdf_pages` and the heal now share them (behavior unchanged).
+- **Tests**: 11 new pins in `tests/unit/test_importer_unchanged_fast_path.py` — heal (pictorial, live NULL-mode), fallback heal (`primary`), healthy no-op (zero writes, zero re-gate log), mode re-stamp, repair arm, kill-switch clear, render-raise fail-soft, empty-render keeps row, count drift, content-digest-hit path, non-strip zero-parse guard (spy).
+- `uv run pytest tests/unit/test_importer_unchanged_fast_path.py tests/unit/test_pdf_page_copies.py --no-cov` → exit 0 (24 + 15 tests); `uv run pytest --cov=app` → **3787 passed**, `app/rag/importer.py` **99%** (only pre-existing misses: 883, 3506-3507, 3793), TOTAL 99%; `uv run ruff check .` → clean; `uv run pyright` → **0 errors**; PDF E2E files pass in isolation (`test_pdf_pages_in_sources`, `test_pictorial_pdf`, `test_garbage_pdf`, `test_sync_changed_only`, exit 0).
+- **Docs**: README — new "The strip repairs itself (phase 163)" viewer bullet, the "Updating the documents" fast-path qualifier, and the `BOR_PDF_PAGE_DIR` env-row note (dedup avoided). **AGENTS.md: no change** (per the phase overview its viewer bullet lands in the phase's final pass).
+- **Notable**: (1) running all of `tests/e2e` in one process fails on a **pre-existing** design artifact — `tests/e2e/test_extensionless_import.py:70` mutates `os.environ["BOR_IMPORT_EXTENSIONS"]` at import, so other suites' 13-doc seed drops to 5; unaffected by this task, and the project's gate is per-file isolation (flagging for the owner/final pass). (2) Restored `.agents/screenshots/doc_modal_scrollbar_1280.png` (a tracked artifact my E2E runs overwrote) from HEAD so the phase commit stays clean.
+- **Next pending task**: `163_pdf_strip_additive_selfheal/04_viewer_additive_toggle.md`.

.agents/reports/163_pdf_strip_additive_selfheal/163_pdf_strip_additive_selfheal__03_fast_path_regate.a1.md.status

@@ -0,0 +1 @@
+ok

.agents/reports/163_pdf_strip_additive_selfheal/163_pdf_strip_additive_selfheal__03_fast_path_regate.a1.validate

@@ -0,0 +1,124 @@
+........................................................................ [  1%]
+........................................................................ [  3%]
+........................................................................ [  5%]
+........................................................................ [  7%]
+........................................................................ [  9%]
+........................................................................ [ 11%]
+........................................................................ [ 13%]
+........................................................................ [ 15%]
+........................................................................ [ 17%]
+........................................................................ [ 19%]
+........................................................................ [ 20%]
+........................................................................ [ 22%]
+........................................................................ [ 24%]
+........................................................................ [ 26%]
+........................................................................ [ 28%]
+........................................................................ [ 30%]
+........................................................................ [ 32%]
+........................................................................ [ 34%]
+........................................................................ [ 36%]
+........................................................................ [ 38%]
+........................................................................ [ 39%]
+........................................................................ [ 41%]
+........................................................................ [ 43%]
+........................................................................ [ 45%]
+........................................................................ [ 47%]
+........................................................................ [ 49%]
+........................................................................ [ 51%]
+........................................................................ [ 53%]
+........................................................................ [ 55%]
+........................................................................ [ 57%]
+........................................................................ [ 58%]
+........................................................................ [ 60%]
+........................................................................ [ 62%]
+........................................................................ [ 64%]
+........................................................................ [ 66%]
+........................................................................ [ 68%]
+........................................................................ [ 70%]
+........................................................................ [ 72%]
+........................................................................ [ 74%]
+........................................................................ [ 76%]
+........................................................................ [ 77%]
+........................................................................ [ 79%]
+........................................................................ [ 81%]
+........................................................................ [ 83%]
+........................................................................ [ 85%]
+........................................................................ [ 87%]
+........................................................................ [ 89%]
+........................................................................ [ 91%]
+........................................................................ [ 93%]
+........................................................................ [ 95%]
+........................................................................ [ 96%]
+........................................................................ [ 98%]
+...........................................                              [100%]
+=============================== warnings summary ===============================
+.venv/lib/python3.13/site-packages/fastapi/testclient.py:1
+  /var/home/ducoterra/Projects/Personal/brain_of_reese/.venv/lib/python3.13/site-packages/fastapi/testclient.py:1: StarletteDeprecationWarning: Using `httpx` with `starlette.testclient` is deprecated; install `httpx2` instead.
+    from starlette.testclient import TestClient as TestClient  # noqa
+
+-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
+================================ tests coverage ================================
+_______________ coverage: platform linux, python 3.13.13-final-0 _______________
+
+Name                           Stmts   Miss  Cover
+--------------------------------------------------
+app/__init__.py                    1      0   100%
+app/api/__init__.py                0      0   100%
+app/api/auth.py                   52      0   100%
+app/api/chat.py                  277      1    99%
+app/api/chat_images.py            50      0   100%
+app/api/chats.py                 110      0   100%
+app/api/config.py                 13      0   100%
+app/api/doc_drafts.py             99      0   100%
+app/api/docs.py                  196      1    99%
+app/api/git_sources.py           259      0   100%
+app/api/health.py                 10      0   100%
+app/api/steering.py               42      0   100%
+app/api/suggestions.py            33      0   100%
+app/api/sync.py                  204      0   100%
+app/api/tokens.py                 40      0   100%
+app/api/ui_settings.py            55      0   100%
+app/config.py                    266      0   100%
+app/core/__init__.py               0      0   100%
+app/core/auth.py                  45      0   100%
+app/core/caching.py              124      0   100%
+app/core/debugging.py             29      2    93%
+app/core/docs_push.py             39      0   100%
+app/core/errors.py                 5      0   100%
+app/core/logging.py               13      0   100%
+app/core/rate_limit.py            44      0   100%
+app/core/security_headers.py      20      0   100%
+app/core/theming.py               38      0   100%
+app/core/tokens.py                44      0   100%
+app/db.py                         22      0   100%
+app/main.py                       68      0   100%
+app/models.py                    136      0   100%
+app/rag/__init__.py                0      0   100%
+app/rag/agent.py                 454      1    99%
+app/rag/archive_upload.py        134      0   100%
+app/rag/chunker.py               220      4    98%
+app/rag/doc_dates.py              18      0   100%
+app/rag/folder_summaries.py      134      0   100%
+app/rag/git_sources.py            77      0   100%
+app/rag/importer.py              841      4    99%
+app/rag/llm.py                   246      1    99%
+app/rag/md_images.py              63      0   100%
+app/rag/overview.py               78      0   100%
+app/rag/pdf.py                   192      0   100%
+app/rag/prompts.py               156      0   100%
+app/rag/retriever.py             252      3    99%
+app/rag/scaffolding.py            55      0   100%
+app/rag/source_removal.py         41      0   100%
+app/rag/sources_meta.py           16      0   100%
+app/rag/suggestions.py            27      0   100%
+app/rag/summarizer.py             58      0   100%
+app/schemas.py                   473      0   100%
+--------------------------------------------------
+TOTAL                           5869     17    99%
+coverage gate: app/ 99% (>90%) OK
+All checks passed!
+0 errors, 0 warnings, 0 informations
+WARNING: there is a new pyright version available (v1.1.411 -> v1.1.414).
+Please install the new version or set PYRIGHT_PYTHON_FORCE_VERSION to `latest`
+
+validation OK

.agents/reports/163_pdf_strip_additive_selfheal/163_pdf_strip_additive_selfheal__04_viewer_additive_toggle.a1.md

@@ -0,0 +1,18 @@
+**Task 04 complete — report**
+
+**Implemented**
+- `frontend/assets/document.js`: `renderDocument` now branches on `doc.page_strip_mode` — absent/`"primary"` keeps the phase-148 strip-only branch verbatim (byte-identical); `"additive"` renders the text doc (Summary panel + content) inside a `.doc-views` wrapper with a `Text`/`Pages` toggle (`role=group`, `aria-pressed`, Text default per LOCKED A6); strip mounts once, hidden via CSS `data-view` swap (lazy imgs never fetch while hidden, no refetch on flip-flop); never-build-HTML rule kept (module still has exactly one `innerHTML`); module docstring gained the phase-163 paragraph and re-scoped the phase-148 "INSTEAD of" sentence to the primary class.
+- `frontend/assets/styles.css`: `.doc-view-toggle` ghost-pill pair (house `.auth-link`/`.share-chat-btn` idiom, chat-column cap, ≥44px targets, global 3px focus ring, pressed = brand fill + weight 700 — text + style, never color alone) + `.doc-views[data-view]` show/hide rules.
+- New `tests/unit/test_viewer_additive_toggle.py`: 9 pins (source-contract + node behavioral runs of the real core: additive shape, primary/absent byte-identity, text-doc untouched, flip semantics with strip-node persistence, lazy imgs, XSS, CSS contract).
+- Re-cut stale/newly-affected pins: `test_summary_edit_ui.py` (slot mount), `test_wide_column_css.py` (7th reading-column selector), E2E additive-view contracts in `test_pdf_pages_in_sources.py` and `test_pictorial_pdf.py` (text-first default, Pages flip reveals loaded strip, description rides the shown text).
+
+**Gates**
+- `uv run pytest` → **3796 passed**, 0 failed; coverage `app/` **99%** (>90%).
+- `uv run ruff check .` → clean; `uv run pyright` → 0 errors.
+- E2E in isolation: `test_pdf_pages_in_sources.py` 4✓, `test_pictorial_pdf.py` 3✓, `test_garbage_pdf.py` 3✓, plus `test_pdf_support.py` 6✓, `test_document_viewer.py` 8✓, `test_summary_in_viewer.py` 4✓.
+
+**Docs**: README — phase-148 viewer bullet scope-fixed to "primary class only" (the phase-163 bullets from tasks 01–03 already describe the toggle); AGENTS.md — no doc impact (viewer bullet note is reserved for the phase final pass).
+
+**Decisions**: additive E2E witnesses flipped from "descriptions not displayed" to "descriptions ride the shown text" (intended consequence of LOCKED A1/A6); toggle lives in a `.doc-views` wrapper only on additive docs — all other DOM untouched.
+
+**Next pending task**: `163_pdf_strip_additive_selfheal/05_e2e_selfheal_pins.md`.

.agents/reports/163_pdf_strip_additive_selfheal/163_pdf_strip_additive_selfheal__04_viewer_additive_toggle.a1.md.status

@@ -0,0 +1 @@
+ok

.agents/reports/163_pdf_strip_additive_selfheal/163_pdf_strip_additive_selfheal__04_viewer_additive_toggle.a1.validate

@@ -0,0 +1,124 @@
+........................................................................ [  1%]
+........................................................................ [  3%]
+........................................................................ [  5%]
+........................................................................ [  7%]
+........................................................................ [  9%]
+........................................................................ [ 11%]
+........................................................................ [ 13%]
+........................................................................ [ 15%]
+........................................................................ [ 17%]
+........................................................................ [ 18%]
+........................................................................ [ 20%]
+........................................................................ [ 22%]
+........................................................................ [ 24%]
+........................................................................ [ 26%]
+........................................................................ [ 28%]
+........................................................................ [ 30%]
+........................................................................ [ 32%]
+........................................................................ [ 34%]
+........................................................................ [ 36%]
+........................................................................ [ 37%]
+........................................................................ [ 39%]
+........................................................................ [ 41%]
+........................................................................ [ 43%]
+........................................................................ [ 45%]
+........................................................................ [ 47%]
+........................................................................ [ 49%]
+........................................................................ [ 51%]
+........................................................................ [ 53%]
+........................................................................ [ 55%]
+........................................................................ [ 56%]
+........................................................................ [ 58%]
+........................................................................ [ 60%]
+........................................................................ [ 62%]
+........................................................................ [ 64%]
+........................................................................ [ 66%]
+........................................................................ [ 68%]
+........................................................................ [ 70%]
+........................................................................ [ 72%]
+........................................................................ [ 73%]
+........................................................................ [ 75%]
+........................................................................ [ 77%]
+........................................................................ [ 79%]
+........................................................................ [ 81%]
+........................................................................ [ 83%]
+........................................................................ [ 85%]
+........................................................................ [ 87%]
+........................................................................ [ 89%]
+........................................................................ [ 91%]
+........................................................................ [ 92%]
+........................................................................ [ 94%]
+........................................................................ [ 96%]
+........................................................................ [ 98%]
+....................................................                     [100%]
+=============================== warnings summary ===============================
+.venv/lib/python3.13/site-packages/fastapi/testclient.py:1
+  /var/home/ducoterra/Projects/Personal/brain_of_reese/.venv/lib/python3.13/site-packages/fastapi/testclient.py:1: StarletteDeprecationWarning: Using `httpx` with `starlette.testclient` is deprecated; install `httpx2` instead.
+    from starlette.testclient import TestClient as TestClient  # noqa
+
+-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
+================================ tests coverage ================================
+_______________ coverage: platform linux, python 3.13.13-final-0 _______________
+
+Name                           Stmts   Miss  Cover
+--------------------------------------------------
+app/__init__.py                    1      0   100%
+app/api/__init__.py                0      0   100%
+app/api/auth.py                   52      0   100%
+app/api/chat.py                  277      1    99%
+app/api/chat_images.py            50      0   100%
+app/api/chats.py                 110      0   100%
+app/api/config.py                 13      0   100%
+app/api/doc_drafts.py             99      0   100%
+app/api/docs.py                  196      1    99%
+app/api/git_sources.py           259      0   100%
+app/api/health.py                 10      0   100%
+app/api/steering.py               42      0   100%
+app/api/suggestions.py            33      0   100%
+app/api/sync.py                  204      0   100%
+app/api/tokens.py                 40      0   100%
+app/api/ui_settings.py            55      0   100%
+app/config.py                    266      0   100%
+app/core/__init__.py               0      0   100%
+app/core/auth.py                  45      0   100%
+app/core/caching.py              124      0   100%
+app/core/debugging.py             29      2    93%
+app/core/docs_push.py             39      0   100%
+app/core/errors.py                 5      0   100%
+app/core/logging.py               13      0   100%
+app/core/rate_limit.py            44      0   100%
+app/core/security_headers.py      20      0   100%
+app/core/theming.py               38      0   100%
+app/core/tokens.py                44      0   100%
+app/db.py                         22      0   100%
+app/main.py                       68      0   100%
+app/models.py                    136      0   100%
+app/rag/__init__.py                0      0   100%
+app/rag/agent.py                 454      1    99%
+app/rag/archive_upload.py        134      0   100%
+app/rag/chunker.py               220      4    98%
+app/rag/doc_dates.py              18      0   100%
+app/rag/folder_summaries.py      134      0   100%
+app/rag/git_sources.py            77      0   100%
+app/rag/importer.py              841      4    99%
+app/rag/llm.py                   246      1    99%
+app/rag/md_images.py              63      0   100%
+app/rag/overview.py               78      0   100%
+app/rag/pdf.py                   192      0   100%
+app/rag/prompts.py               156      0   100%
+app/rag/retriever.py             252      3    99%
+app/rag/scaffolding.py            55      0   100%
+app/rag/source_removal.py         41      0   100%
+app/rag/sources_meta.py           16      0   100%
+app/rag/suggestions.py            27      0   100%
+app/rag/summarizer.py             58      0   100%
+app/schemas.py                   473      0   100%
+--------------------------------------------------
+TOTAL                           5869     17    99%
+coverage gate: app/ 99% (>90%) OK
+All checks passed!
+0 errors, 0 warnings, 0 informations
+WARNING: there is a new pyright version available (v1.1.411 -> v1.1.414).
+Please install the new version or set PYRIGHT_PYTHON_FORCE_VERSION to `latest`
+
+validation OK

.agents/reports/163_pdf_strip_additive_selfheal/163_pdf_strip_additive_selfheal__05_e2e_selfheal_pins.a1.md

@@ -0,0 +1,16 @@
+**Task 05 complete — report**
+
+**Implemented**
+- New `tests/e2e/test_pdf_strip_additive_selfheal.py` (5 tests, phase-163 A16 suite): fixture `strip_fixture` = `manual.pdf` (3 dense pages + 1 sparse map page → additive), `scanned.pdf` (XML-garbage layer → primary), `plain.pdf` (no strip); asserts 1 additive view (payload stamp, Text default, Pages flip loads all 4 imgs, flip-back keeps the mounted strip node — dataset-probed, zero refetch); 2 primary/plain DOM byte-identity incl. the key-omission payload pins; 3 heal (wipe copies → route 404 → one unchanged-byte sync → files restored, route 200, viewer complete); 4 repair (direct-DB stale `page_count`/mode + fake copies → cleared + deleted); 5 zero-LLM (mock `/v1/e2e/captured`: index burned both vision passes, heal/repair captured zero content calls — the sync prelude's fixed `ping` health probe is the one documented exemption).
+- Docs pass: README verified in full (viewer + importer + env-table claims cross-checked against code; defaults script-audited vs `Settings` — no drift); AGENTS.md rule 5 gained the phase-163 viewer-class note.
+
+**Gates** (all run this session, DB up)
+- `uv run pytest tests/e2e/test_pdf_strip_additive_selfheal.py -v --no-cov` → **5 passed** in isolation.
+- `test_pdf_pages_in_sources / test_pictorial_pdf / test_garbage_pdf / test_document_viewer / test_document_viewer_context / test_image_documents` → all exit 0, unmodified.
+- `uv run pytest --cov=app --cov-report=term-missing` → **3796 passed**, coverage **99%** (>90%); `uv run ruff check .` clean; `uv run pyright` 0 errors; alembic **0027 downgrade→upgrade round-trip clean**.
+
+**Docs**: README — no further edits needed (phase bullets from tasks 01–03 verified accurate); AGENTS.md — viewer-class note added.
+
+**Decisions**: heal/repair zero-call pins exempt only `check_models`' `ping` (every sync pays it, predates the phase); anything else fails the pin. Pre-existing `.agents/screenshots/doc_modal_scrollbar_1280.png` modification (earlier E2E run) left for the pipeline commit.
+
+**Next pending task**: `164_chat_error_honesty/01_caps_long_answer_chain.md`.

.agents/reports/163_pdf_strip_additive_selfheal/163_pdf_strip_additive_selfheal__05_e2e_selfheal_pins.a1.md.status

@@ -0,0 +1 @@
+ok

.agents/reports/163_pdf_strip_additive_selfheal/163_pdf_strip_additive_selfheal__05_e2e_selfheal_pins.a1.validate

@@ -0,0 +1,124 @@
+........................................................................ [  1%]
+........................................................................ [  3%]
+........................................................................ [  5%]
+........................................................................ [  7%]
+........................................................................ [  9%]
+........................................................................ [ 11%]
+........................................................................ [ 13%]
+........................................................................ [ 15%]
+........................................................................ [ 17%]
+........................................................................ [ 18%]
+........................................................................ [ 20%]
+........................................................................ [ 22%]
+........................................................................ [ 24%]
+........................................................................ [ 26%]
+........................................................................ [ 28%]
+........................................................................ [ 30%]
+........................................................................ [ 32%]
+........................................................................ [ 34%]
+........................................................................ [ 36%]
+........................................................................ [ 37%]
+........................................................................ [ 39%]
+........................................................................ [ 41%]
+........................................................................ [ 43%]
+........................................................................ [ 45%]
+........................................................................ [ 47%]
+........................................................................ [ 49%]
+........................................................................ [ 51%]
+........................................................................ [ 53%]
+........................................................................ [ 55%]
+........................................................................ [ 56%]
+........................................................................ [ 58%]
+........................................................................ [ 60%]
+........................................................................ [ 62%]
+........................................................................ [ 64%]
+........................................................................ [ 66%]
+........................................................................ [ 68%]
+........................................................................ [ 70%]
+........................................................................ [ 72%]
+........................................................................ [ 73%]
+........................................................................ [ 75%]
+........................................................................ [ 77%]
+........................................................................ [ 79%]
+........................................................................ [ 81%]
+........................................................................ [ 83%]
+........................................................................ [ 85%]
+........................................................................ [ 87%]
+........................................................................ [ 89%]
+........................................................................ [ 91%]
+........................................................................ [ 92%]
+........................................................................ [ 94%]
+........................................................................ [ 96%]
+........................................................................ [ 98%]
+....................................................                     [100%]
+=============================== warnings summary ===============================
+.venv/lib/python3.13/site-packages/fastapi/testclient.py:1
+  /var/home/ducoterra/Projects/Personal/brain_of_reese/.venv/lib/python3.13/site-packages/fastapi/testclient.py:1: StarletteDeprecationWarning: Using `httpx` with `starlette.testclient` is deprecated; install `httpx2` instead.
+    from starlette.testclient import TestClient as TestClient  # noqa
+
+-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
+================================ tests coverage ================================
+_______________ coverage: platform linux, python 3.13.13-final-0 _______________
+
+Name                           Stmts   Miss  Cover
+--------------------------------------------------
+app/__init__.py                    1      0   100%
+app/api/__init__.py                0      0   100%
+app/api/auth.py                   52      0   100%
+app/api/chat.py                  277      1    99%
+app/api/chat_images.py            50      0   100%
+app/api/chats.py                 110      0   100%
+app/api/config.py                 13      0   100%
+app/api/doc_drafts.py             99      0   100%
+app/api/docs.py                  196      1    99%
+app/api/git_sources.py           259      0   100%
+app/api/health.py                 10      0   100%
+app/api/steering.py               42      0   100%
+app/api/suggestions.py            33      0   100%
+app/api/sync.py                  204      0   100%
+app/api/tokens.py                 40      0   100%
+app/api/ui_settings.py            55      0   100%
+app/config.py                    266      0   100%
+app/core/__init__.py               0      0   100%
+app/core/auth.py                  45      0   100%
+app/core/caching.py              124      0   100%
+app/core/debugging.py             29      2    93%
+app/core/docs_push.py             39      0   100%
+app/core/errors.py                 5      0   100%
+app/core/logging.py               13      0   100%
+app/core/rate_limit.py            44      0   100%
+app/core/security_headers.py      20      0   100%
+app/core/theming.py               38      0   100%
+app/core/tokens.py                44      0   100%
+app/db.py                         22      0   100%
+app/main.py                       68      0   100%
+app/models.py                    136      0   100%
+app/rag/__init__.py                0      0   100%
+app/rag/agent.py                 454      1    99%
+app/rag/archive_upload.py        134      0   100%
+app/rag/chunker.py               220      4    98%
+app/rag/doc_dates.py              18      0   100%
+app/rag/folder_summaries.py      134      0   100%
+app/rag/git_sources.py            77      0   100%
+app/rag/importer.py              841      4    99%
+app/rag/llm.py                   246      1    99%
+app/rag/md_images.py              63      0   100%
+app/rag/overview.py               78      0   100%
+app/rag/pdf.py                   192      0   100%
+app/rag/prompts.py               156      0   100%
+app/rag/retriever.py             252      3    99%
+app/rag/scaffolding.py            55      0   100%
+app/rag/source_removal.py         41      0   100%
+app/rag/sources_meta.py           16      0   100%
+app/rag/suggestions.py            27      0   100%
+app/rag/summarizer.py             58      0   100%
+app/schemas.py                   473      0   100%
+--------------------------------------------------
+TOTAL                           5869     17    99%
+coverage gate: app/ 99% (>90%) OK
+All checks passed!
+0 errors, 0 warnings, 0 informations
+WARNING: there is a new pyright version available (v1.1.411 -> v1.1.414).
+Please install the new version or set PYRIGHT_PYTHON_FORCE_VERSION to `latest`
+
+validation OK

AGENTS.md

@@ -21,7 +21,11 @@
    panels — `max-height: min(70vh, 64rem)` — with a sticky header row;
    a clipped action label shows an ellipsis hint (`.cell-clip`) with
    the full text kept reachable via the control's aria-label
-   (phase 156)) and
+   (phase 156); the document viewer's page strip is CLASS-scoped —
+   the primary class shows the strip in place of the text, the
+   additive class never hides the text: Text/Pages toggle group
+   (`aria-pressed`, Text default) with the strip behind it
+   (phase 163)) and
    meets WCAG 2.1 AA basics (semantic
    landmarks, labels, contrast ≥4.5:1, focus-visible, aria-live for the
    stream).

README.md

@@ -239,9 +239,11 @@ idempotent and delta-based: each file's RAW BYTES are hashed (sha256,
 streamed) and checked against the last index's stamp *before* any
 extraction or vision work — an unchanged file costs one hash + one row
 lookup (no re-extract, no vision, no re-embed), while a changed file
-re-indexes fully. A refresh after a normal editing session takes
+re-indexes fully. The one exception is a PDF row that still claims a
-seconds — pictorial PDFs included (their expensive first-index vision
+page strip: it is re-gated from its current bytes (no vision, no LLM —
-burns only once per file version):
+see "The strip repairs itself" below). A refresh after a normal editing
+session takes seconds — pictorial PDFs included (their expensive
+first-index vision burns only once per file version):
 
 ```bash
 # After editing/adding/removing notes:
@@ -844,11 +846,42 @@ uploaded archives):
   shows the PAGES: a vertical strip of lazy page images with
   "Page N" captions, in place of the text content and the Summary
   panel (the per-page descriptions stay in the DB for Q&A — the
-  embeddings and summary are untouched, just not displayed). A lost
+  embeddings and summary are untouched, just not displayed). Since
+  phase 163 this strip-only shape is the PRIMARY class only — the
+  pictorial class keeps its text behind a Text/Pages toggle, per the
+  next bullet. A lost
   copy shows a "Page N unavailable" note for that page only. Text-
   heavy PDFs (usable layer, zero pictorial candidates — and the whole
   sparse class while `BOR_PDF_PICTORIAL=0`) keep the text view,
   byte-identical to pre-phase.
+- **Which kind of strip it is (phase 163).** The two qualifying classes
+  are not the same problem: the screenshot class has no usable text to
+  show, while a book with one sparse page (an illustration map page in a
+  36-page text) does. The document row therefore records the strip's
+  CLASS beside the page count (`documents.page_strip_mode` —
+  `primary` = the screenshots ARE the content, so the strip is the view;
+  `additive` = a usable text layer, so the text + Summary view stays and
+  the pages sit behind a Text/Pages toggle), and the document-content
+  payload carries it as `page_strip_mode`. The key is ABSENT — never
+  `null` — when there is no strip, which is also how a row written before
+  this phase serializes; the viewer treats an absent class as `primary`,
+  so every existing document renders exactly as before until the next
+  sync stamps it (the sync's self-heal pass, phase 163).
+- **The strip repairs itself (phase 163).** An unchanged PDF is not
+  assumed to still have its page images on disk. On every sync, a row
+  that claims a strip is re-gated deterministically from the current
+  file bytes and the current settings — the same gate the changed path
+  uses (`BOR_PDF_PICTORIAL` and `BOR_PDF_PICTORIAL_MIN_CHARS`
+  included), with **zero vision or LLM calls** (rendering is CPU-only):
+  the class is re-stamped, any `<doc-id>-p{N}.png` missing from
+  `BOR_PDF_PAGE_DIR` is re-rendered (write-before-delete, atomically per
+  file), and a strip that no longer qualifies — the kill switch turned
+  off, the sparse-page bar lowered past it — is cleared along with its
+  copies. So a wiped pages directory, or a copy set lost to an
+  interrupted sync, repairs itself on the next sync instead of leaving
+  "Page N unavailable" forever. It logs one `PDF strip re-gate …` line
+  only when it actually acts, and a failure there never fails the sync.
+  PDFs without a strip pay nothing extra (no extra read at all).
 
 ---
 
@@ -975,7 +1008,7 @@ to `.env` and adjust.
 | `BOR_PDF_PICTORIAL` | `1` | Master switch for the pictorial PDF pass (phase 140, on by default): a usable-text PDF's sparse pages (stripped text under `BOR_PDF_PICTORIAL_MIN_CHARS`) are rendered + vision-described into the indexed content (`[Page N illustration: …]`); `0` ⇒ content byte-identical to pre-phase-140 (no render, no vision call) |
 | `BOR_PDF_PICTORIAL_MIN_CHARS` | `100` | Sparseness bar for the pictorial pass (phase 140): a page whose STRIPPED text is STRICTLY shorter than this many chars is a candidate (a page with exactly this many is not); `0` = every page (A21); a negative value fails startup loudly |
 | `BOR_PDF_IMAGE_SKIP_BLANK` | `1` | Skip blank/decorative images embedded in PDFs (phase 151, on by default): near-white, nearly-uniform images (the table-shading "blank white" underlays) are dropped from the embedded-image pass — no row, no copy, no vision call; `0` ⇒ every extracted image is indexed again (pre-phase behavior; dedupe + soft-mask compositing unaffected) |
-| `BOR_PDF_PAGE_DIR` | `~/bor-sources/pages` | Persistent home for the rendered PDF page PNGs (phase 148): a page-qualifying PDF (the phase-133 screenshot class or the phase-140 pictorial class) gets every page rendered at index time as `<doc-id>-p{N}.png` here, so the document viewer shows the actual pages (the descriptions stay embedded for Q&A, just not displayed); text-heavy PDFs keep the text view |
+| `BOR_PDF_PAGE_DIR` | `~/bor-sources/pages` | Persistent home for the rendered PDF page PNGs (phase 148): a page-qualifying PDF (the phase-133 screenshot class or the phase-140 pictorial class) gets every page rendered at index time as `<doc-id>-p{N}.png` here, so the document viewer shows the actual pages (the descriptions stay embedded for Q&A, just not displayed); text-heavy PDFs keep the text view. Clearing the directory is self-healing — the next sync re-gates any strip that still claims pages and re-renders what is missing, with zero vision calls (phase 163) |
 | `BOR_GIT_SOURCES` | — (empty) | CSV of git repo URLs — **fallback while the admin Git sources page's list is empty** |
 | `BOR_SOURCES_DIR` | `~/bor-sources` | Where git repos are cloned/pulled |
 | `BOR_UPLOAD_DIR` | `~/bor-sources/uploads` | Where uploaded source archives are unpacked |

alembic/versions/0027_documents_page_strip_mode.py

@@ -0,0 +1,68 @@
+"""documents.page_strip_mode: the strip CLASS behind page_count (phase 163, task 01)
+
+Revision ID: 0027
+Revises: 0026
+Create Date: 2026-10-08
+
+Phase 163 (the PDF page strip is ADDITIVE for a text-bearing book —
+task 01, storage only):
+
+* ``documents.page_strip_mode`` — VARCHAR(16) NULLABLE, NO server
+  default, NO index: the CLASS of the page strip a non-NULL
+  ``page_count`` describes, so the viewer knows whether the strip
+  REPLACES the text or sits one toggle behind it.
+
+  ``"primary"`` — the phase-133 FALLBACK class (an unusable text layer:
+  the screenshots ARE the content), i.e. exactly the pre-phase view: the
+  strip is the document's view and no text is shown. ``"additive"`` —
+  the phase-140 PICTORIAL class (a usable text layer with ≥1 sparse page
+  described): the document keeps its text view and its Summary panel,
+  and the strip is one Text/Pages toggle away.
+
+  NULL means NO strip (every non-qualifying document — text documents,
+  images, text-heavy PDFs — and every qualifying PDF whose render/write
+  failed). A row written before this phase keeps NULL while its
+  ``page_count`` stands; the wire omits the key in that case and the
+  viewer treats an absent mode as ``"primary"`` (today's DOM), so
+  existing rows render exactly as before until the fast-path re-gate
+  (task 03) stamps them. Stamped and cleared EXACTLY with
+  ``page_count`` — the importer's persist (task 02) and the unchanged-
+  path re-gate (task 03) are the only writers.
+
+  A display fact, never a query surface, like the 0026 column it
+  accompanies: the content endpoint reads it from the ROW after the
+  ``(source, path)`` lookup, so it carries no index of its own (the
+  ``file_hash`` convention, 0025).
+
+One additive, fully reversible migration (A13); no other schema change.
+The ordering of the copies on disk (task 02) and the re-gate itself
+(task 03) are importer code — this revision only carries the column.
+"""
+from __future__ import annotations
+
+import sqlalchemy as sa
+
+from alembic import op
+
+revision = "0027"
+down_revision = "0026"
+branch_labels = None
+depends_on = None
+
+
+def upgrade() -> None:
+    # No index (a display fact, never a query surface — the content
+    # endpoint reads it from the row) and no server default (NULL is the
+    # meaningful state for every non-qualifying document, every fail-soft
+    # miss, and every pre-phase row).
+    op.add_column(
+        "documents", sa.Column("page_strip_mode", sa.String(length=16), nullable=True)
+    )
+
+
+def downgrade() -> None:
+    # The page_strip_mode column is the only 0027 artefact — dropping it
+    # leaves 0026's schema byte-identical (A13, fully reversible). The
+    # class is re-derivable from the source file on the next sync (the
+    # deterministic re-gate of task 03), so no data is lost by the drop.
+    op.drop_column("documents", "page_strip_mode")

app/api/docs.py

@@ -206,6 +206,16 @@ def get_document_content(
     basename matches). ``None`` — a non-git source, no row, or an
     ambiguous basename — keeps the payload byte-identical to pre-phase
     (both fields are dropped by the serializer, never ``null``).
+
+    Phase 163 (task 01): a page-qualifying PDF's payload additionally
+    carries ``page_strip_mode`` — ``"primary"`` (the phase-133 fallback
+    class: the strip IS the view) or ``"additive"`` (the phase-140
+    pictorial class: the text + Summary view with the strip behind a
+    Text/Pages toggle). Both strip keys are built off ``page_count`` and
+    dropped when it is NULL, so every non-qualifying payload stays
+    byte-identical to pre-phase (and a pre-phase row, whose mode column
+    is NULL, keeps the strip-only DOM the viewer assumes for an absent
+    mode).
     """
     row = db.execute(
         select(Document, func.count(Chunk.id).label("chunks"))
@@ -254,6 +264,16 @@ def get_document_content(
             if doc.page_count is not None
             else None
         ),
+        # Phase 163 (task 01): the strip CLASS (``"primary"`` = the
+        # fallback class, the strip IS the view; ``"additive"`` = the
+        # pictorial class, text + a Text/Pages toggle). The whole strip
+        # block keys off ``page_count`` — a row with a stray mode but no
+        # strip serializes NEITHER key — and a pre-phase row (mode NULL)
+        # serializes the key ABSENT (never null), which the viewer reads
+        # as ``"primary"`` (task 04).
+        page_strip_mode=(
+            doc.page_strip_mode if doc.page_count is not None else None
+        ),
     )
 
 

app/models.py

@@ -198,6 +198,24 @@ class Document(Base):
     #: cleanup deletes the copies with the row. No index: a display
     #: fact, never a query surface (the ``file_hash`` convention, 0025).
     page_count: Mapped[int | None] = mapped_column(Integer, default=None)
+    #: The STRIP CLASS behind a non-NULL :py:attr:`page_count` (phase 163,
+    #: LOCKED A1/A3) — the signal the viewer branches on so a text-bearing
+    #: book never loses its text:
+    #:
+    #: * ``"primary"`` — the phase-133 FALLBACK class (unusable text layer:
+    #:   the screenshots ARE the content): the strip IS the view, exactly
+    #:   as phase 148 shipped it;
+    #: * ``"additive"`` — the phase-140 PICTORIAL class (usable text layer,
+    #:   ≥1 sparse page described): the text view + the Summary panel, with
+    #:   the strip one Text/Pages toggle away (Text is the default).
+    #:
+    #: NULL = no strip (every non-qualifying document) — and a pre-phase-163
+    #: row keeps NULL while its ``page_count`` stands, which the wire treats
+    #: as ``"primary"`` (today's DOM) until the fast-path re-gate stamps it.
+    #: Stamped/cleared EXACTLY with ``page_count``: written by the importer's
+    #: persist (task 02) and the unchanged-path re-gate (task 03) only. Like
+    #: ``page_count`` it is a display fact, never a query surface — no index.
+    page_strip_mode: Mapped[str | None] = mapped_column(String(16), default=None)
 
     chunks: Mapped[list[Chunk]] = relationship(
         back_populates="document", cascade="all, delete-orphan"

app/rag/importer.py

@@ -190,15 +190,37 @@ dpi, every page) and persisted to ``settings.pdf_page_dir``
 ``<doc-id>-p{N}.png`` (N 1-based), with ``documents.page_count``
 (migration 0026) recording the page count — a per-page render
 failure leaves a GAP (no file — the route 404s that N, the viewer's
-note covers it). The existing copies are deleted before replacement
+note covers it). WRITE-FIRST (phase 163, task 02, LOCKED A4): the
-(a shrunk PDF's old ``p{N}`` tail cannot leak), a document that
+fresh render's bytes land ATOMICALLY per file (``<name>.png.tmp`` +
-STOPS qualifying is cleared (``page_count`` NULL + copies deleted),
+``os.replace``) BEFORE any old copy is removed — only the old-extent
-and the prune deletes the copies with the row
+files the fresh render did NOT write (a shrunk PDF's ``p{N}`` tail)
+go, so a rollback or crash can never restore a stale ``page_count``
+over ZERO files. The row is stamped with the strip's CLASS beside the
+count (``documents.page_strip_mode``, migration 0027 — ``primary`` for
+the fallback class, ``additive`` for the pictorial class, phase 163);
+a document that STOPS qualifying is cleared (both columns NULL +
+copies deleted — deletion-first is correct where there is no
+replacement set), and the prune deletes the copies with the row
 (:func:`_delete_page_copies`). FAIL-SOFT: a render or write failure
-leaves ``page_count`` NULL and logs a warning — the document indexes
+leaves ``page_count``/``page_strip_mode`` NULL and logs a warning —
-EXACTLY as pre-phase (the render is a display concern, it never
+the document indexes EXACTLY as pre-phase (the render is a display
-gates the indexing). The UNCHANGED fast path renders nothing (the
+concern, it never gates the indexing).
-copies persist — zero render work, zero orphans).
+
+PDF strip re-gate (phase 163, task 03, LOCKED A5): the UNCHANGED paths
+never assume "the copies persist" any more — a row that still claims a
+strip (``page_count`` set, a ``.pdf`` path) is re-gated
+deterministically from the CURRENT bytes + CURRENT settings
+(:func:`_regate_pdf_strip`, run at the tail of
+:func:`_unchanged_doc`): the class is re-stamped (``primary`` for the
+unusable-layer fallback class, ``additive`` for the sparse-page
+class), any missing ``<doc-id>-p{N}.png`` is RE-RENDERED through the
+same write-first atomic writer, and a strip that no longer qualifies
+(lowered ``pdf_pictorial_min_chars``, the kill switch off) is cleared
+with its copies. ZERO vision/LLM calls (the render is pure CPU, PLAN
+A21-safe), one greppable INFO line only when the re-gate ACTS, and
+fail-soft on every exception — the loss state (an external wipe, a
+pre-phase delete-first rollback) cannot survive one sync. Every other
+file's fast path still performs ZERO PDF parsing (the guard).
 
 Image-anchor context (phase 125, task 03, LOCKED A1/A2/A3): BEFORE the
 description, the same source's importable markdown files are scanned for
@@ -298,6 +320,7 @@ from __future__ import annotations
 
 import hashlib
 import logging
+import os
 import posixpath
 import re
 import uuid
@@ -1021,6 +1044,7 @@ async def _unchanged_doc(
     summary: ImportSummary,
     raw_date: datetime,
     file_digest: str,
+    full_path: Path | None = None,
 ) -> None:
     """Phase 147 (task 02): the ONE shared unchanged path — the three
     branches' near-identical unchanged blocks, one code site, reached
@@ -1064,7 +1088,22 @@ async def _unchanged_doc(
        the date block (the stamp set ``file_hash`` and the date did
        not change): quiet bookkeeping — no count, no log (the
        ``unchanged`` line already fired; a date-change run already
-       committed in step 5).
+       committed in step 5);
+    7. the phase-163 (task 03) PDF strip RE-GATE — a row that still
+       claims a strip (``page_count`` set) whose path ends in ``.pdf``
+       goes through :func:`_regate_pdf_strip`: the strip is re-derived
+       from the CURRENT bytes + CURRENT settings (deterministic, ZERO
+       LLM/vision calls) — a missing/changed ``page_strip_mode`` is
+       re-stamped, copies that vanished from disk are RE-RENDERED, a
+       strip that no longer qualifies is cleared — and whatever it
+       wrote commits in this helper's transaction. Unchanged files
+       re-gate the PDF strip deterministically (phase 163): the
+       pre-phase "the copies persist" assumption is gone, so a wipe of
+       ``pdf_page_dir`` or a pre-phase delete-first rollback cannot
+       outlive one sync. Every other file (``page_count`` NULL) pays
+       ZERO extra read — the guard keeps the fast path's contract, and
+       BOTH exits (the D1 manual-date return and the normal tail) run
+       the re-gate.
     """
     summary.unchanged += 1
     logger.info("import: unchanged source=%s path=%s", source, rel)
@@ -1078,10 +1117,32 @@ async def _unchanged_doc(
             session, doc=doc, source=source, rel=rel, content=doc.content,
             llm=llm, summary=summary, backfill=True,
         )
+
+    def _regate_strip() -> None:
+        """Step 7 (phase 163, task 03): the GUARDED strip re-gate — only
+        a row that carries a strip (``page_count`` set) on a ``.pdf``
+        path reaches :func:`_regate_pdf_strip`, so every other file's
+        fast path stays byte-identical (zero extra read, zero PDF
+        parse). The re-gate never raises (fail-soft inside); its writes
+        ride this helper's transaction."""
+        if doc.page_count is None or not str(doc.path).lower().endswith(".pdf"):
+            return
+        _regate_pdf_strip(
+            session,
+            doc,
+            full_path=full_path if full_path is not None else Path(doc.full_path),
+            llm=llm,
+            summary=summary,
+        )
+        if session.is_modified(doc):
+            session.commit()
+
     # 4 — the phase-106 (D1) manual-date lock: no date write; the
     # stamp, if dirty, is the only write (committed — see step 2's
-    # "next sync takes the fast path").
+    # "next sync takes the fast path") — plus whatever the phase-163
+    # re-gate repaired.
     if doc.created_at_manual:
+        _regate_strip()
         if session.is_modified(doc):
             session.commit()
         return
@@ -1103,6 +1164,165 @@ async def _unchanged_doc(
         # 6 — the stamp's commit (the date did not move): quiet
         # bookkeeping — the ``unchanged`` line already fired.
         session.commit()
+    # 7 — the phase-163 (task 03) strip re-gate (see the local helper).
+    _regate_strip()
+
+
+def _regate_pdf_strip(
+    session: Session,
+    doc: Document,
+    *,
+    full_path: Path,
+    llm: Embedder,
+    summary: ImportSummary,
+) -> None:
+    """Phase 163 (task 03, LOCKED A5): the DETERMINISTIC re-gate of a
+    PDF's display strip — run from the unchanged paths
+    (:func:`_unchanged_doc`) whenever a row still claims a strip
+    (``page_count`` set), so a strip can never die silently.
+
+    WHY (the 2026-10-07 live state, owner report): the ``Letumdei`` row
+    carried ``page_count = 36`` over ZERO files — the pre-phase-163
+    :func:`_persist_pdf_pages` deleted the old copies BEFORE rendering,
+    OUTSIDE the DB transaction, so a later per-file rollback (the
+    import loop's ``EmbeddingError`` handler) restored the stale count
+    over an empty directory; an external wipe of ``pdf_page_dir`` lands
+    exactly the same state. The phase-147 fast path believed "the
+    copies persist" and never re-rendered, so every
+    ``GET /api/documents/<id>/pages/N`` 404'd and the viewer showed
+    "Page N unavailable." on all 36 pages — on every sync, forever.
+
+    WHAT IT DOES — the changed-path gate mirrored exactly, from the
+    CURRENT file bytes + CURRENT settings, with ZERO vision/LLM calls
+    (:func:`app.rag.pdf.render_pdf_pages` is pure CPU — PLAN A21-safe,
+    no model, no budget):
+
+    * ``not is_usable_pdf_text(extract_pdf_text(full_path))`` → the
+      phase-133 fallback class → mode ``"primary"`` (the strip IS the
+      view — the same doc as today);
+    * else ``settings.pdf_pictorial`` ON and
+      :func:`app.rag.pdf.pictorial_pages` NON-EMPTY over
+      :func:`app.rag.pdf.extract_pdf_page_texts` → the phase-140
+      pictorial class → mode ``"additive"`` (the text view + the strip
+      behind a toggle). The candidates are the existing gate's
+      VERBATIM (text-carrying pages under ``min_chars``; a page with no
+      text has no key in the map and is not a candidate) — the re-gate
+      adds NO new eligibility semantics;
+    * else the strip no longer qualifies (the owner lowered
+      ``pdf_pictorial_min_chars``, the kill switch went off, the layer
+      gained substance) → the REPAIR arm: the copies go
+      (``_delete_page_copies`` reads ``doc.page_count``, so it runs
+      while the count still describes the extent) and BOTH columns are
+      cleared.
+
+    For a re-gated-ELIGIBLE doc the copies are VERIFIED: any missing
+    ``<doc-id>-p{N}.png`` in 1..``page_count`` re-renders EVERY page
+    through the changed path's own write-first atomic writer
+    (:func:`_write_page_copies` — tmp + ``os.replace``), then the
+    stale tail beyond the new count goes and ``page_count`` follows the
+    fresh render's length (a drift names both counts in a WARNING — on
+    unchanged bytes a drift is pathological). A fresh render producing
+    NO pages (unopenable right after the fast path hashed the file)
+    leaves the row exactly as it was + a WARNING: the re-gate never
+    clears a strip on a transient.
+
+    The mode re-stamp is SILENT when it only fills a NULL (a pre-phase
+    row learning its class is beneficial bookkeeping — it rides the
+    same INFO line only when real work followed); a change of an
+    ALREADY-SET mode (the kill switch, a class flip) says so
+    (``mode <old>-><new>``). The ONE greppable INFO line fires only
+    when the re-gate ACTS — never per unchanged doc.
+
+    FAIL-SOFT: the whole body is wrapped — any exception is a WARNING
+    and the sync continues (the strip is a display concern). A repair's
+    DB writes ride the caller's transaction, so an exception raised
+    after them rolls back with the file like any other per-file
+    failure. *summary* keeps the unchanged path's signature shape (and
+    a future counter site); the re-gate books nothing there — its stats
+    are log-only (the phase-133/140 convention).
+    """
+    source, rel = doc.source, doc.path
+    try:
+        page_dir = Path(llm.settings.pdf_page_dir).expanduser()
+        # The gate, mirrored from the changed path (the phase-133
+        # precedence over the phase-140 pass, the kill switch included).
+        if not is_usable_pdf_text(extract_pdf_text(full_path)):
+            strip_class: str | None = "fallback"
+        elif llm.settings.pdf_pictorial and pictorial_pages(
+            extract_pdf_page_texts(full_path),
+            min_chars=llm.settings.pdf_pictorial_min_chars,
+        ):
+            strip_class = "pictorial"
+        else:
+            strip_class = None
+        if strip_class is None:
+            # The REPAIR arm — the strip no longer qualifies: delete
+            # while the count still names the extent, then clear both
+            # columns (the same clearing the changed path does for a
+            # doc that stopped qualifying).
+            _delete_page_copies(doc, page_dir)
+            doc.page_count = None
+            doc.page_strip_mode = None
+            logger.info(
+                "import: PDF strip re-gate source=%s path=%s — cleared stale strip",
+                source, rel,
+            )
+            return
+        mode = "primary" if strip_class == "fallback" else "additive"
+        old_mode = doc.page_strip_mode
+        old_count = doc.page_count or 0
+        actions: list[str] = []
+        missing = [
+            n
+            for n in range(1, old_count + 1)
+            if not (page_dir / f"{doc.id}-p{n}.png").is_file()
+        ]
+        if missing:
+            pages = render_pdf_pages(full_path)
+            if not pages or not any(pages):
+                # Unopenable (or every page a placeholder) — right
+                # after the fast path hashed the file this is
+                # pathological: KEEP the row as-is (never clear a strip
+                # on a transient) and let the next sync try again.
+                logger.warning(
+                    "import: PDF strip re-gate source=%s path=%s — re-render "
+                    "produced no pages (state kept)",
+                    source, rel,
+                )
+                return
+            page_dir.mkdir(parents=True, exist_ok=True)
+            written_ns = _write_page_copies(doc, page_dir, pages)
+            _delete_stale_page_tail(
+                doc_id=doc.id,
+                page_dir=page_dir,
+                old_extent=old_count,
+                written_ns=written_ns,
+            )
+            doc.page_count = len(pages)
+            actions.append(f"healed {len(missing)} pages")
+            if doc.page_count != old_count:
+                logger.warning(
+                    "import: PDF strip re-gate source=%s path=%s — page_count "
+                    "drift %s->%s on unchanged bytes",
+                    source, rel, old_count, doc.page_count,
+                )
+        if doc.page_strip_mode != mode:
+            # NULL -> stamped stays SILENT (it rides the line only when
+            # work followed); a change of an already-set mode is news.
+            doc.page_strip_mode = mode
+            if old_mode is not None:
+                actions.append(f"mode {old_mode}->{mode}")
+        if actions:
+            logger.info(
+                "import: PDF strip re-gate source=%s path=%s — %s",
+                source, rel, " | ".join(actions),
+            )
+    except Exception as e:
+        # Fail-soft: the strip never breaks a sync.
+        logger.warning(
+            "import: PDF strip re-gate failed source=%s path=%s — %s",
+            source, rel, e,
+        )
 
 
 async def _index_file(
@@ -1186,7 +1406,7 @@ async def _index_file(
         # helper, nothing else does.
         await _unchanged_doc(
             session, doc, source=source, rel=rel, llm=llm, summary=summary,
-            raw_date=raw_date, file_digest=file_digest,
+            raw_date=raw_date, file_digest=file_digest, full_path=full_path,
         )
         return
     # Phase 122 (task 02): the image branch FIRST. Only reachable while
@@ -1223,7 +1443,7 @@ async def _index_file(
         # file hash, so the NEXT sync takes the fast path above.
         await _unchanged_doc(
             session, doc, source=source, rel=rel, llm=llm, summary=summary,
-            raw_date=raw_date, file_digest=file_digest,
+            raw_date=raw_date, file_digest=file_digest, full_path=full_path,
         )
         return
 
@@ -1506,19 +1726,24 @@ async def _index_pdf_file(
       (:func:`_unchanged_doc`) — no re-extract, no re-embed, no image
       pass, the backfill/manual-date/date-refresh semantics in one
       code site — and the helper stamps the new file hash (the next
-      sync takes the fast path). This path renders NO pages and
+      sync takes the fast path). This path renders NO pages through
-      touches NO ``page_count`` (the copies persist — zero render
+      THIS branch, but a row that carries a strip is re-gated by the
-      work).
+      helper's phase-163 tail (:func:`_regate_pdf_strip` — the class
+      re-stamped, missing copies healed, a stale strip cleared; zero
+      vision/LLM work).
     * Phase 148 (task 01): in this NEW/CHANGED path, RIGHT AFTER the
       row's flush and BEFORE the chunk pipeline, a page-QUALIFYING
       PDF (the phase-133 fallback fired, or the phase-140 pictorial
       pass described ≥1 page — the signals recorded above) gets its
       every page rendered to ``settings.pdf_page_dir`` as
-      ``<doc-id>-p{N}.png`` and ``documents.page_count`` set
+      ``<doc-id>-p{N}.png`` and ``documents.page_count`` +
-      (:func:`_persist_pdf_pages` — stale copies deleted first, a
+      ``documents.page_strip_mode`` set
-      stopped-qualifier cleared, a per-page ``b""`` a gap). FAIL-SOFT
+      (:func:`_persist_pdf_pages` — WRITE-FIRST (phase 163, LOCKED
-      — a render/write exception leaves ``page_count`` NULL + a
+      A4): the new copies land atomically per file BEFORE any old copy
-      warning; the document indexes exactly as pre-phase.
+      is removed, a stopped-qualifier cleared, a per-page ``b""`` a
+      gap). FAIL-SOFT — a render/write exception leaves
+      ``page_count``/``page_strip_mode`` NULL + a warning; the
+      document indexes exactly as pre-phase.
     * THEN the embedded images: :func:`_index_pdf_embedded_images`
       (A3 — the phase-122 pipeline, gated by the ``images`` toggle).
 
@@ -1631,7 +1856,7 @@ async def _index_pdf_file(
         # fast path).
         await _unchanged_doc(
             session, doc, source=source, rel=rel, llm=llm, summary=summary,
-            raw_date=raw_date, file_digest=file_digest,
+            raw_date=raw_date, file_digest=file_digest, full_path=full_path,
         )
         return
 
@@ -1673,23 +1898,34 @@ async def _index_pdf_file(
     # its id — the copy names need it) and BEFORE the chunk pipeline
     # (the count commits ATOMICALLY with the doc + chunks). Runs ONLY
     # in this new/changed path: the phase-147 fast path returned
-    # before the branch (an unchanged PDF re-renders NOTHING — the
+    # before the branch and the content-digest hit above took the
-    # copies persist) and the content-digest hit above took the shared
+    # shared unchanged path — neither renders here; the phase-163
-    # unchanged path (no page work there either). FAIL-SOFT: the
+    # task-03 re-gate (:func:`_regate_pdf_strip`, run by
-    # render is a DISPLAY concern — any exception leaves
+    # :func:`_unchanged_doc`) owns the repair of an unchanged doc's
-    # ``page_count`` NULL, the warning is the signal, and the document
+    # strip. FAIL-SOFT: the render is a DISPLAY
-    # proceeds to chunk/embed/summary EXACTLY as pre-phase.
+    # concern — any exception leaves ``page_count``/``page_strip_mode``
+    # NULL, the warning is the signal, and the document proceeds to
+    # chunk/embed/summary EXACTLY as pre-phase.
+    # Phase 163 (task 02): the eligibility decision carried as a CLASS
+    # (LOCKED A7 — the truth table is the phase-148 rule verbatim: a
+    # non-None class ⇔ fallback_fired or pictorial_described). The
+    # class decides the stamped `page_strip_mode` ("fallback" →
+    # "primary", "pictorial" → "additive"), nothing else.
+    strip_class = (
+        "fallback" if fallback_fired else ("pictorial" if pictorial_described else None)
+    )
     try:
         _persist_pdf_pages(
             doc,
             full_path=full_path,
             page_dir=Path(llm.settings.pdf_page_dir).expanduser(),
-            eligible=fallback_fired or pictorial_described,
+            strip_class=strip_class,
             source=source,
             rel=rel,
         )
     except Exception as e:
         doc.page_count = None
+        doc.page_strip_mode = None
         logger.warning(
             "import: PDF page render failed source=%s path=%s — %s",
             source, rel, e,
@@ -1991,84 +2227,186 @@ def _persist_pdf_pages(
     *,
     full_path: Path,
     page_dir: Path,
-    eligible: bool,
+    strip_class: str | None,
     source: str,
     rel: str,
 ) -> None:
-    """Phase 148 (task 01): the persistent page copies of one NEW/CHANGED
+    """Phase 148 (task 01) + phase 163 (task 02): the persistent page
-    PDF — the display strip's bytes (task 02's public route serves them).
+    copies of one NEW/CHANGED PDF — the display strip's bytes (task 02's
+    public route serves them) — written WRITE-FIRST and stamped with the
+    strip's CLASS.
 
-    * ELIGIBLE (the phase-133 fallback fired, or the phase-140
+    *strip_class* is the sync's eligibility decision carried as a CLASS
-      pictorial pass described ≥1 page — the sync's own decisions, see
+    (phase 163, LOCKED A7): ``"fallback"`` (the phase-133 fallback fired
-      the caller's signals): EVERY page of the PDF is rendered (150 dpi,
+    — the screenshots ARE the content → stamped mode ``"primary"``),
+    ``"pictorial"`` (the phase-140 pass described ≥1 sparse page →
+    stamped mode ``"additive"``), or ``None`` (no strip). The truth
+    table is the phase-148 eligibility rule VERBATIM: a non-None class
+    ⇔ the old ``fallback_fired or pictorial_described`` — this phase
+    changes only what the class DOES (the stamp), never who qualifies.
+
+    * ELIGIBLE: EVERY page of the PDF is rendered (150 dpi,
       :func:`render_pdf_pages`' default — the phase-133 constant; the
       ``pages=None`` "list index *i* is page *i*+1" contract) and each
-      NON-EMPTY PNG is written to *page_dir* as
+      NON-EMPTY PNG lands in *page_dir* as ``<doc-id>-p{N}.png``
-      ``<doc-id>-p{N}.png`` (N 1-based), then ``doc.page_count`` is set
+      (N 1-based) through an ATOMIC per-file write — the shared writer
-      to the PDF's page count (``len(pages)`` — one slot per page, so
+      :func:`_write_page_copies` puts the bytes on ``<name>.png.tmp``
-      the list IS the page count for an openable file). A per-page
+      first and ``os.replace`` moves them onto the final name, so a
+      concurrent reader of the route NEVER sees a torn PNG (the phase
+      163 task-03 re-gate heals a strip through THE SAME writer) — and
+      the success tail stamps ``doc.page_count = len(pages)``
+      (one slot per page — the list IS the page count for an openable
+      file) and ``doc.page_strip_mode`` from the class. A per-page
       ``b""`` (an unreadable page) is a GAP, not an error: no file is
       written for it (the route 404s that N, the viewer's note covers
       it), and the count still covers the full range.
-    * the STALE-REPLACEMENT convention (phase 122's image copies): the
+    * WRITE-FIRST (phase 163, LOCKED A4 — the 2026-10-07 live loss): the
-      doc's EXISTING copies (1..``doc.page_count`` — the row knows its
+      OLD copies are deleted LAST, never first. The pre-phase order
-      own extent, read BEFORE any change) are deleted FIRST, so a
+      (blanket delete → render → write) ran OUTSIDE the DB transaction,
-      changed PDF's old page set (e.g. 30 pages → 24) cannot leak
+      so any later per-file rollback (the import loop's
-      ``p25..p30``; a fresh doc's NULL count is a no-op.
+      ``EmbeddingError`` handler) — or a crash between delete and write —
-    * NOT eligible, or an unopenable file (:func:`render_pdf_pages` →
+      restored the stale ``page_count`` over ZERO files and the strip
-      ``[]``), or a render with ZERO non-empty pages: ``page_count`` is
+      404ed forever. Now only the old-extent files (1..the count read
-      None + the existing copies deleted — a document that STOPPED
+      BEFORE any change) that the fresh render did NOT write — a shrunk
-      qualifying loses its strip (e.g. a new version gained a real text
+      PDF's ``p{N}`` tail, a new gap — are removed, and only once every
-      layer).
+      replacement byte is durably in place; a file about to be
-    * FAIL-SOFT: a write failure removes the copies THIS CALL wrote
+      overwritten is never deleted at all.
-      (no partial-write orphans) and re-raises — the caller
+    * NOT eligible (``strip_class is None``), an unopenable file
-      (:func:`_index_pdf_file`) catches, leaves ``page_count`` NULL,
+      (:func:`render_pdf_pages` → ``[]``), or a render with ZERO
-      logs the warning, and the document indexes exactly as pre-phase
+      non-empty pages: the existing copies are deleted (the helper reads
-      (the render is a display concern — it never gates indexing).
+      ``doc.page_count``, so it runs BEFORE the nulling) and BOTH
+      ``page_count`` and ``page_strip_mode`` are cleared — the
+      deletion-first order is correct HERE, there is no replacement set
+      to protect; a document that STOPPED qualifying loses its strip
+      (e.g. a new version gained a real text layer).
+    * FAIL-SOFT: a write failure removes this call's ``.tmp`` files and
+      the final-name files it CREATED — but NEVER a file it OVERWROTE
+      (the replaced copy keeps standing as the durable new bytes; a
+      failed call may change a copy's bytes, it may never remove a
+      copy) — and re-raises: the caller (:func:`_index_pdf_file`)
+      catches, leaves ``page_count`` AND ``page_strip_mode`` NULL, logs
+      the warning, and the document indexes exactly as pre-phase (the
+      render is a display concern — it never gates indexing).
 
     *page_dir* is ``Path(settings.pdf_page_dir).expanduser()`` (the
     ``image_dir`` convention — raw string, expanduser'd by the
     importer); the directory is created on demand, only when a page is
     actually written (an ineligible doc never creates it).
     """
-    # 1 — the stale-replacement: this doc's EXISTING copies go first —
+    if strip_class is None:
-    # a changed PDF's old page set must not leak, and a doc that
+        # No strip — the DELETION-first order is correct here (there is
-    # stopped qualifying loses its strip (a fresh doc's NULL count is
+        # no replacement set to protect); the helper reads
-    # a no-op; missing files are a silent no-op).
+        # doc.page_count, so it runs BEFORE the nulling (a fresh doc's
-    _delete_page_copies(doc, page_dir)
+        # NULL count is a no-op; missing files are a silent no-op).
-    if not eligible:
+        _delete_page_copies(doc, page_dir)
         doc.page_count = None
+        doc.page_strip_mode = None
         return
     pages = render_pdf_pages(full_path)
     if not pages or not any(pages):
         # Unopenable file (the render's ``[]``) or every page a
-        # placeholder: no strip — fail-soft, the index is untouched
+        # placeholder: no strip and NO replacement set — the existing
-        # (the render helper already logged the per-page failures).
+        # copies go with the cleared row (helper BEFORE the nulling),
+        # fail-soft, the index is untouched (the render helper already
+        # logged the per-page failures).
         logger.warning(
             "import: PDF page render produced no pages source=%s path=%s",
             source, rel,
         )
+        _delete_page_copies(doc, page_dir)
         doc.page_count = None
+        doc.page_strip_mode = None
         return
+    # The row's OWN old extent, read BEFORE any change — the stale-tail
+    # delete below walks it (the write-first contract never deletes a
+    # file outside this extent, and never before its replacement is
+    # durable).
+    old_extent = doc.page_count or 0
+    # 1 — WRITE FIRST, atomically per file (the shared writer, phase
+    # 163 task 02 — the phase-163 task-03 re-gate heals through it).
+    written_ns = _write_page_copies(doc, page_dir, pages)
+    # 2 — the stale tail goes LAST.
+    _delete_stale_page_tail(
+        doc_id=doc.id,
+        page_dir=page_dir,
+        old_extent=old_extent,
+        written_ns=written_ns,
+    )
+    doc.page_count = len(pages)
+    doc.page_strip_mode = "primary" if strip_class == "fallback" else "additive"
+
+
+def _write_page_copies(doc: Document, page_dir: Path, pages: Sequence[bytes]) -> set[int]:
+    """Phase 163 (task 02, LOCKED A4; extracted for the task-03 re-gate,
+    which heals a strip through the SAME writer): write every NON-EMPTY
+    PNG of *pages* to ``page_dir/<doc-id>-p{N}.png`` (N 1-based) and
+    return the set of page numbers written.
+
+    * ATOMIC per file — the bytes land on ``<name>.png.tmp`` and
+      ``os.replace`` moves them onto the final name, so a concurrent
+      reader of the ``/pages/{n}`` route never sees a torn PNG.
+    * A per-page ``b""`` (the render's failure placeholder) is the GAP:
+      no file, no entry in the returned set.
+    * The directory is created on demand (a doc that never writes never
+      creates it).
+    * A failure leaves NO orphan: this call's ``.tmp`` files and the
+      finals it CREATED are removed — an OVERWRITTEN final keeps
+      standing (deleting it would re-open the 2026-10-07 hole with one
+      page instead of the whole strip) — and the exception re-raises so
+      the caller keeps its fail-soft contract.
+    """
     page_dir.mkdir(parents=True, exist_ok=True)
-    written: list[Path] = []
+    written_tmps: list[Path] = []
+    created_finals: list[Path] = []  # finals that did NOT exist before
+    written_ns: set[int] = set()
     try:
         for n, png in enumerate(pages, start=1):
             if not png:
                 # A per-page render failure: the GAP (no file — the
                 # route 404s that N, the viewer's note covers it).
                 continue
-            path = page_dir / f"{doc.id}-p{n}.png"
+            final = page_dir / f"{doc.id}-p{n}.png"
-            path.write_bytes(png)
+            tmp = page_dir / f"{doc.id}-p{n}.png.tmp"
-            written.append(path)
+            # Track the tmp BEFORE the write: a failure mid-write may
+            # leave it on disk, and the cleanup must find it either way.
+            written_tmps.append(tmp)
+            existed = final.exists()
+            tmp.write_bytes(png)
+            os.replace(tmp, final)  # atomic per file — no torn reads
+            written_ns.add(n)
+            if not existed:
+                created_finals.append(final)
     except Exception:
-        # A partial write leaves no orphan: the copies this call wrote
+        # A partial write leaves no orphan: this call's tmp files and
-        # go, then the re-raise hands the failure to the caller's
+        # the finals it CREATED go (an OVERWRITTEN final keeps standing
-        # fail-soft handler (page_count NULL + the warning).
+        # — deleting it would re-open the 2026-10-07 hole with one page
-        for path in written:
+        # instead of the whole strip), then the re-raise hands the
+        # failure to the caller's fail-soft handler (page_count +
+        # page_strip_mode NULL + the warning).
+        for path in [*written_tmps, *created_finals]:
             path.unlink(missing_ok=True)
         raise
-    doc.page_count = len(pages)
+    return written_ns
+
+
+def _delete_stale_page_tail(
+    *, doc_id: uuid.UUID, page_dir: Path, old_extent: int, written_ns: set[int]
+) -> None:
+    """Phase 163 (task 02, LOCKED A4 — the LAST step of the write-first
+    order): remove the files of the row's OLD extent (1..*old_extent*)
+    the fresh render did NOT write — a shrunk PDF's ``p{N}`` tail, a new
+    gap. Runs only once every replacement byte is durable, never touches
+    a file outside the old extent, and never touches a file that was just
+    overwritten. Best-effort like :func:`_delete_page_copies`: an
+    unreadable file is logged, never raised (the sync survives)."""
+    for n in range(1, old_extent + 1):
+        if n in written_ns:
+            continue
+        path = page_dir / f"{doc_id}-p{n}.png"
+        try:
+            path.unlink(missing_ok=True)
+        except OSError as e:
+            logger.warning("import: could not delete PDF page copy %s — %s", path, e)
 
 
 async def _index_pdf_embedded_images(
@@ -3271,7 +3609,7 @@ async def _index_image_file(
         # ``_store_summary`` backfill re-stores it verbatim.
         await _unchanged_doc(
             session, doc, source=source, rel=rel, llm=llm, summary=summary,
-            raw_date=raw_date, file_digest=file_digest,
+            raw_date=raw_date, file_digest=file_digest, full_path=full_path,
         )
         return
 

app/schemas.py

@@ -636,6 +636,19 @@ class DocContent(BaseModel):
     #: ``null``, the :attr:`image_url` omission precedent), so a
     #: non-qualifying doc's payload is byte-identical to pre-phase.
     page_urls: list[str] | None = None
+    #: Phase 163 (task 01, LOCKED A2/A3) — the STRIP CLASS behind
+    #: :attr:`page_urls`: ``"primary"`` = the phase-133 fallback class
+    #: (an unusable text layer — the screenshots ARE the content, so the
+    #: strip IS the view, exactly as phase 148 shipped it); ``"additive"``
+    #: = the phase-140 pictorial class (a usable text layer with ≥1 sparse
+    #: page described — the text view + the Summary panel, the strip one
+    #: Text/Pages toggle away). ABSENT from the wire whenever there is no
+    #: strip (``None`` → dropped by the serializer — never ``null``, the
+    #: :attr:`image_url` omission rule), which is ALSO how a pre-phase row
+    #: (``page_count`` set, the column NULL) serializes: the viewer treats
+    #: an absent mode as ``"primary"`` (task 04), so every existing row
+    #: keeps today's DOM until the fast-path re-gate stamps it.
+    page_strip_mode: str | None = None
 
     @model_serializer(mode="wrap")
     def _serialize(self, handler: SerializerFunctionWrapHandler) -> Any:
@@ -648,6 +661,11 @@ class DocContent(BaseModel):
             data.pop("origin_branch", None)
         if self.page_urls is None:
             data.pop("page_urls", None)
+        # Phase 163 (task 01): the strip class rides the strip — ABSENT
+        # (never null) with no strip, so a non-qualifying doc's payload is
+        # byte-identical to pre-phase.
+        if self.page_strip_mode is None:
+            data.pop("page_strip_mode", None)
         return data
 
 

frontend/assets/document.js

@@ -144,8 +144,9 @@
  * wire for every other document, so they render the EXACT pre-phase
  * DOM on BOTH surfaces — the load-bearing byte-identity property),
  * #doc-content renders a .doc-pages strip INSTEAD of the text
- * content AND the .doc-summary panel: one <figure class="doc-page">
+ * content AND the .doc-summary panel FOR THE PRIMARY CLASS ONLY
- * per URL — a lazy <img> (alt = "Page N of <title>", the WCAG alt
+ * (phase 163 — the absent/`primary` mode; see the next paragraph):
+ * one <figure class="doc-page"> per URL — a lazy <img> (alt = "Page N of <title>", the WCAG alt
  * contract — the title as the fallback context) + a "Page N"
  * <figcaption> — while the meta badge row renders exactly as today.
  * The owner's contract: the per-page descriptions stay in the DB for
@@ -160,6 +161,26 @@
  * per-surface copy); createElement + properties/textContent only (the
  * never-build-HTML rule — no innerHTML).
  *
+ * Phase 163 (task 04, LOCKED A1/A2/A6): the strip's CLASS decides the
+ * view. page_strip_mode === "additive" (the pictorial class — a USABLE
+ * text layer, so the pages are a convenience, not the content) keeps
+ * the TEXT doc EXACTLY as a non-qualifying doc renders it (Summary
+ * panel + content — the pre-148 branch reused), and the strip sits one
+ * Text/Pages toggle away: a .doc-view-toggle group (role=group,
+ * aria-label="Document view", two <button aria-pressed>) above a
+ * .doc-views wrapper holding the text view + the strip; Text is
+ * pressed by DEFAULT (A6 — never a sparse-ratio flip), Pages flips the
+ * wrapper's data-view (CSS shows the strip / hides the text — the
+ * strip node stays MOUNTED, its lazy imgs never refetch on flip-flop).
+ * The text is NEVER hidden for this class. The absent/"primary" mode
+ * (the fallback class — the screenshots ARE the content, and every
+ * pre-163 stamped row) keeps the strip-only DOM above byte-identically
+ * (the tolerant read: only the exact "additive" string switches on the
+ * toggle); falsy page_urls stay untouched. The wrapper exists ONLY on
+ * additive docs — every other document's DOM is unchanged. Toggle
+ * labels are fixed strings (no doc-derived text); createElement +
+ * properties/textContent only.
+ *
  * Phase 153 (task 02): the meta row's "Copy path" control — the
  * read-tool `source/path` form in the clipboard with one click
  * (ASSUMPTION A1 — the exact `source/path` the read tool, the
@@ -361,7 +382,9 @@ export function renderDocument(doc, { titleEl, metaEl, contentEl }) {
   // Phase 148 (task 03): the page strip — the payload's page_urls
   // (task 02's endpoint: the exact 1-based route list for a
   // page-qualifying PDF; ABSENT on the wire for every other
-  // document) is the ONE branch that replaces the content slot: the
+  // document) is the ONE branch that replaces the content slot (phase
+  // 163: the PRIMARY class only — an "additive" doc falls through to
+  // the text rendering with the strip behind a toggle, D4/A1): the
   // markdown/.doc-raw content block AND the labeled Summary panel
   // (including its admin edit wiring — moot without the panel) are
   // NOT built. The per-page descriptions stay in the DB for Q&A (the
@@ -372,16 +395,40 @@ export function renderDocument(doc, { titleEl, metaEl, contentEl }) {
   // below (byte-identical for every other doc — the load-bearing
   // property).
   const pageUrls = Array.isArray(doc.page_urls) ? doc.page_urls : null;
-  if (pageUrls && pageUrls.length > 0) {
+  const stripEligible = pageUrls !== null && pageUrls.length > 0;
+  // Phase 163 (task 04, LOCKED A2): the strip's CLASS picks the view.
+  // Only the exact "additive" string switches on the toggle — absent
+  // (pre-163 rows, and the omission rule's non-qualifying docs) or
+  // "primary" read as primary (the tolerant-reading house rule), so
+  // the fallback class keeps the strip-only DOM byte-identically.
+  const additive = stripEligible && doc.page_strip_mode === "additive";
+  if (stripEligible && !additive) {
     contentEl.appendChild(docPagesStrip(doc, pageUrls));
     return;
   }
+  // Phase 163 (task 04, LOCKED A1/A6): an ADDITIVE doc renders the
+  // text doc EXACTLY as a non-qualifying one does (Summary panel +
+  // content below, reused unchanged) INSIDE a .doc-views wrapper —
+  // the Text/Pages toggle above it flips the wrapper's data-view
+  // (CSS-driven show/hide, never a re-render); the strip mounts once
+  // AFTER the text view and stays mounted across flips (lazy imgs,
+  // display:none while hidden, never fetch). Every non-additive doc
+  // keeps slot === contentEl — its DOM is byte-identical to pre-phase.
+  let views = null;
+  let slot = contentEl;
+  if (additive) {
+    views = document.createElement("div");
+    views.className = "doc-views";
+    views.setAttribute("data-view", "text"); // LOCKED A6: Text is the default view
+    contentEl.append(docViewToggle(views), views);
+    slot = views;
+  }
   // Phase 122 (task 04): an image document — the persistent bytes
   // render as the <img> block FIRST; the description (= doc.content)
   // follows in the normal content slot below (the plain-content path
   // — the last branch in this function).
   if (doc.is_image) {
-    contentEl.appendChild(docImageBlock(doc));
+    slot.appendChild(docImageBlock(doc));
   }
   // Phase 36: the summary panel — labeled section ABOVE the original
   // content, on BOTH surfaces (page + modal) through this one core.
@@ -405,7 +452,7 @@ export function renderDocument(doc, { titleEl, metaEl, contentEl }) {
     body.className = "doc-summary-text";
     body.textContent = doc.summary; // text node — XSS contract unchanged
     section.append(title, body);
-    contentEl.appendChild(section);
+    slot.appendChild(section);
     // Phase 57 (task 02, D4): the section above is the phase-36 shape
     // for EVERYONE — only an authenticated admin (docAdminReady(), the
     // module-cached whoami promise) then gains the header row + Edit
@@ -419,13 +466,59 @@ export function renderDocument(doc, { titleEl, metaEl, contentEl }) {
     const wrap = document.createElement("div");
     wrap.className = "doc-md";
     wrap.innerHTML = renderMarkdown(doc.content); // escape-first: XSS-safe
-    contentEl.appendChild(wrap);
+    slot.appendChild(wrap);
   } else {
     const pre = document.createElement("pre");
     pre.className = "doc-raw";
     pre.textContent = doc.content; // text node: never parsed as HTML
-    contentEl.appendChild(pre);
+    slot.appendChild(pre);
   }
+  // Phase 163 (task 04): the additive strip mounts once, AFTER the
+  // text view, inside the wrapper — hidden by [data-view="text"]
+  // (the lazy imgs never fetch until Pages is chosen; the node
+  // persists across flips — no refetch).
+  if (views !== null) {
+    views.appendChild(docPagesStrip(doc, pageUrls));
+  }
+}
+
+/* ---------- view toggle (phase 163, task 04 — LOCKED A1/A6) ----------
+ * The Text/Pages toggle of an ADDITIVE strip doc: one
+ * <div class="doc-view-toggle" role="group" aria-label="Document
+ * view"> with two <button type="button" class="doc-view-btn"
+ * aria-pressed>. Text is pressed initially (A6 — no sparse-ratio
+ * default flip). A click mutates ONLY the two aria-pressed states and
+ * the .doc-views wrapper's data-view attribute — the CSS shows/hides
+ * the text view and the strip (never a re-render, never a refetch).
+ * The labels are FIXED strings (no doc-derived text inside the
+ * toggle); createElement + properties/textContent only (the
+ * never-build-HTML rule — no innerHTML). */
+function docViewToggle(viewsEl) {
+  const group = document.createElement("div");
+  group.className = "doc-view-toggle";
+  group.setAttribute("role", "group");
+  group.setAttribute("aria-label", "Document view");
+  const textBtn = viewToggleButton("Text");
+  const pagesBtn = viewToggleButton("Pages");
+  const select = (view) => {
+    viewsEl.setAttribute("data-view", view);
+    textBtn.setAttribute("aria-pressed", view === "text" ? "true" : "false");
+    pagesBtn.setAttribute("aria-pressed", view === "pages" ? "true" : "false");
+  };
+  textBtn.setAttribute("aria-pressed", "true"); // LOCKED A6: Text pressed by default
+  pagesBtn.setAttribute("aria-pressed", "false");
+  textBtn.addEventListener("click", () => select("text"));
+  pagesBtn.addEventListener("click", () => select("pages"));
+  group.append(textBtn, pagesBtn);
+  return group;
+}
+
+function viewToggleButton(label) {
+  const btn = document.createElement("button");
+  btn.type = "button";
+  btn.className = "doc-view-btn";
+  btn.textContent = label; // fixed string — never doc-derived
+  return btn;
 }
 
 /* ---------- image block (phase 122, task 04) ----------

frontend/assets/styles.css

@@ -5311,6 +5311,59 @@ a.answer-image:hover { text-decoration: underline; } /* flat — the caption und
   padding: 0.35rem 0 0;
 }
 
+/* Additive strip view (phase 163, task 04): the pictorial-class PDF
+   — a USABLE text layer — keeps the text doc (Summary panel +
+   content) with the page strip one Text/Pages toggle away (LOCKED
+   A1/A6). The toggle is the house ghost-pill idiom (the .auth-link
+   family): the two pills sit adjacent in the labeled group, ink-soft
+   on the transparent pill (AA on --bg/--surface), hover the
+   brand-soft/brand-ink pair; the PRESSED button is the brand fill +
+   heavier weight — text + style, never color alone (B5) — with --bg
+   ink on --brand (5.2:1, the .share-chat-btn pair). ≥44px targets
+   (§7.1); :focus-visible rides the global 3px ring. The view swap is
+   CSS-driven off the .doc-views wrapper's data-view: the strip is
+   display:none by default (its lazy imgs never fetch while hidden)
+   and shows on [data-view="pages"], which hides the text view instead
+   — the nodes stay mounted across flips (no re-render, no refetch).
+   Phase-08 tokens only (no new hue — the monochrome invariant). */
+.doc-view-toggle {
+  display: flex;
+  gap: 0.4rem;
+  width: 100%;
+  max-width: var(--chat-column);
+  margin-inline: auto;
+  margin-bottom: 0.75rem;
+}
+.doc-view-btn {
+  display: inline-flex;
+  align-items: center;
+  justify-content: center;
+  min-height: 44px;
+  padding: 0.5rem 0.9rem;
+  border-radius: 999px;
+  border: 1px solid var(--line);
+  background: transparent;
+  color: var(--ink-soft);
+  font: inherit;
+  font-weight: 600;
+  font-size: 0.95rem;
+  white-space: nowrap;
+  cursor: pointer;
+}
+.doc-view-btn:hover { background: var(--brand-soft); color: var(--brand-ink); }
+.doc-view-btn[aria-pressed="true"] {
+  background: var(--brand);
+  border-color: var(--brand);
+  color: var(--bg); /* dark ink on brand: 5.2:1 (AA) */
+  font-weight: 700; /* pressed = weight + fill, never color alone (B5) */
+}
+.doc-views .doc-pages { display: none; } /* Text default — hidden lazy imgs never fetch */
+.doc-views[data-view="pages"] .doc-summary,
+.doc-views[data-view="pages"] .doc-md,
+.doc-views[data-view="pages"] .doc-raw,
+.doc-views[data-view="pages"] .doc-image { display: none; }
+.doc-views[data-view="pages"] .doc-pages { display: grid; }
+
 /* Designed not-found state (no emoji — plain SVG mark, phase 08 rule). */
 .doc-not-found {
   width: 100%;

tests/e2e/test_pdf_pages_in_sources.py

@@ -36,33 +36,38 @@ token (the phase-126 convention — the owner's local ``.env`` may lack
 it, and without it the walk is blind to the fixture).
 
 MOCK-ONLY suite: ``E2E_REAL_LLM=1`` is not supported — the described
-blocks (and the seed's DB witness that the descriptions ARE indexed
+blocks (and the seed's DB witness that the descriptions ARE indexed)
-but NOT displayed) are exact only against the mock's deterministic
+are exact only against the mock's deterministic
-``PICTORIAL_PAGE_MODE`` canned reply.
+``PICTORIAL_PAGE_MODE`` canned reply. (Phase 163 re-cut: the viewer
+assertions moved from the phase-148 suppression contract to the
+ADDITIVE contract — the pictorial class shows its text INCLUDING the
+page descriptions, with the strip behind the Text/Pages toggle.)
 
 Test → observable mapping (Playwright Mapping Rule):
 1. ``test_pictorial_pdf_shows_pages_in_modal`` — the seed (module)
    asserts the DB + filesystem first: the pictorial row has
    ``page_count == 3`` and the scratch page dir holds EXACTLY
    ``<id>-p{1..3}.png`` (real PNG magic, non-empty — the textheavy
-   doc's copies do not exist); then the Sources-tree doc modal shows
+   doc's copies do not exist); then the Sources-tree doc modal opens
-   the ``.doc-pages`` strip: exactly three ``.doc-page`` figures,
+   the ADDITIVE view (phase 163 — a usable text layer: TEXT first,
-   each a LOADED lazy ``<img>`` (``complete && naturalWidth > 0`` —
+   the strip one toggle away): the Text/Pages group with ``Text``
-   the route served a real PNG, not a 404) with the
+   pressed, the Summary panel + content visible (the mock's page
-   ``Page N of <title>`` alt, the ``Page N`` caption, the NO
+   descriptions RIDE the shown text — LOCKED A1/A6), the
-   ``.doc-raw`` / NO ``.doc-md`` / NO ``.doc-summary`` suppression
+   ``.doc-pages`` strip mounted but hidden; pressing ``Pages`` shows
-   contract (the mock's page descriptions sit in the DB content but
+   exactly three ``.doc-page`` figures, each a LOADED lazy ``<img>``
-   nowhere in the rendered modal), the meta badge row (source /
+   (``complete && naturalWidth > 0`` — the route served a real PNG,
-   Created / Indexed) untouched.
+   not a 404) with the ``Page N of <title>`` alt and the ``Page N``
+   caption while the text view is hidden (mounted for the flip-
+   back), the meta badge row (source / Created / Indexed) untouched.
 2. ``test_textheavy_pdf_unchanged`` — the textheavy modal renders the
    PRE-PHASE shape: the content in ``pre.doc-raw`` (both pages), the
    Summary panel (the mock's lite digest), NO ``.doc-pages`` — and
    the content payload's ``page_urls`` KEY is ABSENT (never ``null`` —
    the omission pin the viewer's strip gate rides on).
-3. ``test_full_page_surface`` — the same strip on the dedicated
+3. ``test_full_page_surface`` — the same additive view + strip on the
-   ``/document.html`` page (admin-logged-in, the phase-79 gate
+   dedicated ``/document.html`` page (admin-logged-in, the phase-79
-   satisfied): one ``renderDocument`` core — no per-surface drift —
+   gate satisfied): one ``renderDocument`` core — no per-surface
-   same figures, same suppression.
+   drift — same toggle, same figures, same hidden-while-Text shape.
 4. ``test_page_route_serves_png`` — an ANONYMOUS (no-login) direct
    request on ``/api/documents/<id>/pages/1``: 200, ``image/png``,
    the ``Cache-Control: private, max-age=3600`` header, a non-empty
@@ -597,15 +602,25 @@ def _wait_img_loaded(img: Locator) -> None:
 def _assert_page_strip(
     content: Locator, title: str, doc_id: str, page_count: int = 3
 ) -> None:
-    """The phase-148 strip contract inside a content container (the
+    """The strip contract inside a content container (the ONE
-    ONE ``renderDocument`` core — modal and full page, no
+    ``renderDocument`` core — modal and full page, no per-surface
-    per-surface drift): a ``.doc-pages`` strip with exactly
+    drift): a ``.doc-pages`` strip with exactly *page_count*
-    *page_count* ``.doc-page`` figures; each figure's lazy ``<img>``
+    ``.doc-page`` figures; each figure's lazy ``<img>`` is LOADED (the
-    is LOADED (the route served a real PNG) with the ``Page N of
+    route served a real PNG) with the ``Page N of <title>`` alt and
-    <title>`` alt and the exact 1-based route src, and the
+    the exact 1-based route src, and the figcaption reads ``Page N``.
-    figcaption reads ``Page N``; the text content block AND the
+
-    Summary panel are SUPPRESSED (the descriptions stay in the DB,
+    Phase 163 (task 04) re-cut: this fixture is the PICTORIAL class —
-    not displayed)."""
+    an ADDITIVE doc (``page_strip_mode == "additive"``) that opens on
+    its TEXT view, so the strip is revealed by pressing the toggle's
+    ``Pages`` button first (a primary/absent doc has no toggle — the
+    strip already IS the content). The text content block and the
+    Summary panel are HIDDEN while the strip shows (they stay mounted
+    for the flip back — LOCKED A1); a primary-class doc has no such
+    nodes at all, which the same ``not_to_be_visible`` assertion
+    covers."""
+    toggle = content.locator(".doc-view-toggle")
+    if toggle.count() > 0:
+        toggle.get_by_role("button", name="Pages").click()
     strip = content.locator(".doc-pages")
     expect(strip).to_have_count(1)
     figures = strip.locator(".doc-page")
@@ -624,12 +639,12 @@ def _assert_page_strip(
         )
         _wait_img_loaded(img)
         expect(fig.locator("figcaption.doc-page-num")).to_have_text(f"Page {n}")
-    # The suppression contract (the owner's TODO L5): no text content
+    # While the strip shows, the text view is not (phase 163: hidden,
-    # block, no Summary panel — the meta badge row is asserted by the
+    # not unmounted — the additive flip-back; a primary doc has no
-    # caller (it renders exactly as today).
+    # such nodes at all). The meta badge row is asserted by the caller.
-    expect(content.locator("pre.doc-raw")).to_have_count(0)
+    expect(content.locator("pre.doc-raw")).not_to_be_visible()
-    expect(content.locator(".doc-md")).to_have_count(0)
+    expect(content.locator(".doc-md")).not_to_be_visible()
-    expect(content.locator(".doc-summary")).to_have_count(0)
+    expect(content.locator(".doc-summary")).not_to_be_visible()
 
 
 # ---------------------------------------------------------------------------
@@ -640,15 +655,18 @@ def _assert_page_strip(
 def test_pictorial_pdf_shows_pages_in_modal(
     page: Page, app_url: str, seeded: dict[str, Any]
 ) -> None:
-    """The IKEA-shaped PDF opens in the doc modal as a strip of its
+    """The IKEA-shaped PDF opens in the doc modal as an ADDITIVE doc
-    REAL pages (the owner's L5 ask): ``.doc-pages`` with exactly three
+    (phase 163 — a usable text layer: the pictorial class): the TEXT
+    view first (Summary panel + content visible, the toggle's ``Text``
+    pressed — LOCKED A6, the mock's page descriptions ride the shown
+    text), with the strip of its REAL pages (the owner's L5 ask) one
+    ``Pages`` click away: ``.doc-pages`` with exactly three
     ``.doc-page`` figures — each a loaded lazy ``<img>`` (naturalWidth
     > 0 proves the full pipeline: sync → render → persist → route →
     viewer) with the ``Page N of <title>`` alt and the ``Page N``
-    caption — and the suppression contract: NO text content
+    caption — while the text view is hidden (mounted for the
-    (``.doc-raw``), NO Summary panel (the mock's page descriptions are
+    flip-back); the meta badge row (source / Created / Indexed)
-    in the DB content, nowhere in the rendered modal); the meta badge
+    renders as today."""
-    row (source / Created / Indexed) renders as today."""
     login(page, app_url)  # lands on /sources.html
     _drill_to_source(page, SOURCE_NAME)
     expect(page.locator("#docs-tbody tr")).to_have_count(2)
@@ -658,6 +676,29 @@ def test_pictorial_pdf_shows_pages_in_modal(
     )
 
     content = page.locator("#doc-modal-content")
+    # Phase 163 (task 04, LOCKED A1/A6): the pictorial class is an
+    # ADDITIVE doc — it opens on its TEXT view: Summary panel +
+    # content visible, the toggle's Text pressed, the strip mounted
+    # but hidden (its lazy imgs never fetched while hidden).
+    toggle = content.locator(".doc-view-toggle")
+    expect(toggle).to_have_count(1)
+    expect(toggle.get_by_role("button", name="Text")).to_have_attribute(
+        "aria-pressed", "true"
+    )
+    expect(toggle.get_by_role("button", name="Pages")).to_have_attribute(
+        "aria-pressed", "false"
+    )
+    expect(content.locator("pre.doc-raw")).to_be_visible()
+    expect(content.locator(".doc-summary")).to_be_visible()
+    expect(content.locator(".doc-pages")).to_be_hidden()
+    # The text view carries the indexed content — the phase-140 page
+    # descriptions ride the SHOWN text for an additive doc (the text
+    # is never hidden: A1).
+    assert PICTORIAL_PAGE_ANSWER in content.inner_text(), (
+        "the additive text view shows the indexed content including "
+        "the described sparse pages"
+    )
+    # …and the strip is one Pages click away (the helper flips).
     _assert_page_strip(content, seeded["title"], seeded["pictorial_id"])
 
     # The meta badge row is untouched by the strip (the pre-phase row).
@@ -666,13 +707,6 @@ def test_pictorial_pdf_shows_pages_in_modal(
     expect(meta.locator(".doc-created")).to_contain_text("Created")
     expect(meta.locator(".doc-indexed")).to_contain_text("Indexed")
 
-    # The suppression witness: the mock's canned page descriptions sit
-    # in the DB (the seed asserted it) but are NOT displayed.
-    assert PICTORIAL_PAGE_ANSWER not in content.inner_text(), (
-        "the described pages' text is in the DB for Q&A — it must not "
-        "be shown to the user (the owner's contract)"
-    )
-
 
 # ---------------------------------------------------------------------------
 # 2. The text-heavy PDF: byte-identical to pre-phase
@@ -730,10 +764,11 @@ def test_textheavy_pdf_unchanged(
 
 def test_full_page_surface(page: Page, app_url: str, seeded: dict[str, Any]) -> None:
     """The dedicated ``/document.html`` page (admin-logged-in — the
-    phase-79 gate satisfied) renders the SAME strip through the ONE
+    phase-79 gate satisfied) renders the SAME additive view through
-    shared ``renderDocument`` core: the same three loaded figures with
+    the ONE shared ``renderDocument`` core (no per-surface drift): the
-    captions/alt texts, the same suppression (no text content, no
+    text view first, the strip one ``Pages`` click away with the same
-    Summary panel), the meta row as today."""
+    three loaded figures + captions/alt texts and the text view
+    hidden while it shows, the meta row as today."""
     login(page, app_url, next="/")
     page.goto(
         f"{app_url}/document.html?source={quote(SOURCE_NAME)}"

tests/e2e/test_pictorial_pdf.py

@@ -24,11 +24,14 @@ format — the upload gate admits it) and arriving through the REAL admin
 upload flow → the RAG page's Sync button:
 
 * (a) the document appears on the Sources page and the document viewer
-  shows the PDF's ACTUAL pages (phase 148 — a described sparse page
+  opens the PDF in its ADDITIVE view (phase 163 — a described sparse
-  makes the doc page-QUALIFYING: the page strip REPLACES the text
+  page makes the doc page-QUALIFYING, and a usable text layer makes
-  content AND the Summary panel, the owner's TODO L5 contract): two
+  the strip ADDITIVE: the text stays, the pages sit behind a
-  ``.doc-page`` figures with loaded lazy images + ``Page N`` captions,
+  Text/Pages toggle — LOCKED A1/A6): the Text view first (``.doc-raw``
-  NO ``.doc-raw``, NO ``.doc-summary``; the described content itself
+  + ``.doc-summary`` visible, ``Text`` pressed), the strip one
+  ``Pages`` click away — two ``.doc-page`` figures with loaded lazy
+  images + ``Page N`` captions, the text view hidden while it shows;
+  the described content itself
   (page 1's prose WITHOUT an illustration block, page 2 WITH
   ``[Page 2 illustration: <the mock's canned PICTORIAL_PAGE_ANSWER>]``
   — the task-02 ``PICTORIAL_PAGE_MODE`` branch, deterministic) stays in
@@ -83,14 +86,17 @@ Test → observable mapping (Playwright Mapping Rule):
    Sources page reads 1 doc / 3 chunks (2 page-anchored content chunks +
    the summary chunk — no image doc), the drill-down table lists the PDF
    as a plain text row (LOCKED A4 — a text document, no thumbnail), and
-   the same-page document viewer (the row's path link) shows the
+   the same-page document viewer (the row's path link) opens the
-   PHASE-148 PAGE STRIP under the METADATA title (the fixture is
+   PHASE-163 ADDITIVE VIEW under the METADATA title (the fixture is
-   page-qualifying — the pictorial pass described its sparse page):
+   page-qualifying with a usable text layer — the pictorial pass
-   exactly two ``.doc-page`` figures, each a LOADED lazy ``<img>``
+   described its sparse page): the TEXT view first — ``.doc-raw`` +
-   (naturalWidth > 0) with the ``Page N of <title>`` alt + the
+   ``.doc-summary`` visible with the ``[Page 2 illustration: …]``
-   ``Page N`` caption, NO ``.doc-raw`` / NO ``.doc-summary`` (the
+   block among them, the toggle's ``Text`` pressed, the strip mounted
-   page descriptions stay in the DB for Q&A, not displayed — the
+   but hidden; pressing ``Pages`` shows exactly two ``.doc-page``
-   owner's contract).
+   figures, each a LOADED lazy ``<img>`` (naturalWidth > 0) with the
+   ``Page N of <title>`` alt + the ``Page N`` caption while the text
+   view hides (mounted for the flip-back — the text is never
+   destroyed).
 2. ``test_chat_question_grounds_and_echoes_the_illustration`` — the raw
    SSE over the scripted seeded-read question (httpx — the ``done``
    event is the wire lens the task pins; the UI flow's premise is the
@@ -578,8 +584,8 @@ def seeded(
             )
             # The non-markdown doc got its lite summary (the mock's
             # SUMMARY_MODE digest) — the summary chunk exists (the
-            # viewer's Summary panel is suppressed by the phase-148
+            # additive view shows the Summary panel; the summary
-            # strip — the summary itself stays in the DB).
+            # itself stays in the DB regardless).
             assert doc.summary
             # Phase 148: the described sparse page makes this doc
             # page-QUALIFYING — the strip's payload (page_urls) rides
@@ -602,18 +608,19 @@ def test_sources_page_and_viewer_show_the_described_page(
     content chunks + the phase-30 summary chunk — no image doc; the
     vector figure is invisible to the embedded-image pipeline). The
     drill-down table lists the PDF as a plain text row (LOCKED A4), and
-    the same-page document viewer (the row's path link, phase 26) shows
+    the same-page document viewer (the row's path link, phase 26)
-    the PHASE-148 PAGE STRIP under the METADATA title: the fixture is
+    opens the PHASE-163 ADDITIVE VIEW (LOCKED A1/A6): the TEXT view
-    page-qualifying (the pictorial pass described its sparse page), so
+    first — Summary panel + ``.doc-raw`` visible, the Text/Pages
-    the strip REPLACES the text content AND the Summary panel — two
+    toggle's ``Text`` pressed — with the ``[Page 2 illustration:
-    ``.doc-page`` figures, each a loaded lazy ``<img>`` with the
+    <canned>]`` block riding the SHOWN text (the text is never hidden
-    ``Page N of <title>`` alt + the ``Page N`` caption, NO
+    for the pictorial class). Pressing ``Pages`` reveals the strip —
-    ``.doc-raw``, NO ``.doc-summary``. The described content (page 1's
+    two ``.doc-page`` figures, each a loaded lazy ``<img>`` with the
-    prose WITHOUT an illustration block, page 2 WITH
+    ``Page N of <title>`` alt + the ``Page N`` caption — while the
-    ``[Page 2 illustration: <canned>]`` — LOCKED A4) stays in the DB
+    text view is hidden (mounted, not removed: flipping back shows it
-    for Q&A: the seed asserted it byte-exact, the chat test proves it
+    with zero refetch). The described content (page 1's prose WITHOUT
-    reaches the model — it is simply not displayed (the owner's
+    an illustration block, page 2 WITH the block — LOCKED A4) also
-    contract)."""
+    stays in the DB for Q&A: the seed asserted it byte-exact, the chat
+    test proves it reaches the model."""
     login(page, app_url)  # lands on /sources.html
     expect(page.locator("#stat-docs")).to_have_text("1")
     expect(page.locator("#stat-chunks")).to_have_text("3")
@@ -630,13 +637,31 @@ def test_sources_page_and_viewer_show_the_described_page(
     expect(modal).to_be_visible(timeout=30_000)
     expect(page.locator("#doc-modal-title")).to_have_text(PDF_TITLE)
 
-    # Phase 148: the page-qualifying fixture shows its ACTUAL pages —
+    # Phase 163 (task 04, LOCKED A1/A6): the page-qualifying fixture
-    # a .doc-pages strip with exactly two .doc-page figures (the
+    # with a usable text layer is an ADDITIVE doc — it opens on its
-    # 2-page fixture), each a LOADED lazy <img> (naturalWidth > 0 —
+    # TEXT view: Summary panel + content visible, Text pressed, the
-    # the route served a real PNG) with the "Page N of <title>" alt +
+    # strip mounted but hidden (its lazy imgs never fetched).
-    # the "Page N" caption. The text content AND the Summary panel are
+    toggle = modal.locator(".doc-view-toggle")
-    # SUPPRESSED (the owner's contract — the descriptions stay in the
+    expect(toggle).to_have_count(1)
-    # DB for Q&A, not displayed).
+    expect(toggle.get_by_role("button", name="Text")).to_have_attribute(
+        "aria-pressed", "true"
+    )
+    expect(toggle.get_by_role("button", name="Pages")).to_have_attribute(
+        "aria-pressed", "false"
+    )
+    expect(modal.locator("pre.doc-raw")).to_be_visible()
+    expect(modal.locator(".doc-summary")).to_be_visible()
+    expect(modal.locator(".doc-pages")).to_be_hidden()
+    # The canned page description RIDES the shown text (phase-140
+    # blocks are part of the content — the additive view never hides
+    # the text). The node stays in the DOM, so a visibility-scoped
+    # inner_text() check would miss it — use the DOM text.
+    assert PICTORIAL_PAGE_ANSWER in modal.evaluate("el => el.textContent"), (
+        "the additive text view shows the indexed content including "
+        "the described sparse page"
+    )
+    # …and the strip is one Pages click away (phase 148's ask):
+    toggle.get_by_role("button", name="Pages").click()
     strip = modal.locator(".doc-pages")
     expect(strip).to_have_count(1)
     figures = strip.locator(".doc-page")
@@ -657,17 +682,12 @@ def test_sources_page_and_viewer_show_the_described_page(
         expect(figures.nth(i).locator("figcaption.doc-page-num")).to_have_text(
             f"Page {n}"
         )
-    # The suppression contract: no text content block, no Summary
+    # While the strip shows, the text view is HIDDEN — mounted for
-    # panel, no image block (the PDF is not an image document).
+    # the flip back (LOCKED A1), never removed. No image block (the
-    expect(modal.locator("pre.doc-raw")).to_have_count(0)
+    # PDF is not an image document).
-    expect(modal.locator(".doc-summary")).to_have_count(0)
+    expect(modal.locator("pre.doc-raw")).not_to_be_visible()
+    expect(modal.locator(".doc-summary")).not_to_be_visible()
     expect(modal.locator(".doc-image-img")).to_have_count(0)
-    # The canned description is in the DB (the seed asserted it)
-    # but nowhere in the rendered modal.
-    assert PICTORIAL_PAGE_ANSWER not in modal.inner_text(), (
-        "the described page's text is in the DB for Q&A — it must not "
-        "be shown to the user (the owner's contract)"
-    )
 
 
 # ---------------------------------------------------------------------------

tests/integration/test_docs_api.py

@@ -10,7 +10,10 @@ per phase 124 A4) and the phase-148 (task 02) PDF page route
 (``GET /api/documents/{id}/pages/{n}`` — the same trust model and 404
 posture for a page-qualifying document's rendered page PNGs) + the
 content endpoint's ``image_url`` / ``page_urls`` keys (present for the
-qualifying row, ABSENT otherwise — never ``null``).
+qualifying row, ABSENT otherwise — never ``null``), plus the phase-163
+(task 01) ``page_strip_mode`` class signal (``"primary"`` /``"additive"``
+for a stamped strip, ABSENT for a pre-phase row and for every
+non-qualifying document).
 
 Uses the real compose Postgres (``db`` fixture) and FastAPI's TestClient.
 """
@@ -1381,7 +1384,8 @@ def test_content_endpoint_exposes_the_image_affordance(
 
 # ---------------------------------------------------------------------------
 # Phase 148 (task 02) — the PDF page BYTES route + the content payload's
-# ``page_urls`` affordance
+# ``page_urls`` affordance (phase 163 task 01 adds the ``page_strip_mode``
+# class signal beside it)
 # ---------------------------------------------------------------------------
 
 
@@ -1404,6 +1408,7 @@ def _seed_page_doc(
     path: str = "besta.pdf",
     page_count: int | None = 2,
     write_copies: bool = True,
+    page_strip_mode: str | None = None,
 ) -> Document:
     """One PDF document row with ``page_count`` + its rendered page
     copies (the importer's ``pdf_page_dir`` layout —
@@ -1411,7 +1416,8 @@ def _seed_page_doc(
     sentinel bytes, so a page mix-up between the two copies is caught.
     ``page_count=None`` is a non-qualifying PDF (no copies);
     ``write_copies=False`` leaves a qualifying row without its files
-    (the lost-copy corner)."""
+    (the lost-copy corner); ``page_strip_mode`` (phase 163) stamps the
+    strip CLASS — ``None`` is the pre-phase row."""
     doc = Document(
         id=uuid.uuid4(),
         source=source,
@@ -1424,6 +1430,7 @@ def _seed_page_doc(
         created_at=datetime(2026, 1, 1, tzinfo=UTC),
         summary="IKEA BESTA frame assembly manual (page descriptions).",
         page_count=page_count,
+        page_strip_mode=page_strip_mode,
     )
     db.add(doc)
     db.flush()
@@ -1610,6 +1617,79 @@ def test_content_endpoint_exposes_the_page_urls_affordance(
         )
         assert r.status_code == 200
         assert "page_urls" not in r.json()  # absent — never null (non-qualifying PDF)
+        # Phase 163 (task 01): the class key follows the same rule — a
+        # pre-phase row (mode NULL) omits it while keeping page_urls.
+        assert "page_strip_mode" not in r.json()
+    finally:
+        _cleanup_kb(db)
+
+
+@pytest.mark.parametrize("mode", ["primary", "additive"])
+def test_content_endpoint_carries_the_strip_class(
+    admin_client: TestClient, db, tmp_path: Path, mode: str
+) -> None:
+    """Phase 163 (task 01, LOCKED A2/A3) — the real-DB round-trip of the
+    strip CLASS: a stamped row (``page_strip_mode`` = ``"primary"`` — the
+    phase-133 fallback class, the strip IS the view — or ``"additive"`` —
+    the phase-140 pictorial class, text + a Text/Pages toggle) serves the
+    class VERBATIM beside the untouched ``page_urls`` list (the viewer's
+    branch signal, task 04); every other key stands where phase 148 left
+    it."""
+    _cleanup_kb(db)
+    try:
+        pages_dir = tmp_path / "pages"
+        _seed_page_doc(db, pages_dir, page_count=3, page_strip_mode=mode)
+        r = admin_client.get(
+            "/api/documents/content", params={"source": "PdfSrc", "path": "besta.pdf"}
+        )
+        assert r.status_code == 200
+        body = r.json()
+        assert body["page_strip_mode"] == mode
+        assert len(body["page_urls"]) == 3  # the strip rides along, unchanged
+    finally:
+        _cleanup_kb(db)
+
+
+def test_content_strip_class_never_travels_without_the_strip(
+    admin_client: TestClient, db, tmp_path: Path
+) -> None:
+    """Phase 163 (task 01) — the two strip keys are ONE block keyed off
+    ``page_count``: a row with a stray ``page_strip_mode`` and no strip
+    (``page_count`` NULL — the corner a partial repair could leave)
+    serializes NEITHER key, byte-identically to a plain non-qualifying
+    document (the viewer must never be handed a class for pages that do
+    not exist)."""
+    _cleanup_kb(db)
+    try:
+        pages_dir = tmp_path / "pages"
+        _seed_page_doc(
+            db,
+            pages_dir,
+            path="stray.pdf",
+            page_count=None,
+            write_copies=False,
+            page_strip_mode="additive",
+        )
+        r = admin_client.get(
+            "/api/documents/content", params={"source": "PdfSrc", "path": "stray.pdf"}
+        )
+        assert r.status_code == 200
+        body = r.json()
+        assert "page_strip_mode" not in body
+        assert "page_urls" not in body
+        # The exact pre-phase key set (the byte-identity pin).
+        assert set(body) == {
+            "source",
+            "path",
+            "title",
+            "format",
+            "summary",
+            "created_at",
+            "content",
+            "indexed_at",
+            "chunks",
+            "is_image",
+        }
     finally:
         _cleanup_kb(db)
 

tests/integration/test_migration_0027.py

@@ -0,0 +1,370 @@
+"""Integration: migration 0027 (documents.page_strip_mode) schema contract
+(phase 163, task 01).
+
+Drives the **real Alembic engine** against the live dev database
+(``podman compose up -d db``), mirroring the house pattern of
+``test_migration_0026.py`` (information_schema assertions on the state
+the migration must leave). The tests target the 0026 → 0027 step
+explicitly so later migrations cannot break the pins:
+
+* upgrade 0026 → 0027 → ``page_strip_mode`` exists with the full
+  contract — VARCHAR(16), NULLABLE, NO server default (NULL = "no strip"
+  for every pre-phase row, every non-qualifying document and every
+  fail-soft miss) and UNINDEXED (a display fact, never a query surface —
+  the content endpoint reads it from the ROW, the ``file_hash`` /
+  ``page_count`` convention) — while the 0026 ``documents`` schema
+  survives, in particular ``page_count`` keeping its 0026 contract and
+  ``content_hash`` staying NOT NULL and INDEXED;
+* rows inserted while the DB is at 0026 backfill ``page_strip_mode`` to
+  NULL — including a row whose ``page_count`` IS set (the pre-phase strip
+  with no stamped class: the wire omits the key and the viewer keeps
+  today's "primary" DOM until the task-03 re-gate stamps it);
+* the ORM contract agrees: a freshly inserted ``Document`` with no mode
+  reads ``page_strip_mode is None`` through a fresh session, and each of
+  the two classes (``"primary"`` / ``"additive"``) round-trips verbatim;
+* downgrade to 0026 → the column is GONE (A13) while the row + its
+  content + its ``page_count`` survive; upgrade back to 0027 → the column
+  is back (round-trip).
+
+The ``alembic`` fixture guarantees the DB ends at head even if a test
+fails or the process is interrupted.
+"""
+from __future__ import annotations
+
+import hashlib
+import uuid
+from collections.abc import Iterator
+from typing import Any
+
+import pytest
+from alembic.config import Config
+from sqlalchemy import text
+from sqlalchemy.orm import Session
+
+from alembic import command
+from app.db import SessionLocal, db_available
+from app.models import Document
+
+SOURCE = "mig0027"
+CONTENT = "content here"
+CONTENT_HASH = hashlib.sha256(CONTENT.encode()).hexdigest()
+
+
+@pytest.fixture()
+def alembic(db: Session) -> Iterator[Config]:
+    """Real Alembic config bound to the dev DB (URL from app settings).
+
+    Starts at head (repairs an interrupted earlier run); teardown
+    upgrades to head no matter what happened, so the dev DB is never
+    left below head.
+    """
+    if not db_available():
+        pytest.skip("Postgres not reachable — run `podman compose up -d db` first")
+    cfg = Config()  # no alembic.ini file — env.py gets the URL from app config
+    cfg.set_main_option("script_location", "alembic")
+    command.upgrade(cfg, "head")
+    try:
+        yield cfg
+    finally:
+        # Release the test session's open transaction BEFORE the repair
+        # DDL: an idle-in-transaction SELECT holds an ACCESS SHARE lock
+        # on ``documents``, which would deadlock the repair's
+        # ``ALTER TABLE`` forever.
+        db.rollback()
+        command.upgrade(cfg, "head")
+
+
+def _version(db: Session) -> str | None:
+    return db.execute(text("SELECT version_num FROM alembic_version")).scalar()
+
+
+def _column(db: Session, column: str) -> tuple[Any, ...] | None:
+    """(data_type, is_nullable, column_default, character_maximum_length)
+    for one documents column."""
+    row = db.execute(
+        text(
+            "SELECT data_type, is_nullable, column_default, character_maximum_length"
+            " FROM information_schema.columns"
+            " WHERE table_name = 'documents' AND column_name = :c"
+        ),
+        {"c": column},
+    ).fetchone()
+    return tuple(row) if row is not None else None
+
+
+def _index_covering(db: Session, column: str) -> str | None:
+    """The name of any documents index whose definition mentions the
+    column (None = UNINDEXED)."""
+    return db.execute(
+        text(
+            "SELECT indexname FROM pg_indexes"
+            " WHERE tablename = 'documents' AND indexdef ILIKE :like"
+        ),
+        {"like": f"%{column}%"},
+    ).scalar()
+
+
+def _insert_sql(db: Session, path: str, page_count: int | None = None) -> uuid.UUID:
+    """Insert one documents row with the PRE-0027 column set (the 0026
+    shape — the mode column, when present, is omitted so a NULL backfill
+    is what the row reads; ``page_count`` is 0026's own column and may be
+    set to model a pre-phase strip)."""
+    doc_id = uuid.uuid4()
+    db.execute(
+        text(
+            "INSERT INTO documents (id, source, path, full_path, title, content,"
+            " content_hash, page_count)"
+            " VALUES (:id, :s, :p, :fp, :t, :c, :h, :pc)"
+        ),
+        {
+            "id": doc_id,
+            "s": SOURCE,
+            "p": path,
+            "fp": f"/tmp/{path}",
+            "t": f"Doc {path}",
+            "c": CONTENT,
+            "h": CONTENT_HASH,
+            "pc": page_count,
+        },
+    )
+    db.commit()
+    return doc_id
+
+
+def _delete(db: Session, doc_id: uuid.UUID) -> None:
+    db.execute(text("DELETE FROM documents WHERE id = :id"), {"id": doc_id})
+    db.commit()
+
+
+def _assert_0026_schema_survives(db: Session) -> None:
+    """The 0026 documents contract must survive the additive upgrade —
+    ``page_count`` keeps its INTEGER / NULLABLE / NO-default / UNINDEXED
+    shape (the sibling column this one rides) and ``content_hash`` stays
+    NOT NULL and indexed (the retriever's identity)."""
+    page_count = _column(db, "page_count")
+    assert page_count is not None, "documents.page_count (0026) must survive the upgrade"
+    assert page_count[0] == "integer", "page_count must stay INTEGER"
+    assert page_count[1] == "YES", "page_count must stay NULLABLE"
+    assert page_count[2] is None, "page_count must keep NO server default"
+    assert _index_covering(db, "page_count") is None, "page_count must stay UNINDEXED"
+
+    content = _column(db, "content_hash")
+    assert content is not None and content[0] == "character varying" and content[1] == "NO", (
+        "documents.content_hash (0001) must keep its contract"
+    )
+    assert (
+        _index_covering(db, "content_hash") is not None
+    ), "the ix_documents_content_hash index must survive the upgrade"
+
+    indexed = _column(db, "indexed_at")
+    assert indexed is not None and indexed[0] == "timestamp with time zone", (
+        "documents.indexed_at (0001) must survive the upgrade"
+    )
+    is_image = _column(db, "is_image")
+    assert is_image is not None and is_image[0] == "boolean" and is_image[1] == "NO", (
+        "documents.is_image (0022) must survive the upgrade"
+    )
+
+
+def test_upgrade_to_0027_adds_page_strip_mode(db: Session, alembic: Config) -> None:
+    """Upgrade 0026 → 0027: ``page_strip_mode`` exists with the full
+    contract (VARCHAR(16), NULLABLE, NO server default — NULL = "no
+    strip" — and UNINDEXED: a display fact, never a query surface), is
+    ABSENT at 0026, and every pre-0027 row backfills it to NULL —
+    including the row whose ``page_count`` IS set (the pre-phase strip
+    with no stamped class), while the 0026 table contract (the
+    ``page_count`` sibling, the indexed ``content_hash``) survives.
+    """
+    command.downgrade(alembic, "0026")  # start from the pre-0027 state
+    assert _version(db) == "0026"
+    assert _column(db, "page_strip_mode") is None, "page_strip_mode must be absent at 0026"
+
+    plain_id = _insert_sql(db, "plain.pdf")  # the 0026 column set, no strip
+    pre_strip_id = _insert_sql(db, "prephase.pdf", page_count=36)  # a live pre-phase strip
+
+    try:
+        command.upgrade(alembic, "0027")
+        assert _version(db) == "0027", "alembic_version must be at 0027"
+
+        col = _column(db, "page_strip_mode")
+        assert col is not None, "documents.page_strip_mode is missing"
+        assert col[0] == "character varying", "page_strip_mode must be VARCHAR"
+        assert col[3] == 16, "page_strip_mode must be VARCHAR(16) — the two class names"
+        assert col[1] == "YES", (
+            "page_strip_mode must be NULLABLE (NULL = no strip, and the pre-phase shape)"
+        )
+        assert col[2] is None, (
+            "page_strip_mode must carry NO server default — NULL is the meaningful"
+            " pre-phase state"
+        )
+        assert _index_covering(db, "page_strip_mode") is None, (
+            "page_strip_mode must be UNINDEXED — a display fact, never a query"
+            " surface (the content endpoint reads it from the row)"
+        )
+
+        # Every pre-0027 row backfills to NULL — no backfill, no default.
+        assert (
+            db.execute(
+                text("SELECT page_count, page_strip_mode FROM documents WHERE id = :id"),
+                {"id": plain_id},
+            ).fetchone()
+            == (None, None)
+        ), "a pre-0027 row without a strip must read both columns NULL"
+        assert (
+            db.execute(
+                text("SELECT page_count, page_strip_mode FROM documents WHERE id = :id"),
+                {"id": pre_strip_id},
+            ).fetchone()
+            == (36, None)
+        ), (
+            "a pre-phase strip (page_count set) must backfill page_strip_mode to"
+            " NULL — the wire omits the key and the viewer keeps today's DOM"
+        )
+
+        _assert_0026_schema_survives(db)
+    finally:
+        _delete(db, plain_id)
+        _delete(db, pre_strip_id)
+
+
+@pytest.mark.parametrize("mode", ["primary", "additive"])
+def test_orm_page_strip_mode_round_trips(db: Session, alembic: Config, mode: str) -> None:
+    """The ORM contract agrees with the column contract: each class
+    (``"primary"`` = the fallback class, ``"additive"`` = the pictorial
+    class) round-trips VERBATIM through a fresh session, and a row
+    created with no mode reads ``page_strip_mode is None`` (the
+    pre-phase shape — NULL until the importer stamps it)."""
+    command.upgrade(alembic, "head")
+    stamped = Document(
+        source=SOURCE,
+        path="stamped.pdf",
+        full_path="/tmp/stamped.pdf",
+        title="stamped",
+        content=CONTENT,
+        content_hash=CONTENT_HASH,
+        page_count=36,
+        page_strip_mode=mode,
+    )
+    unstamped = Document(
+        source=SOURCE,
+        path="unstamped.md",
+        full_path="/tmp/unstamped.md",
+        title="unstamped",
+        content=CONTENT,
+        content_hash=CONTENT_HASH,
+    )
+    db.add(stamped)
+    db.add(unstamped)
+    db.commit()
+    try:
+        with SessionLocal() as fresh:
+            reloaded = fresh.get(Document, stamped.id)
+            assert reloaded is not None, "the stamped row must be readable"
+            assert reloaded.page_strip_mode == mode, (
+                f"the {mode!r} class must round-trip through the DB verbatim"
+            )
+            assert reloaded.page_count == 36, "the sibling page_count must round-trip too"
+            reloaded_unstamped = fresh.get(Document, unstamped.id)
+            assert reloaded_unstamped is not None, "the unstamped row must be readable"
+            assert reloaded_unstamped.page_strip_mode is None, (
+                "a row created without a mode must read NULL (the pre-phase shape)"
+            )
+            assert reloaded_unstamped.content_hash == CONTENT_HASH, (
+                "content_hash must be untouched by the page_strip_mode addition"
+            )
+    finally:
+        _delete(db, stamped.id)
+        _delete(db, unstamped.id)
+
+
+def test_downgrade_to_0026_drops_the_column(db: Session, alembic: Config) -> None:
+    """Downgrade 0027 → 0026: the page_strip_mode column is gone (A13 —
+    fully reversible, the class is re-derivable from the source file by
+    the deterministic re-gate) while the row + its content + its 0026
+    ``page_count`` survive, and the rest of the table contract is intact."""
+    command.upgrade(alembic, "head")
+    row = Document(
+        source=SOURCE,
+        path="down.pdf",
+        full_path="/tmp/down.pdf",
+        title="down",
+        content=CONTENT,
+        content_hash=CONTENT_HASH,
+        page_count=36,
+        page_strip_mode="additive",
+    )
+    db.add(row)
+    db.commit()
+    try:
+        command.downgrade(alembic, "0026")
+        assert _version(db) == "0026"
+        assert _column(db, "page_strip_mode") is None, "page_strip_mode must be dropped"
+
+        surviving = db.execute(
+            text(
+                "SELECT source, path, content, content_hash, page_count, is_image"
+                " FROM documents WHERE id = :id"
+            ),
+            {"id": row.id},
+        ).fetchone()
+        assert surviving is not None and surviving[0] == SOURCE, (
+            "the row must survive the column drop"
+        )
+        assert surviving[1] == "down.pdf" and surviving[2] == CONTENT, (
+            "path + content must survive the drop"
+        )
+        assert surviving[3] == CONTENT_HASH, (
+            "content_hash (the retriever's identity) must survive the drop"
+        )
+        assert surviving[4] == 36, (
+            "page_count (the 0026 sibling) must survive the drop — the strip"
+            " itself is untouched by this migration's reversal"
+        )
+        assert surviving[5] is False, "is_image must survive the drop"
+        assert _index_covering(db, "content_hash") is not None, (
+            "the content_hash index must survive the downgrade"
+        )
+    finally:
+        _delete(db, row.id)
+    # Repair: the fixture teardown re-upgrades to head.
+
+
+def test_upgrade_round_trip_restores_the_column(db: Session, alembic: Config) -> None:
+    """Downgrade to 0026, then upgrade back to 0027: ``page_strip_mode``
+    is back with the full contract (VARCHAR(16), NULLABLE, NO server
+    default, UNINDEXED), a row seeded at 0026 — strip included — reads
+    the mode as NULL (the pre-phase shape on the round-trip), and the
+    0026 schema (the ``page_count`` sibling, the indexed
+    ``content_hash``) is intact."""
+    command.downgrade(alembic, "0026")
+    pre_id = _insert_sql(db, "round.pdf", page_count=36)
+    try:
+        command.upgrade(alembic, "0027")
+        assert _version(db) == "0027", "round-trip upgrade must land at 0027"
+
+        col = _column(db, "page_strip_mode")
+        assert col is not None, "documents.page_strip_mode must be back"
+        assert col[0] == "character varying", "page_strip_mode must be VARCHAR after the round-trip"
+        assert col[3] == 16, "page_strip_mode must be VARCHAR(16) after the round-trip"
+        assert col[1] == "YES", "page_strip_mode must be NULLABLE after the round-trip"
+        assert col[2] is None, (
+            "page_strip_mode must still carry NO server default after the round-trip"
+        )
+        assert _index_covering(db, "page_strip_mode") is None, (
+            "page_strip_mode must stay UNINDEXED after the round-trip"
+        )
+
+        assert (
+            db.execute(
+                text("SELECT page_count, page_strip_mode FROM documents WHERE id = :id"),
+                {"id": pre_id},
+            ).fetchone()
+            == (36, None)
+        ), (
+            "a row seeded at 0026 must read page_strip_mode NULL after the"
+            " round-trip while its page_count stands"
+        )
+
+        _assert_0026_schema_survives(db)
+    finally:
+        _delete(db, pre_id)

tests/unit/test_document_viewer.py

@@ -30,6 +30,17 @@ page-qualifying PDF row (``page_count`` set) adds ``page_urls`` = the
 exact 1-based ``/api/documents/<id>/pages/<i>`` list over
 ``1..page_count``.
 
+Phase 163 (task 01) — the content payload's STRIP CLASS signal
+(``page_strip_mode``: ``"primary"`` = the fallback class, the strip IS
+the view; ``"additive"`` = the pictorial class, text + a Text/Pages
+toggle): the model-level omission rule (ABSENT when ``None``, never
+``null``), and the endpoint (stubbed session) — the key rides
+``page_urls`` for a qualifying row, a PRE-PHASE row (``page_count`` set,
+mode NULL — the exact live ``Letumdei`` state) serializes it ABSENT while
+keeping ``page_urls``, and the whole strip block keys off
+``page_count``, so a row with a stray mode and no strip carries neither
+key (every non-qualifying payload byte-identical to pre-phase).
+
 Phase 145 (task 03) — the content payload's git origin fields + the
 ``renderDocument`` "Original" anchor (modal AND /document.html, one
 site):
@@ -316,10 +327,11 @@ def _text_doc_content(
     origin_url: str | None = None,
     origin_branch: str | None = None,
     page_urls: list[str] | None = None,
+    page_strip_mode: str | None = None,
 ) -> DocContent:
     """One text doc's content payload (the pre-phase shape) with the
-    phase-145 origin fields + the phase-148 ``page_urls`` optionally
+    phase-145 origin fields + the phase-148 ``page_urls`` + the
-    set."""
+    phase-163 ``page_strip_mode`` optionally set."""
     return DocContent(
         source="S",
         path="a.md",
@@ -332,6 +344,7 @@ def _text_doc_content(
         origin_url=origin_url,
         origin_branch=origin_branch,
         page_urls=page_urls,
+        page_strip_mode=page_strip_mode,
     )
 
 
@@ -367,9 +380,29 @@ def test_doc_content_page_urls_absent_on_the_wire_when_unset() -> None:
     assert dumped["page_urls"] == urls  # the exact 1-based route list
 
 
-def _page_pdf_doc() -> Document:
+def test_doc_content_page_strip_mode_absent_on_the_wire_when_unset() -> None:
+    """Phase 163 (task 01, LOCKED A2) — the model-level omission rule
+    (the ``image_url`` / ``page_urls`` precedent): no strip class
+    (``page_strip_mode`` ``None`` — every non-qualifying doc AND every
+    pre-phase row) serializes the key ABSENT — never ``null`` — so those
+    payloads are byte-identical to pre-phase. A stamped row carries the
+    class verbatim, beside ``page_urls``."""
+    dumped = _text_doc_content().model_dump()
+    assert "page_strip_mode" not in dumped  # absent — never null
+    for mode in ("primary", "additive"):
+        dumped = _text_doc_content(
+            page_urls=["/api/documents/d/pages/1"], page_strip_mode=mode
+        ).model_dump()
+        assert dumped["page_strip_mode"] == mode  # the class verbatim
+
+
+def _page_pdf_doc(
+    page_count: int | None = 2, page_strip_mode: str | None = None
+) -> Document:
     """A page-QUALIFYING PDF row (phase 148): ``page_count`` set — the
-    endpoint derives ``page_urls`` from it (the exact 1-based list)."""
+    endpoint derives ``page_urls`` from it (the exact 1-based list).
+    ``page_strip_mode`` (phase 163) is the strip CLASS; ``None`` is the
+    PRE-PHASE row shape (the column NULL while the strip stands)."""
     doc = Document(
         id=uuid.uuid4(),
         source="Homelab",
@@ -379,7 +412,8 @@ def _page_pdf_doc() -> Document:
         content="[Page 1 illustration: the parts tray]…",
         content_hash="e" * 64,
         is_image=False,  # phase 122: the stub session never applies defaults
-        page_count=2,
+        page_count=page_count,
+        page_strip_mode=page_strip_mode,
     )
     doc.indexed_at = datetime(2026, 8, 22, 1, 2, 3, tzinfo=UTC)
     doc.created_at = datetime(2026, 8, 20, 9, 0, 0, tzinfo=UTC)
@@ -410,6 +444,79 @@ database): a page-qualifying PDF row (``page_count`` set) carries
     ]  # the exact 1-based list, count == page_count
     assert body["is_image"] is False
     assert "image_url" not in body  # absent — never null (not an image doc)
+    # Phase 163 (task 01): the pre-phase row (mode NULL) keeps the strip
+    # and serializes the class ABSENT — the viewer treats that as
+    # "primary", i.e. today's DOM.
+    assert "page_strip_mode" not in body
+
+
+@pytest.mark.parametrize("mode", ["primary", "additive"])
+def test_content_page_strip_mode_carried_for_a_stamped_strip(mode: str) -> None:
+    """Phase 163 (task 01, LOCKED A2/A3) — the endpoint (stubbed
+session, no
+database): a page-qualifying PDF whose row carries the class
+    (``"primary"`` = the fallback class, ``"additive"`` = the pictorial
+    class) carries ``page_strip_mode`` VERBATIM next to ``page_urls`` —
+    the signal the viewer branches on so an additive strip never hides
+    the text (task 04). Nothing else in the payload moves."""
+    doc = _page_pdf_doc(page_strip_mode=mode)
+    with _client_with_row((doc, 4)) as client:
+        r = client.get(
+            "/api/documents/content",
+            params={"source": "Homelab", "path": "ikea/besta.pdf"},
+        )
+    assert r.status_code == 200
+    body = r.json()
+    assert body["page_strip_mode"] == mode  # the class verbatim
+    assert body["page_urls"] == [
+        f"/api/documents/{doc.id}/pages/1",
+        f"/api/documents/{doc.id}/pages/2",
+    ]  # the strip rides along, unchanged
+
+
+def test_content_strip_block_keys_off_page_count_never_mode_alone() -> None:
+    """Phase 163 (task 01) — the class never appears alone: the whole
+    strip block keys off ``page_count``, so a row whose column somehow
+    holds a mode while ``page_count`` is NULL serializes NEITHER
+    ``page_strip_mode`` nor ``page_urls`` — the payload is byte-identical
+    to a plain non-qualifying document (there is no strip to show, and
+    the viewer must never be told a class for pages that do not exist)."""
+    doc = _page_pdf_doc(page_count=None, page_strip_mode="additive")
+    with _client_with_row((doc, 1)) as client:
+        r = client.get(
+            "/api/documents/content",
+            params={"source": "Homelab", "path": "ikea/besta.pdf"},
+        )
+    assert r.status_code == 200
+    body = r.json()
+    assert "page_strip_mode" not in body  # absent — the strip block is off
+    assert "page_urls" not in body  # absent — never null
+
+
+def test_content_non_qualifying_payload_key_set_is_byte_identical() -> None:
+    """Phase 163 (task 01) — the byte-identity pin: a non-qualifying PDF
+    row (``page_count`` NULL, mode NULL) carries EXACTLY the pre-phase
+    key set — neither strip key joins it (``is_image`` always present,
+    ``image_url`` / the origin keys / both strip keys absent)."""
+    doc = _page_pdf_doc(page_count=None)
+    with _client_with_row((doc, 1)) as client:
+        r = client.get(
+            "/api/documents/content",
+            params={"source": "Homelab", "path": "ikea/besta.pdf"},
+        )
+    assert r.status_code == 200
+    assert set(r.json()) == {
+        "source",
+        "path",
+        "title",
+        "format",
+        "summary",
+        "created_at",
+        "content",
+        "indexed_at",
+        "chunks",
+        "is_image",
+    }
 
 
 def test_content_origin_url_carried_for_a_matching_git_row() -> None:

tests/unit/test_importer_unchanged_fast_path.py

@@ -31,6 +31,16 @@ the second sync):
 * the streamed hash is transparent (a >1 MiB file hashes identically
   to ``hashlib.sha256(bytes).hexdigest()``).
 
+Phase 163 (task 03) extends this file with the PDF strip RE-GATE the
+unchanged paths now run (``_regate_pdf_strip``) — the pre-phase "the
+copies persist" assumption is what broke in production (a row claiming
+``page_count = 36`` over ZERO files, every page 404ing forever): the
+re-gate re-derives the strip from the current bytes + settings (ZERO
+vision/LLM calls), heals missing copies, re-stamps a changed class,
+clears a strip that no longer qualifies (kill switch included), stays
+fail-soft, logs only when it ACTS, and costs a non-strip file zero PDF
+parsing (spy-pinned).
+
 The E2E surface (the idempotent second sync through the real UI +
 mock LLM) is pinned in ``tests/e2e/test_sync_changed_only.py``
 (task 03).
@@ -43,6 +53,7 @@ import logging
 import os
 from datetime import UTC, datetime, timedelta
 from pathlib import Path
+from typing import Any
 
 import pymupdf as fitz
 import pytest
@@ -50,6 +61,7 @@ from sqlalchemy import select
 
 from app.config import Settings
 from app.models import Document
+from app.rag import importer
 from app.rag.importer import _file_hash, import_sources
 from tests.fakes import FakeEmbedder
 from tests.unit.test_importer import (
@@ -60,6 +72,7 @@ from tests.unit.test_importer import (
     _VisionLLM,
 )
 from tests.unit.test_pictorial_import import (
+    _build_dense_pdf,
     _build_ikea_like_pdf,
     _PictorialLLM,
     _settings,
@@ -454,6 +467,520 @@ def test_pre_phase_null_file_hash_image_runs_branch_and_stamps(
         _cleanup_source(db, root.name)
 
 
+# ---------- phase 163 (task 03): the deterministic PDF strip re-gate ----------
+#
+# The live defect (owner report, 2026-10-07): the `Letumdei` row carried
+# `page_count = 36` over ZERO files on disk (the pre-phase-163 persist
+# deleted the old copies BEFORE rendering, outside the DB transaction —
+# a later per-file rollback restored the stale count over an empty
+# directory; an external wipe of `pdf_page_dir` lands the same state).
+# The phase-147 fast path believed "the copies persist" and never
+# re-rendered, so the strip 404'd forever. `_regate_pdf_strip` now runs
+# at the tail of `_unchanged_doc`: the changed-path gate mirrored from
+# the current bytes + settings (ZERO vision/LLM calls), missing copies
+# healed, a stale strip cleared.
+
+#: The PNG signature (a healed copy is a real render, not a placeholder).
+PNG_MAGIC = b"\x89PNG\r\n\x1a\n"
+
+
+def _regate_settings(tmp_path: Path, **kwargs: Any) -> Settings:
+    """Settings for the re-gate matrix: the ``pdf`` token admitted, the
+    phase-140 knobs at their CODE DEFAULTS (``pdf_pictorial`` ON), and
+    the strip's home ``pdf_page_dir`` pinned under *tmp_path* (never the
+    real home)."""
+    kwargs.setdefault("_env_file", None)
+    kwargs.setdefault("import_extensions", "md,txt,pdf")
+    kwargs.setdefault("images", False)
+    kwargs.setdefault("image_dir", str(tmp_path / "images"))
+    kwargs.setdefault("pdf_page_dir", str(tmp_path / "pages"))
+    return Settings(**kwargs)  # pyright: ignore[reportCallIssue]
+
+
+def _pages_dir(tmp_path: Path) -> Path:
+    return tmp_path / "pages"
+
+
+def _strip_copies(doc: Document, page_dir: Path, count: int) -> list[Path]:
+    return [page_dir / f"{doc.id}-p{n}.png" for n in range(1, count + 1)]
+
+
+def test_regate_heals_deleted_copies_of_a_pictorial_strip(
+    db, tmp_path: Path, caplog: pytest.LogCaptureFixture
+) -> None:
+    """THE live-loss reproduction (heal arm): a pictorial-class PDF is
+    indexed (``page_count`` set, copies written), its copies are deleted
+    out from under the row (the exact live state — mode NULL on a
+    pre-phase row), and an UNCHANGED re-sync restores every file with
+    real PNG bytes, stamps the mode ``additive``, keeps ``page_count``
+    (no drift), burns ZERO llm/vision calls, and logs the one greppable
+    INFO line. The NULL-mode stamp is SILENT (no ``mode`` segment — it
+    rides the line only when work followed)."""
+    root = tmp_path / "rgheal"
+    root.mkdir()
+    _build_ikea_like_pdf(root / "ikea.pdf")
+    llm = _PictorialLLM()
+    llm.settings = _regate_settings(tmp_path)
+    try:
+        asyncio.run(import_sources([root], llm, session=db))
+        doc = _doc(db, "rgheal", "ikea.pdf")
+        assert (doc.page_count, doc.page_strip_mode) == (10, "additive")
+        page_dir = _pages_dir(tmp_path)
+        doc.page_strip_mode = None  # the pre-phase row (the live shape)
+        db.commit()
+        for path in _strip_copies(doc, page_dir, 10):
+            path.unlink()
+        assert not list(page_dir.glob(f"{doc.id}-*.png"))
+        embeds1, chats1, vision1 = len(llm.calls), len(llm.chat_calls), llm.pictorial_calls
+
+        with caplog.at_level(logging.INFO, logger="app.importer"):
+            second = asyncio.run(import_sources([root], llm, session=db))
+
+        assert (second.added, second.updated, second.unchanged) == (0, 0, 1)
+        # ZERO llm work of any kind (the re-gate is deterministic, A5).
+        assert len(llm.calls) == embeds1, "the re-gate never re-embeds"
+        assert len(llm.chat_calls) == chats1, "the re-gate never chats"
+        assert llm.pictorial_calls == vision1, "the re-gate issues ZERO vision calls"
+
+        db.expire_all()
+        doc2 = _doc(db, "rgheal", "ikea.pdf")
+        assert doc2.page_count == 10, "no drift — the count is untouched"
+        assert doc2.page_strip_mode == "additive", "the mode is stamped"
+        for path in _strip_copies(doc2, page_dir, 10):
+            assert path.is_file(), f"copy {path.name} was not healed"
+            assert path.read_bytes()[:8] == PNG_MAGIC, f"{path.name} is not a PNG"
+        assert not list(page_dir.glob("*.tmp")), "no .tmp orphan from the heal"
+        messages = [r.getMessage() for r in caplog.records]
+        assert (
+            "import: PDF strip re-gate source=rgheal path=ikea.pdf — "
+            "healed 10 pages"
+        ) in messages, messages
+        assert not any("mode " in m for m in messages if "re-gate" in m), (
+            "a NULL->stamp mode re-stamp is silent-beneficial " + str(messages)
+        )
+    finally:
+        _cleanup_source(db, "rgheal")
+
+
+def test_regate_heals_a_fallback_strip_as_primary(
+    db, tmp_path: Path, caplog: pytest.LogCaptureFixture
+) -> None:
+    """Heal arm, the FALLBACK class (phase 133's unusable text layer):
+    the copies of a garbage-layer PDF's strip are restored on an
+    unchanged re-sync and the mode lands ``primary`` (the strip IS the
+    view) — and the phase-133 fallback vision does NOT re-fire (the
+    class is re-derived from the bytes, no vision call)."""
+    root = tmp_path / "rgfall"
+    root.mkdir()
+    _build_garbage_pdf(root / "garbage.pdf")  # 2 pages
+    llm = _PageVisionLLM()
+    llm.settings = _regate_settings(tmp_path)
+    try:
+        asyncio.run(import_sources([root], llm, session=db))
+        doc = _doc(db, "rgfall", "garbage.pdf")
+        assert (doc.page_count, doc.page_strip_mode) == (2, "primary")
+        page_dir = _pages_dir(tmp_path)
+        for path in _strip_copies(doc, page_dir, 2):
+            path.unlink()
+        vision1, chats1, embeds1 = llm.vision_calls, len(llm.chat_calls), len(llm.calls)
+
+        with caplog.at_level(logging.INFO, logger="app.importer"):
+            second = asyncio.run(import_sources([root], llm, session=db))
+
+        assert (second.added, second.updated, second.unchanged) == (0, 0, 1)
+        assert second.pdfs_screenshotted == 0, "the fallback gate did not re-run"
+        assert llm.vision_calls == vision1, "ZERO vision calls on the re-gate"
+        assert len(llm.chat_calls) == chats1 and len(llm.calls) == embeds1
+
+        db.expire_all()
+        doc2 = _doc(db, "rgfall", "garbage.pdf")
+        assert doc2.page_strip_mode == "primary", "the unusable layer is primary"
+        assert doc2.page_count == 2
+        for path in _strip_copies(doc2, page_dir, 2):
+            assert path.is_file() and path.read_bytes()[:8] == PNG_MAGIC
+        assert any(
+            "healed 2 pages" in m and "re-gate" in m for m in caplog.messages
+        ), caplog.messages
+    finally:
+        _cleanup_source(db, "rgfall")
+
+
+def test_regate_noop_when_the_strip_is_healthy(
+    db, tmp_path: Path, caplog: pytest.LogCaptureFixture
+) -> None:
+    """No-op arm (the IKEA-class doc whose copies all stand): an
+    unchanged re-sync writes NOTHING (every copy's mtime and bytes are
+    untouched), renders NOTHING, and logs NO re-gate line — the INFO is
+    never per-unchanged-doc — while the pre-existing ``import:
+    unchanged`` line still fires."""
+    root = tmp_path / "rgnoop"
+    root.mkdir()
+    _build_ikea_like_pdf(root / "ikea.pdf")
+    llm = _PictorialLLM()
+    llm.settings = _regate_settings(tmp_path)
+    try:
+        asyncio.run(import_sources([root], llm, session=db))
+        doc = _doc(db, "rgnoop", "ikea.pdf")
+        page_dir = _pages_dir(tmp_path)
+        before = {
+            p: (p.stat().st_mtime_ns, p.read_bytes())
+            for p in _strip_copies(doc, page_dir, 10)
+        }
+        assert len(before) == 10
+
+        with caplog.at_level(logging.INFO, logger="app.importer"):
+            second = asyncio.run(import_sources([root], llm, session=db))
+
+        assert (second.added, second.updated, second.unchanged) == (0, 0, 1)
+        db.expire_all()
+        doc2 = _doc(db, "rgnoop", "ikea.pdf")
+        assert (doc2.page_count, doc2.page_strip_mode) == (10, "additive")
+        after = {
+            p: (p.stat().st_mtime_ns, p.read_bytes())
+            for p in _strip_copies(doc2, page_dir, 10)
+        }
+        assert after == before, "a healthy strip is not written at all"
+        messages = [r.getMessage() for r in caplog.records]
+        assert any(m.startswith("import: unchanged ") for m in messages), messages
+        assert not any("re-gate" in m for m in messages), (
+            f"the re-gate logged without acting: {messages}"
+        )
+    finally:
+        _cleanup_source(db, "rgnoop")
+
+
+def test_regate_restamps_a_changed_mode(
+    db, tmp_path: Path, caplog: pytest.LogCaptureFixture
+) -> None:
+    """Mode re-stamp (the class flip): a row whose stamped mode no
+    longer matches the gate's class (here a pictorial-class row carrying
+    ``primary`` — a legacy/mis-stamped row) is re-stamped to
+    ``additive``, and because the mode was ALREADY SET the change is
+    reported (``mode primary->additive``) — the only thing the re-gate
+    did, so that is the whole INFO line (no copies were missing, none
+    were written)."""
+    root = tmp_path / "rgmode"
+    root.mkdir()
+    _build_ikea_like_pdf(root / "ikea.pdf")
+    llm = _PictorialLLM()
+    llm.settings = _regate_settings(tmp_path)
+    try:
+        asyncio.run(import_sources([root], llm, session=db))
+        doc = _doc(db, "rgmode", "ikea.pdf")
+        assert doc.page_strip_mode == "additive"
+        doc.page_strip_mode = "primary"  # the mis-stamp the re-gate fixes
+        db.commit()
+        page_dir = _pages_dir(tmp_path)
+        before = {
+            p: p.stat().st_mtime_ns for p in _strip_copies(doc, page_dir, 10)
+        }
+
+        with caplog.at_level(logging.INFO, logger="app.importer"):
+            second = asyncio.run(import_sources([root], llm, session=db))
+
+        assert (second.added, second.unchanged) == (0, 1)
+        db.expire_all()
+        doc2 = _doc(db, "rgmode", "ikea.pdf")
+        assert doc2.page_strip_mode == "additive", "the class is re-stamped"
+        assert doc2.page_count == 10
+        assert {
+            p: p.stat().st_mtime_ns for p in _strip_copies(doc2, page_dir, 10)
+        } == before, "a mode fix writes no files"
+        assert (
+            "import: PDF strip re-gate source=rgmode path=ikea.pdf — "
+            "mode primary->additive"
+        ) in [r.getMessage() for r in caplog.records], caplog.messages
+    finally:
+        _cleanup_source(db, "rgmode")
+
+
+def test_regate_clears_a_strip_that_no_longer_qualifies(
+    db, tmp_path: Path, caplog: pytest.LogCaptureFixture
+) -> None:
+    """REPAIR arm: a row claims a strip (``page_count`` set, copies on
+    disk) on a PDF whose GATE now says ineligible (a clean, dense text
+    layer — zero sparse pages). The unchanged re-sync clears BOTH
+    columns and deletes the copies (the strip must not linger over a
+    document the viewer shows as text) + the ``cleared stale strip``
+    INFO line, with zero llm calls."""
+    root = tmp_path / "rgclear"
+    root.mkdir()
+    pdf = root / "clean.pdf"
+    _build_dense_pdf(pdf)  # usable layer, zero candidates — never qualifies
+    llm = _PictorialLLM()
+    llm.settings = _regate_settings(tmp_path)
+    try:
+        asyncio.run(import_sources([root], llm, session=db))
+        doc = _doc(db, "rgclear", "clean.pdf")
+        assert (doc.page_count, doc.page_strip_mode) == (None, None)
+        # The stale state the re-gate must repair (a strip stamped by an
+        # older gate/settings, copies still on disk).
+        page_dir = _pages_dir(tmp_path)
+        page_dir.mkdir(parents=True, exist_ok=True)
+        for path in _strip_copies(doc, page_dir, 2):
+            path.write_bytes(b"STALE-COPY")
+        doc.page_count = 2
+        doc.page_strip_mode = "additive"
+        db.commit()
+        embeds1, chats1 = len(llm.calls), len(llm.chat_calls)
+
+        with caplog.at_level(logging.INFO, logger="app.importer"):
+            second = asyncio.run(import_sources([root], llm, session=db))
+
+        assert (second.added, second.updated, second.unchanged) == (0, 0, 1)
+        assert len(llm.calls) == embeds1 and len(llm.chat_calls) == chats1
+
+        db.expire_all()
+        doc2 = _doc(db, "rgclear", "clean.pdf")
+        assert doc2.page_count is None, "the stale strip is cleared"
+        assert doc2.page_strip_mode is None, "the class is cleared with it"
+        assert not list(page_dir.glob(f"{doc2.id}-p*.png")), "the copies are deleted"
+        assert (
+            "import: PDF strip re-gate source=rgclear path=clean.pdf — "
+            "cleared stale strip"
+        ) in [r.getMessage() for r in caplog.records], caplog.messages
+    finally:
+        _cleanup_source(db, "rgclear")
+
+
+def test_regate_kill_switch_clears_the_pictorial_strip(
+    db, tmp_path: Path, caplog: pytest.LogCaptureFixture
+) -> None:
+    """REPAIR arm via the kill switch (``pdf_pictorial=False``): the
+    changed-path semantics mirrored — with the pass off a sparse-page
+    PDF is NOT eligible, so its existing strip is cleared with its
+    copies on the next (unchanged) sync. The bytes never moved: the
+    setting alone decided."""
+    root = tmp_path / "rgkill"
+    root.mkdir()
+    _build_ikea_like_pdf(root / "ikea.pdf")
+    llm = _PictorialLLM()
+    llm.settings = _regate_settings(tmp_path)
+    try:
+        asyncio.run(import_sources([root], llm, session=db))
+        doc = _doc(db, "rgkill", "ikea.pdf")
+        assert (doc.page_count, doc.page_strip_mode) == (10, "additive")
+        page_dir = _pages_dir(tmp_path)
+        assert len(list(page_dir.glob(f"{doc.id}-p*.png"))) == 10
+        vision1 = llm.pictorial_calls
+
+        llm.settings = _regate_settings(tmp_path, pdf_pictorial=False)
+        with caplog.at_level(logging.INFO, logger="app.importer"):
+            second = asyncio.run(import_sources([root], llm, session=db))
+
+        assert (second.added, second.updated, second.unchanged) == (0, 0, 1)
+        assert llm.pictorial_calls == vision1, "ZERO vision calls (the pass is off)"
+        db.expire_all()
+        doc2 = _doc(db, "rgkill", "ikea.pdf")
+        assert doc2.page_count is None and doc2.page_strip_mode is None
+        assert not list(page_dir.glob(f"{doc2.id}-p*.png")), "the copies went too"
+        assert any(
+            "cleared stale strip" in m for m in caplog.messages
+        ), caplog.messages
+    finally:
+        _cleanup_source(db, "rgkill")
+
+
+def test_regate_render_exception_is_fail_soft(
+    db,
+    tmp_path: Path,
+    monkeypatch: pytest.MonkeyPatch,
+    caplog: pytest.LogCaptureFixture,
+) -> None:
+    """Fail-soft: the heal's render RAISES — the sync completes
+    (``unchanged``, zero errors), the row keeps its prior state
+    (``page_count`` + mode untouched — the re-gate never clears a strip
+    on a failure), no copies are touched, and the WARNING is the only
+    signal."""
+    root = tmp_path / "rgboom"
+    root.mkdir()
+    _build_ikea_like_pdf(root / "ikea.pdf")
+    llm = _PictorialLLM()
+    llm.settings = _regate_settings(tmp_path)
+    try:
+        asyncio.run(import_sources([root], llm, session=db))
+        doc = _doc(db, "rgboom", "ikea.pdf")
+        page_dir = _pages_dir(tmp_path)
+        for path in _strip_copies(doc, page_dir, 10):
+            path.unlink()
+
+        def boom(*args: object, **kwargs: object) -> list[bytes]:
+            raise RuntimeError("simulated render failure (test sentinel)")
+
+        monkeypatch.setattr(importer, "render_pdf_pages", boom)
+        with caplog.at_level(logging.WARNING, logger="app.importer"):
+            second = asyncio.run(import_sources([root], llm, session=db))
+
+        assert (second.added, second.updated, second.unchanged, second.errors) == (
+            0, 0, 1, 0,
+        )
+        db.expire_all()
+        doc2 = _doc(db, "rgboom", "ikea.pdf")
+        assert (doc2.page_count, doc2.page_strip_mode) == (10, "additive"), (
+            "the row keeps its pre-re-gate state"
+        )
+        assert not list(page_dir.glob(f"{doc.id}-p*.png")), "no copies were touched"
+        assert any(
+            "PDF strip re-gate failed" in m and "test sentinel" in m
+            for m in caplog.messages
+        ), caplog.messages
+    finally:
+        _cleanup_source(db, "rgboom")
+
+
+def test_regate_empty_render_keeps_the_row(
+    db,
+    tmp_path: Path,
+    monkeypatch: pytest.MonkeyPatch,
+    caplog: pytest.LogCaptureFixture,
+) -> None:
+    """Fail-soft on the no-pages render (unopenable right after the fast
+    path hashed the file — pathological): the row is kept EXACTLY as it
+    was (never cleared on a transient), nothing is written, and the
+    WARNING names the no-pages state."""
+    root = tmp_path / "rgempty"
+    root.mkdir()
+    _build_ikea_like_pdf(root / "ikea.pdf")
+    llm = _PictorialLLM()
+    llm.settings = _regate_settings(tmp_path)
+    try:
+        asyncio.run(import_sources([root], llm, session=db))
+        doc = _doc(db, "rgempty", "ikea.pdf")
+        page_dir = _pages_dir(tmp_path)
+        for path in _strip_copies(doc, page_dir, 10):
+            path.unlink()
+        monkeypatch.setattr(importer, "render_pdf_pages", lambda *a, **k: [])
+
+        with caplog.at_level(logging.WARNING, logger="app.importer"):
+            second = asyncio.run(import_sources([root], llm, session=db))
+
+        assert (second.added, second.unchanged, second.errors) == (0, 1, 0)
+        db.expire_all()
+        doc2 = _doc(db, "rgempty", "ikea.pdf")
+        assert (doc2.page_count, doc2.page_strip_mode) == (10, "additive")
+        assert not page_dir.exists() or not list(page_dir.glob(f"{doc.id}-*"))
+        assert any(
+            "re-render produced no pages (state kept)" in m
+            for m in caplog.messages
+        ), caplog.messages
+    finally:
+        _cleanup_source(db, "rgempty")
+
+
+def test_regate_reports_a_page_count_drift_on_the_heal(
+    db,
+    tmp_path: Path,
+    monkeypatch: pytest.MonkeyPatch,
+    caplog: pytest.LogCaptureFixture,
+) -> None:
+    """Drift (pathological on unchanged bytes, so it is LOUD): the heal
+    renders 6 pages for a row that claimed 10 — ``page_count`` follows
+    the fresh render, the ``p7..p10`` tail beyond it is deleted, and a
+    WARNING names both counts beside the ``healed`` INFO line."""
+    root = tmp_path / "rgdrift"
+    root.mkdir()
+    _build_ikea_like_pdf(root / "ikea.pdf")
+    llm = _PictorialLLM()
+    llm.settings = _regate_settings(tmp_path)
+    try:
+        asyncio.run(import_sources([root], llm, session=db))
+        doc = _doc(db, "rgdrift", "ikea.pdf")
+        page_dir = _pages_dir(tmp_path)
+        for path in _strip_copies(doc, page_dir, 10):
+            path.unlink()
+        monkeypatch.setattr(
+            importer,
+            "render_pdf_pages",
+            lambda *a, **k: [PNG_MAGIC + f"{i}".encode() for i in range(6)],
+        )
+
+        with caplog.at_level(logging.INFO, logger="app.importer"):
+            second = asyncio.run(import_sources([root], llm, session=db))
+
+        assert (second.added, second.unchanged) == (0, 1)
+        db.expire_all()
+        doc2 = _doc(db, "rgdrift", "ikea.pdf")
+        assert doc2.page_count == 6, "the count follows the fresh render"
+        for n in range(1, 7):
+            assert (page_dir / f"{doc2.id}-p{n}.png").is_file()
+        for n in range(7, 11):
+            assert not (page_dir / f"{doc2.id}-p{n}.png").exists(), f"p{n} leaked"
+        messages = [r.getMessage() for r in caplog.records]
+        assert any("page_count drift 10->6" in m for m in messages), messages
+        assert any("healed 10 pages" in m for m in messages), messages
+    finally:
+        monkeypatch.undo()  # pyright: ignore[reportUnknownMemberType]
+        _cleanup_source(db, "rgdrift")
+
+
+def test_regate_fires_on_the_content_digest_hit_path(
+    db, tmp_path: Path, caplog: pytest.LogCaptureFixture
+) -> None:
+    """The SECOND unchanged path (the PDF branch's content-digest hit —
+    a pre-phase row with ``file_hash`` NULL): the branch re-runs its own
+    gate (the byte-change's honest cost) and lands unchanged, and the
+    shared helper's tail still heals the strip deleted out from under
+    the row — the re-gate is not fast-path-only."""
+    root = tmp_path / "rgbranch"
+    root.mkdir()
+    _build_ikea_like_pdf(root / "ikea.pdf")
+    llm = _PictorialLLM()
+    llm.settings = _regate_settings(tmp_path)
+    try:
+        asyncio.run(import_sources([root], llm, session=db))
+        doc = _doc(db, "rgbranch", "ikea.pdf")
+        doc.file_hash = None  # the pre-phase row: the fast path cannot fire
+        db.commit()
+        page_dir = _pages_dir(tmp_path)
+        for path in _strip_copies(doc, page_dir, 10):
+            path.unlink()
+
+        with caplog.at_level(logging.INFO, logger="app.importer"):
+            second = asyncio.run(import_sources([root], llm, session=db))
+
+        assert (second.added, second.updated, second.unchanged) == (0, 0, 1)
+        db.expire_all()
+        doc2 = _doc(db, "rgbranch", "ikea.pdf")
+        assert doc2.file_hash is not None, "the helper stamped the hash"
+        for path in _strip_copies(doc2, page_dir, 10):
+            assert path.is_file(), f"copy {path.name} was not healed"
+        assert any("healed 10 pages" in m for m in caplog.messages), caplog.messages
+    finally:
+        _cleanup_source(db, "rgbranch")
+
+
+def test_regate_guard_costs_a_non_strip_file_nothing(
+    db, tmp_path: Path, monkeypatch: pytest.MonkeyPatch
+) -> None:
+    """The GUARD (A5): a non-PDF unchanged file and a PDF whose
+    ``page_count`` is NULL reach the tail of ``_unchanged_doc`` and
+    ``extract_pdf_text`` is NEVER called for either — every non-strip
+    file's fast path keeps its zero-extra-read contract (one sync of
+    each class, then the spy covers the second sync only)."""
+    root = tmp_path / "rgguard"
+    root.mkdir()
+    (root / "note.md").write_text("# Note\n\nplain text body\n")
+    _build_dense_pdf(root / "clean.pdf")  # usable layer, no strip
+    llm = _PictorialLLM()
+    llm.settings = _regate_settings(tmp_path)
+    seen: list[str] = []
+    real_extract = importer.extract_pdf_text
+
+    def spy(path: Path) -> str:
+        seen.append(str(path))
+        return real_extract(path)
+
+    try:
+        asyncio.run(import_sources([root], llm, session=db))
+        monkeypatch.setattr(importer, "extract_pdf_text", spy)
+        second = asyncio.run(import_sources([root], llm, session=db))
+        assert (second.added, second.updated, second.unchanged) == (0, 0, 2)
+        assert seen == [], f"the guard let a PDF parse through: {seen}"
+    finally:
+        monkeypatch.undo()  # pyright: ignore[reportUnknownMemberType]
+        _cleanup_source(db, "rgguard")
+
+
 # ---------- the preserved semantics on the fast path --------------------------
 
 

tests/unit/test_models.py

@@ -293,6 +293,82 @@ def test_git_source_python_default_include_hidden_false() -> None:
     assert col.default.arg is False, "the default must resolve to False"
 
 
+def test_documents_page_strip_mode_column_contract() -> None:
+    """Phase 163 (task 01): ``documents.page_strip_mode`` — the CLASS of
+    the page strip a non-NULL ``page_count`` describes (``"primary"`` =
+    the phase-133 fallback class, the strip IS the view; ``"additive"``
+    = the phase-140 pictorial class, text + a Text/Pages toggle):
+    VARCHAR(16), NULLABLE (NULL = no strip — and the pre-phase shape for
+    a row whose strip stands), NO server default, and NO index (the
+    content endpoint reads it from the ROW — a display fact, never a
+    query surface, the ``file_hash`` / ``page_count`` convention).
+    ``page_count`` (0026) stays exactly as phase 148 left it."""
+    from sqlalchemy import String as SAString
+
+    documents = Base.metadata.tables["documents"]
+    assert "page_strip_mode" in documents.c, (
+        "documents must have the page_strip_mode column (phase 163)"
+    )
+    col = documents.c["page_strip_mode"]
+    assert col.nullable is True, (
+        "page_strip_mode must be NULLABLE — NULL means no strip (and the"
+        " pre-phase row whose strip stands)"
+    )
+    assert isinstance(col.type, SAString), "page_strip_mode must be a String column"
+    assert col.type.length == 16, (
+        "page_strip_mode must be String(16) — the two class names, bounded"
+    )
+    assert col.server_default is None, (
+        "page_strip_mode must carry NO server default — NULL is the"
+        " meaningful pre-phase state"
+    )
+    assert not col.index, (
+        "page_strip_mode must be UNINDEXED — the content endpoint reads it"
+        " from the row, no query filters on it"
+    )
+    covering = [
+        idx
+        for idx in documents.indexes
+        if any(c.name == "page_strip_mode" for c in idx.columns)
+    ]
+    assert not covering, "no table-level index may cover page_strip_mode"
+
+    # page_count is untouched: still the nullable, unindexed sibling.
+    page_count = documents.c["page_count"]
+    assert page_count.nullable is True, "page_count must stay NULLABLE (0026)"
+    assert not page_count.index, "page_count must stay UNINDEXED (0026)"
+
+
+def test_document_page_strip_mode_orm_defaults_to_none() -> None:
+    """A freshly constructed ``Document`` (no DB) exposes
+    ``page_strip_mode``: an omitted kwarg and an explicit ``None`` both
+    read NULL (no strip — the shape every non-qualifying document keeps,
+    and the shape a pre-phase-163 row keeps WHILE its ``page_count``
+    stands, which the wire serializes ABSENT), and each class name
+    round-trips verbatim as a plain string (the real DB round-trip is
+    pinned in ``test_migration_0027.py``)."""
+    from app.models import Document
+
+    base = dict(
+        source="s",
+        path="p.pdf",
+        full_path="/tmp/s/p.pdf",
+        title="p",
+        content="c",
+        content_hash="1" * 64,
+        page_count=36,
+    )
+    for mode in ("primary", "additive"):
+        assert Document(**base, page_strip_mode=mode).page_strip_mode == mode
+
+    assert Document(**base, page_strip_mode=None).page_strip_mode is None
+    omitted = Document(**base)
+    assert omitted.page_strip_mode is None, (
+        "an omitted page_strip_mode must read NULL while page_count stands"
+        " (the pre-phase strip the viewer keeps showing as 'primary')"
+    )
+
+
 def test_doc_drafts_column_contract() -> None:
     """Phase 59: the editable triple (title/path/body) + status +
     timestamps are NOT NULL; ``branch`` / ``commit_sha`` are NULL

tests/unit/test_pdf_page_copies.py

@@ -21,14 +21,20 @@ builder for the negative; the copies land in a tmp ``pdf_page_dir``):
   layer, ≥1 sparse page) and the fallback class (unusable layer) get
   ``<doc-id>-p{N}.png`` for EVERY page (not just the described
   candidates — the strip shows the whole manual) + ``page_count`` =
-  the page count + real PNG bytes; a text-heavy PDF gets neither
+  the page count + real PNG bytes + the phase-163 CLASS stamp
-  (no copies, NULL count, ZERO render calls, byte-identical content);
+  (``page_strip_mode`` ``additive`` / ``primary``); a text-heavy PDF
+  gets neither (no copies, NULL count, NULL mode, ZERO render calls,
+  byte-identical content);
 * the phase-147 fast path — an unchanged qualifying PDF's second sync
   re-renders NOTHING (zero render calls — the copies persist,
   ``page_count`` and the bytes untouched);
-* stale replacement — a changed (shrunk) PDF's old copy set is
+* WRITE-FIRST replacement (phase 163, task 02, LOCKED A4) — a changed
-  replaced: the new set covers the new pages, the old ``p{N}`` tail
+  (shrunk) PDF's new copies land atomically (tmp + ``os.replace``,
-  is GONE (no orphans), ``page_count`` follows the new count;
+  pinned) BEFORE any old copy is removed; the old ``p{N}`` tail is
+  GONE only after the replacement bytes are durable (no orphans),
+  ``page_count`` follows the new count; a MID-WRITE failure leaves the
+  prior copy set intact (the rollback-orphan state — the row over
+  zero files — is structurally unrepresentable);
 * qualifier loss — a qualifying PDF whose new version gained a dense
   text layer loses its strip: ``page_count`` NULL + copies deleted;
 * unopenable — a render that yields no pages (``render_pdf_pages`` →
@@ -40,9 +46,10 @@ builder for the negative; the copies land in a tmp ``pdf_page_dir``):
   the FULL count (the gap is a display-level 404 — task 02 pins the
   route's side);
 * fail-soft — the render raising (warning is the signal, ``page_count``
-  NULL, no copies, the doc indexed fully — chunks + summary) and the
+  + ``page_strip_mode`` NULL, no copies, the doc indexed fully — chunks
-  WRITER raising (the copies this call wrote are removed — no partial
+  + summary) and the WRITER raising (this call's ``.tmp`` files and the
-  orphans, ``page_count`` NULL, the doc indexed fully);
+  finals it CREATED are removed — no partial orphans, ``page_count`` +
+  ``page_strip_mode`` NULL, the doc indexed fully);
 * prune — a qualifying doc's file removed from the source: the row
   goes and its page copies go with it (no orphans in the tmp dir).
 
@@ -54,6 +61,7 @@ from __future__ import annotations
 
 import asyncio
 import logging
+import os
 import uuid
 from collections.abc import Callable, Sequence
 from pathlib import Path
@@ -161,6 +169,35 @@ def _render_spy(
     return calls
 
 
+# ---------------------------------------------------------------------------
+# Phase 163 (task 01) — the strip-class column beside page_count
+# ---------------------------------------------------------------------------
+
+
+def test_pictorial_index_stamps_the_additive_class(db, tmp_path: Path) -> None:
+    """Phase 163 (task 02, re-cut of the task-01 pin) — the STAMP: a
+    qualifying PDF indexed through the PICTORIAL class gets
+    ``page_strip_mode == "additive"`` beside ``page_count`` (the text +
+    toggle view); the strip's copies themselves are unchanged by the
+    phase (task 01's shape, mode stamped exactly with the count).
+    """
+    root = tmp_path / "pgm"
+    root.mkdir()
+    _build_ikea_like_pdf(root / "ikea.pdf")
+    llm = _PictorialLLM()
+    llm.settings = _settings(tmp_path)
+    try:
+        summary = asyncio.run(import_sources([root], llm, session=db))
+        assert (summary.files, summary.added, summary.errors) == (1, 1, 0)
+        doc = _doc(db, "pgm", "ikea.pdf")
+        assert doc.page_count == 10, "the strip itself is unchanged by task 02"
+        assert doc.page_strip_mode == "additive", (
+            "the pictorial class stamps the additive mode (LOCKED A1/A3)"
+        )
+    finally:
+        _cleanup_source(db, "pgm")
+
+
 # ---------------------------------------------------------------------------
 # Eligibility × new index
 # ---------------------------------------------------------------------------
@@ -219,6 +256,10 @@ def test_fallback_class_garbage_pdf_writes_every_page_copy(db, tmp_path: Path) -
         assert summary.pdfs_screenshotted == 1  # the fallback fired (the class)
         doc = _doc(db, "pgf", "garbage.pdf")
         assert doc.page_count == 2
+        assert doc.page_strip_mode == "primary", (
+            "the fallback class stamps the primary mode — the strip IS the "
+            "view, byte-identical to phase 148 otherwise (LOCKED A1)"
+        )
         page_dir = _pages_dir(tmp_path)
         for n in (1, 2):
             path = page_dir / f"{doc.id}-p{n}.png"
@@ -253,6 +294,7 @@ def test_text_heavy_pdf_keeps_the_text_view(
         assert summary.pdfs_pictorial == 0  # zero candidates — the pass no-ops
         doc = _doc(db, "pgd", "clean.pdf")
         assert doc.page_count is None, "a text-heavy PDF keeps the text view"
+        assert doc.page_strip_mode is None, "no strip — no class either"
         assert not _pages_dir(tmp_path).exists(), "no copies — the dir is never created"
         assert calls == [], "zero render calls (no fallback, zero candidates)"
         assert doc.content == expected, "byte-identical to the extraction"
@@ -315,9 +357,10 @@ def test_changed_pdf_replaces_the_copy_set_without_orphans(
 ) -> None:
     """A CHANGED (shrunk, 10 → 6 pages) qualifying PDF: the re-index
     replaces the copy set — the new set covers p1..p6, the old
-    version's ``p7..p10`` tail is GONE (the stale-replacement runs
+    version's ``p7..p10`` tail is GONE (the stale-tail delete runs
-    before the write), the doc keeps its id, and ``page_count``
+    LAST, after the write-first loop — phase 163, task 02), the doc
-    follows the new page count."""
+    keeps its id, ``page_count`` follows the new page count, and
+    ``page_strip_mode`` is (re-)stamped."""
     root = tmp_path / "pgc"
     root.mkdir()
     pdf = root / "ikea.pdf"
@@ -336,6 +379,7 @@ def test_changed_pdf_replaces_the_copy_set_without_orphans(
         doc2 = _doc(db, "pgc", "ikea.pdf")
         assert doc2.id == doc1.id, "the doc keeps its id (the copy path is stable)"
         assert doc2.page_count == 6, "page_count follows the new page count"
+        assert doc2.page_strip_mode == "additive", "the class is (re-)stamped"
         page_dir = _pages_dir(tmp_path)
         for n in range(1, 7):
             assert (page_dir / f"{doc2.id}-p{n}.png").is_file(), f"p{n} is missing"
@@ -379,6 +423,7 @@ def test_qualifier_loss_clears_page_count_and_copies(
         doc2 = _doc(db, "pgl", "ikea.pdf")
         assert doc2.id == doc1.id
         assert doc2.page_count is None, "a doc that stopped qualifying loses its strip"
+        assert doc2.page_strip_mode is None, "the class is cleared with the count"
         assert not list(_pages_dir(tmp_path).glob(f"{doc2.id}-p*.png")), (
             "the copies must be deleted on qualifier loss"
         )
@@ -411,6 +456,7 @@ def test_unopenable_pdf_page_render_is_fail_soft(
         assert (summary.files, summary.added, summary.errors) == (1, 1, 0)
         doc = _doc(db, "pgn", "garbage.pdf")
         assert doc.page_count is None, "no pages rendered — no strip"
+        assert doc.page_strip_mode is None, "the class is cleared with the count"
         assert not _pages_dir(tmp_path).exists(), "no copies — the dir is never created"
         assert doc.content == "", "the fallback's all-failed render lands empty content"
         assert summary.summaries == 1, "the document is indexed exactly as pre-phase"
@@ -493,10 +539,11 @@ def test_write_exception_is_fail_soft_and_leaves_no_orphans(
     monkeypatch: pytest.MonkeyPatch,
     caplog: pytest.LogCaptureFixture,
 ) -> None:
-    """The PNG WRITER raises (a full disk) after the FIRST copy
+    """The PNG WRITER raises (a full disk) after the FIRST atomic copy
-    landed: the copies this call wrote are removed (no partial
+    landed: this call's ``.tmp`` files and the final-name files it
-    orphans), ``page_count`` NULL, the warning is the signal, and the
+    CREATED are removed (no partial orphans — phase 163, task 02),
-    document indexes exactly as pre-phase."""
+    ``page_count`` AND ``page_strip_mode`` NULL, the warning is the
+    signal, and the document indexes exactly as pre-phase."""
     root = tmp_path / "pgw"
     root.mkdir()
     _build_ikea_like_pdf(root / "ikea.pdf")
@@ -507,22 +554,29 @@ def test_write_exception_is_fail_soft_and_leaves_no_orphans(
 
     def boom(self: Path, data: bytes) -> int:
         written_paths.append(self)
+        n = original_write(self, data)  # the tmp file really lands first
         if len(written_paths) > 1:
             raise OSError("simulated disk full (test sentinel)")
-        return original_write(self, data)
+        return n
 
     monkeypatch.setattr(Path, "write_bytes", boom)  # pyright: ignore[reportArgumentType]
     try:
         with caplog.at_level(logging.WARNING, logger="app.importer"):
             summary = asyncio.run(import_sources([root], llm, session=db))
         assert (summary.files, summary.added, summary.errors) == (1, 1, 0)
-        assert len(written_paths) == 2, "p1 landed, p2's write failed"
+        assert len(written_paths) == 2, "p1's tmp landed + replaced, p2's write failed"
+        assert all(p.name.endswith(".png.tmp") for p in written_paths), (
+            "the writer only ever touches .tmp paths (atomic replace)"
+        )
         doc = _doc(db, "pgw", "ikea.pdf")
         assert doc.page_count is None
+        assert doc.page_strip_mode is None, "the class is cleared with the count"
         assert not list(_pages_dir(tmp_path).glob(f"{doc.id}-p*.png")), (
             "the partial write was cleaned up — no orphans"
         )
-        assert not written_paths[0].exists(), "the landed p1 copy was removed"
+        assert not list(_pages_dir(tmp_path).glob("*.tmp")), (
+            "the failed call's .tmp file was cleaned up too"
+        )
         assert summary.summaries == 1, "the document is indexed exactly as pre-phase"
         assert any("PDF page render failed" in m for m in caplog.messages), (
             caplog.messages
@@ -531,6 +585,246 @@ def test_write_exception_is_fail_soft_and_leaves_no_orphans(
         _cleanup_source(db, "pgw")
 
 
+# ---------------------------------------------------------------------------
+# Phase 163 (task 02) — WRITE-FIRST, atomic per file (LOCKED A4)
+# ---------------------------------------------------------------------------
+
+
+def test_write_first_new_bytes_are_durable_before_any_stale_delete(
+    db, tmp_path: Path, monkeypatch: pytest.MonkeyPatch
+) -> None:
+    """THE order pin (LOCKED A4 — the 2026-10-07 live loss): re-rendering
+    a CHANGED (shrunk, 10 → 6) PDF writes every replacement page FIRST
+    (tmp + ``os.replace``) and only then deletes the stale tail — the
+    event log shows ALL replaces of the fresh bytes before the FIRST
+    unlink of an old-extent file, the overwritten pages keep the NEW
+    bytes, the shrunk tail (p7..p10) is gone, and nothing else is
+    deleted."""
+    root = tmp_path / "pgord"
+    root.mkdir()
+    pdf = root / "ikea.pdf"
+    _build_ikea_like_pdf(pdf)
+    llm = _PictorialLLM()
+    llm.settings = _settings(tmp_path)
+    asyncio.run(import_sources([root], llm, session=db))
+    doc1 = _doc(db, "pgord", "ikea.pdf")
+    page_dir = _pages_dir(tmp_path)
+    sentinel = b"OLD-COPY-SENTINEL"
+    for n in range(1, 11):  # make the old copies unmistakably distinguishable
+        (page_dir / f"{doc1.id}-p{n}.png").write_bytes(sentinel)
+
+    events: list[tuple[str, str]] = []
+    real_replace = os.replace
+
+    def replace_spy(src, dst, *args, **kwargs):  # pyright: ignore[reportUnknownParameterType]
+        events.append(("replace", Path(os.fspath(dst)).name))
+        return real_replace(src, dst, *args, **kwargs)
+
+    real_unlink = Path.unlink
+
+    def unlink_spy(self: Path, **kwargs: Any) -> None:
+        if self.parent == page_dir and self.name.startswith(str(doc1.id)):
+            events.append(("unlink", self.name))
+        return real_unlink(self, **kwargs)
+
+    monkeypatch.setattr(importer.os, "replace", replace_spy)
+    monkeypatch.setattr(Path, "unlink", unlink_spy)  # pyright: ignore[reportArgumentType]
+    try:
+        _build_ikea_shrunk(pdf)  # same path, 6 pages — still qualifying
+        s2 = asyncio.run(import_sources([root], llm, session=db))
+        assert (s2.added, s2.updated, s2.unchanged) == (0, 1, 0)
+    finally:
+        monkeypatch.undo()  # pyright: ignore[reportUnknownMemberType]
+        _cleanup_source(db, "pgord")
+    replaces = [name for kind, name in events if kind == "replace"]
+    unlinks = [name for kind, name in events if kind == "unlink"]
+    assert len(replaces) == 6, "the 6 fresh pages are written via os.replace"
+    assert set(unlinks) == {f"{doc1.id}-p{n}.png" for n in range(7, 11)}, (
+        "the stale-tail delete touches ONLY the old extent the render did "
+        "not rewrite — a file about to be overwritten is never deleted"
+    )
+    first_unlink = next(i for i, e in enumerate(events) if e[0] == "unlink")
+    last_replace = max(i for i, e in enumerate(events) if e[0] == "replace")
+    assert last_replace < first_unlink, (
+        f"write-first violated — event order: {events}"
+    )
+    for n in range(1, 7):
+        blob = (page_dir / f"{doc1.id}-p{n}.png").read_bytes()
+        assert blob != sentinel and blob[:8] == PNG_MAGIC, (
+            f"p{n} keeps the NEW rendered bytes"
+        )
+    for n in range(7, 11):
+        assert not (page_dir / f"{doc1.id}-p{n}.png").exists(), f"stale p{n} leaked"
+    assert not list(page_dir.glob("*.tmp")), "no .tmp orphans"
+
+
+def test_mid_write_failure_keeps_the_prior_copies_intact(
+    db,
+    tmp_path: Path,
+    monkeypatch: pytest.MonkeyPatch,
+    caplog: pytest.LogCaptureFixture,
+) -> None:
+    """THE failure pin (LOCKED A4): the writer raises on the SECOND
+    page's tmp write of a changed re-render — the prior copy set stays
+    on disk INTACT (the rollback-orphan state — the row over zero
+    files — is structurally unrepresentable): every one of the 10 prior
+    files still exists (p1 with the new bytes already atomically
+    replaced before the failure, p2..p10 byte-for-byte the prior
+    copies), zero ``.tmp`` orphans from this call, and the caller's
+    fail-soft fires (``page_count``/``page_strip_mode`` NULL + the
+    warning, the doc indexed fully)."""
+    root = tmp_path / "pgfail"
+    root.mkdir()
+    pdf = root / "ikea.pdf"
+    _build_ikea_like_pdf(pdf)
+    llm = _PictorialLLM()
+    llm.settings = _settings(tmp_path)
+    asyncio.run(import_sources([root], llm, session=db))
+    doc1 = _doc(db, "pgfail", "ikea.pdf")
+    page_dir = _pages_dir(tmp_path)
+    prior = {
+        n: (page_dir / f"{doc1.id}-p{n}.png").read_bytes() for n in range(1, 11)
+    }
+    # Tag the re-render's bytes so the new p1 is recognizable, then
+    # blow up the SECOND tmp write (page 2).
+    _render_spy(
+        monkeypatch,
+        mutate_full=lambda pages: [
+            PNG_MAGIC + f"NEW-{i}".encode() for i in range(len(pages))
+        ],
+    )
+    original_write = Path.write_bytes
+    tmp_writes: list[Path] = []
+
+    def boom(self: Path, data: bytes) -> int:
+        if self.suffix == ".tmp":
+            tmp_writes.append(self)
+            if len(tmp_writes) > 1:
+                raise OSError("simulated disk full (test sentinel)")
+        return original_write(self, data)
+
+    monkeypatch.setattr(Path, "write_bytes", boom)  # pyright: ignore[reportArgumentType]
+    _build_ikea_shrunk(pdf)
+    try:
+        with caplog.at_level(logging.WARNING, logger="app.importer"):
+            s2 = asyncio.run(import_sources([root], llm, session=db))
+    finally:
+        monkeypatch.undo()  # pyright: ignore[reportUnknownMemberType]
+    assert (s2.added, s2.updated, s2.unchanged) == (0, 1, 0)
+    assert len(tmp_writes) == 2, "p1's tmp wrote, p2's tmp write raised"
+    doc2 = _doc(db, "pgfail", "ikea.pdf")
+    assert doc2.page_count is None, "the caller's fail-soft nulled the count"
+    assert doc2.page_strip_mode is None, "the caller's fail-soft nulled the class"
+    assert any("PDF page render failed" in m for m in caplog.messages), caplog.messages
+    # The prior set is INTACT: exactly the 10 names, none lost, none
+    # orphaned on top.
+    assert {p.name for p in page_dir.iterdir()} == {
+        f"{doc1.id}-p{n}.png" for n in range(1, 11)
+    }, "files on disk exactly the prior set"
+    assert (
+        page_dir / f"{doc1.id}-p1.png"
+    ).read_bytes() == PNG_MAGIC + b"NEW-0", (
+        "p1's replacement bytes were durable before the failure (atomic "
+        "replace) and were NOT removed by the cleanup"
+    )
+    for n in range(2, 11):
+        assert (
+            page_dir / f"{doc1.id}-p{n}.png"
+        ).read_bytes() == prior[n], f"prior copy p{n} was disturbed"
+    assert not list(page_dir.glob("*.tmp")), "zero .tmp orphans from this call"
+    assert doc2.chunks, "the document is indexed fully (fail-soft)"
+
+
+def test_page_copies_are_written_atomically_tmp_then_replace(
+    db, tmp_path: Path, monkeypatch: pytest.MonkeyPatch
+) -> None:
+    """THE atomic pin (LOCKED A4): the writer NEVER writes a final
+    ``<doc-id>-p{N}.png`` path directly — every page's bytes land on
+    ``<name>.png.tmp`` and ``os.replace`` moves them onto the final
+    name in order (a reader opening the final path never sees truncated
+    bytes), and no ``.tmp`` survives the successful call."""
+    root = tmp_path / "pgatom"
+    root.mkdir()
+    _build_ikea_like_pdf(root / "ikea.pdf")
+    llm = _PictorialLLM()
+    llm.settings = _settings(tmp_path)
+    written: list[Path] = []
+    original_write = Path.write_bytes
+
+    def write_spy(self: Path, data: bytes) -> int:
+        written.append(self)
+        return original_write(self, data)
+
+    replaced: list[tuple[str, str]] = []
+    real_replace = os.replace
+
+    def replace_spy(src, dst, *args, **kwargs):  # pyright: ignore[reportUnknownParameterType]
+        replaced.append((os.fspath(src), os.fspath(dst)))
+        return real_replace(src, dst, *args, **kwargs)
+
+    monkeypatch.setattr(Path, "write_bytes", write_spy)  # pyright: ignore[reportArgumentType]
+    monkeypatch.setattr(importer.os, "replace", replace_spy)
+    try:
+        asyncio.run(import_sources([root], llm, session=db))
+    finally:
+        monkeypatch.undo()  # pyright: ignore[reportUnknownMemberType]
+    doc = _doc(db, "pgatom", "ikea.pdf")
+    page_dir = _pages_dir(tmp_path)
+    assert len(written) == 10 and len(replaced) == 10
+    assert all(p.name.endswith(".png.tmp") for p in written), (
+        "no final path is ever written directly"
+    )
+    for n, (src, dst) in enumerate(replaced, start=1):
+        assert dst == str(page_dir / f"{doc.id}-p{n}.png")
+        assert src == f"{dst}.tmp", "the tmp name is <final>.tmp"
+    assert not list(page_dir.glob("*.tmp")), "no .tmp survives the success path"
+
+
+def test_stale_tail_delete_failure_is_best_effort(
+    db,
+    tmp_path: Path,
+    monkeypatch: pytest.MonkeyPatch,
+    caplog: pytest.LogCaptureFixture,
+) -> None:
+    """An unreadable STALE-TAIL copy (the write-first tail pass, the
+    ``_delete_page_copies`` convention): the failure is logged, never
+    raised — the sync completes, the row is stamped with the fresh
+    count, and the reachable tail files still go."""
+    root = tmp_path / "pgtail"
+    root.mkdir()
+    pdf = root / "ikea.pdf"
+    _build_ikea_like_pdf(pdf)
+    llm = _PictorialLLM()
+    llm.settings = _settings(tmp_path)
+    asyncio.run(import_sources([root], llm, session=db))
+    doc1 = _doc(db, "pgtail", "ikea.pdf")
+    page_dir = _pages_dir(tmp_path)
+    p10 = page_dir / f"{doc1.id}-p10.png"
+    real_unlink = Path.unlink
+
+    def boom(self: Path, **kwargs: Any) -> None:
+        if self == p10:
+            raise OSError("simulated unreadable file (test sentinel)")
+        return real_unlink(self, **kwargs)
+
+    monkeypatch.setattr(Path, "unlink", boom)  # pyright: ignore[reportArgumentType]
+    _build_ikea_shrunk(pdf)
+    try:
+        with caplog.at_level(logging.WARNING, logger="app.importer"):
+            s2 = asyncio.run(import_sources([root], llm, session=db))
+    finally:
+        monkeypatch.undo()  # pyright: ignore[reportUnknownMemberType]
+    assert (s2.added, s2.updated, s2.unchanged) == (0, 1, 0)
+    doc2 = _doc(db, "pgtail", "ikea.pdf")
+    assert doc2.page_count == 6 and doc2.page_strip_mode == "additive"
+    assert p10.exists(), "the unreadable tail file survived (never raised)"
+    for n in range(7, 10):
+        assert not (page_dir / f"{doc1.id}-p{n}.png").exists(), f"p{n} still leaked"
+    assert any(
+        "could not delete PDF page copy" in m for m in caplog.messages
+    ), caplog.messages
+
+
 # ---------------------------------------------------------------------------
 # The shared cleanup helper (the _delete_image_copy convention)
 # ---------------------------------------------------------------------------

tests/unit/test_summary_edit_ui.py

@@ -117,7 +117,10 @@ def test_gate_runs_after_the_phase36_base_construction() -> None:
     label = js.find('title.textContent = "Summary"')
     text_node = js.find('body.className = "doc-summary-text"')
     append = js.find("section.append(title, body)")
-    mount = js.find("contentEl.appendChild(section)")
+    # Phase 163 (task 04): the mount rides the `slot` variable —
+    # slot === contentEl for every non-additive doc (the byte-identity
+    # property), views only for the additive strip class.
+    mount = js.find("slot.appendChild(section)")
     gate = js.find("void docAdminReady().then(", mount)
     wiring = js.find("if (admin) wireSummaryEdit(section, doc);")
     assert 0 < base < label < text_node < append < mount < gate < wiring, (

tests/unit/test_viewer_additive_toggle.py

@@ -0,0 +1,431 @@
+"""Unit: the additive PDF strip view (phase 163, task 04 — D4, LOCKED A1/A6).
+
+Frontend source-contract pins + node behavioral runs (the
+``test_document_viewer.py`` house pattern) over ``document.js``'s
+``renderDocument`` — the ONE shared core (Sources modal AND
+/document.html):
+
+* the MODE branch — ``page_urls`` + ``page_strip_mode === "additive"``
+  falls through to the text-doc rendering (Summary panel + content,
+  the pre-148 branch reused) wrapped in a ``.doc-views`` container
+  with the ``.doc-view-toggle`` group above it (Text pressed by
+  default, LOCKED A6); ``page_urls`` with the mode ABSENT or
+  ``"primary"`` keeps the phase-148 strip-only branch VERBATIM (append
+  the strip to ``contentEl`` + return — no toggle, no wrapper, no
+  Summary panel); falsy ``page_urls`` is untouched;
+* the toggle semantics — ``role="group"`` + ``aria-label="Document
+  view"``, two ``<button type="button">`` labeled ``Text`` / ``Pages``
+  (fixed strings, no doc-derived text), ``aria-pressed`` true/false
+  initial state, and a click mutating ONLY the two ``aria-pressed``
+  states + the wrapper's ``data-view`` attribute (CSS-driven show/
+  hide — the strip node stays mounted, no re-render, no refetch);
+* the never-build-HTML rule — createElement + properties/textContent
+  only in the new code path (the module's ONE innerHTML stays the
+  escape-first markdown render); the strip imgs keep
+  ``loading = "lazy"`` (the hidden strip never fetches);
+* the ``.doc-view-toggle`` CSS — the house ghost-pill idiom, the
+  pressed state as text + style (never color alone, B5), ≥44px
+  targets, and the ``data-view`` show/hide contract.
+"""
+from __future__ import annotations
+
+import json
+import re
+import shutil
+import subprocess
+from pathlib import Path
+
+import pytest
+
+FRONTEND = Path(__file__).resolve().parents[2] / "frontend"
+DOCUMENT_JS = FRONTEND / "assets" / "document.js"
+STYLES_CSS = FRONTEND / "assets" / "styles.css"
+
+HAVE_NODE = shutil.which("node") is not None
+
+
+def _read(path: Path) -> str:
+    return path.read_text(encoding="utf-8")
+
+
+def _fn(js: str, name: str) -> str:
+    """The source of a top-level function via balanced-brace counting
+    (the ``test_document_viewer.py`` pattern)."""
+    start = js.find(f"function {name}(")
+    assert start != -1, f"{name}() must exist in document.js"
+    i = js.find("(", start)
+    depth = 0
+    close = i
+    while i < len(js):
+        if js[i] == "(":
+            depth += 1
+        elif js[i] == ")":
+            depth -= 1
+            if depth == 0:
+                close = i
+                break
+        i += 1
+    brace = js.find("{", close)
+    depth = 0
+    for j in range(brace, len(js)):
+        if js[j] == "{":
+            depth += 1
+        elif js[j] == "}":
+            depth -= 1
+            if depth == 0:
+                return js[start : j + 1]
+    raise AssertionError(f"unbalanced braces in {name}()")
+
+
+def _code(js: str) -> str:
+    """The source with comments stripped (the never-build-HTML pins)."""
+    return re.sub(r"//.*?$|/\*.*?\*/", "", js, flags=re.S | re.M)
+
+
+def _run_node(script: str) -> str:
+    proc = subprocess.run(
+        ["node", "-e", script], capture_output=True, text=True, timeout=60
+    )
+    assert proc.returncode == 0, f"node failed: {proc.stderr}"
+    return proc.stdout
+
+
+# ---------------------------------------------------------------------------
+# Source pins — the renderDocument mode branch
+# ---------------------------------------------------------------------------
+
+
+def test_render_document_branches_on_strip_mode() -> None:
+    """The mode branch (LOCKED A2 — the tolerant read): the additive
+    flag keys on the EXACT ``doc.page_strip_mode === "additive"``
+    string (absent / "primary" / anything else ⇒ primary), and the
+    primary/absent strip branch stays the phase-148 shape — append
+    ``docPagesStrip`` to ``contentEl`` and RETURN — evaluated BEFORE
+    any wrapper/toggle construction."""
+    render = _fn(_read(DOCUMENT_JS), "renderDocument")
+    assert 'doc.page_strip_mode === "additive"' in render, (
+        "only the exact additive string switches on the toggle"
+    )
+    gate_i = render.find('const additive = stripEligible && doc.page_strip_mode === "additive";')
+    primary_i = render.find("if (stripEligible && !additive) {")
+    strip_i = render.find("contentEl.appendChild(docPagesStrip(doc, pageUrls));")
+    return_i = render.find("return;", strip_i)
+    wrapper_i = render.find('views.className = "doc-views";')
+    assert -1 < gate_i < primary_i < strip_i < return_i < wrapper_i, (
+        "the primary/absent branch (strip into contentEl + return) precedes "
+        "the additive wrapper — its DOM stays the phase-148 shape verbatim"
+    )
+    # The wrapper + toggle: ONE .doc-views container, the toggle BEFORE
+    # it, both into contentEl; the text view lands in the wrapper.
+    append_i = render.find("contentEl.append(docViewToggle(views), views);")
+    assert wrapper_i < append_i, "the wrapper is built, then appended after its toggle"
+    assert 'views.setAttribute("data-view", "text");' in render, (
+        "LOCKED A6 — Text is the default view"
+    )
+    # The strip mounts ONCE inside the wrapper, after the text view —
+    # never rebuilt on a flip (no docPagesStrip call outside the two
+    # branches).
+    assert render.count("docPagesStrip(doc, pageUrls)") == 2
+    mount_i = render.rfind("views.appendChild(docPagesStrip(doc, pageUrls));")
+    assert mount_i > render.find('slot.appendChild(pre);') or mount_i > render.rfind(
+        "slot.appendChild(wrap);"
+    ), "the strip mounts AFTER the text view inside the wrapper"
+    # The non-additive path is untouched: slot === contentEl (the
+    # pre-148 appends keep their exact shape through the slot var).
+    assert "let slot = contentEl;" in render
+
+
+def test_primary_absent_branch_builds_no_toggle() -> None:
+    """The primary/absent branch is toggle-free by construction: the
+    strip-branch block (from the ``pageUrls`` read to its ``return``)
+    contains NO ``doc-view-toggle`` / ``doc-views`` construction — the
+    strip-only DOM stays byte-identical to pre-phase-163."""
+    render = _fn(_read(DOCUMENT_JS), "renderDocument")
+    primary_start = render.find("if (stripEligible && !additive)")
+    branch = render[render.find("const pageUrls =") : render.find("return;", primary_start)]
+    assert "doc-view-toggle" not in branch
+    assert "doc-views" not in branch
+    assert branch.count("appendChild") == 1, (
+        "the primary branch is exactly the phase-148 append + return"
+    )
+
+
+def test_toggle_never_builds_html_and_labels_are_fixed() -> None:
+    """The never-build-HTML rule: the toggle builders use
+    createElement + properties/textContent only — no innerHTML /
+    outerHTML — and the labels are the FIXED strings ``Text`` /
+    ``Pages`` (no doc-derived text inside the toggle). The module's
+    ONE innerHTML (the escape-first markdown render) survives."""
+    js = _read(DOCUMENT_JS)
+    toggle = _fn(js, "docViewToggle")
+    btn = _fn(js, "viewToggleButton")
+    for body in (toggle, btn):
+        code = _code(body)
+        assert "innerHTML" not in code and "outerHTML" not in code
+    assert 'document.createElement("div")' in toggle
+    assert 'group.setAttribute("role", "group")' in toggle
+    assert 'group.setAttribute("aria-label", "Document view")' in toggle
+    assert 'document.createElement("button")' in btn
+    assert 'btn.type = "button"' in btn
+    assert 'btn.className = "doc-view-btn"' in btn
+    assert 'viewToggleButton("Text")' in toggle
+    assert 'viewToggleButton("Pages")' in toggle
+    # The module-wide XSS contract: exactly one innerHTML, the
+    # renderMarkdown line.
+    whole = _code(js)
+    assert whole.count("innerHTML") == 1
+    assert "outerHTML" not in whole
+
+
+def test_strip_imgs_stay_lazy() -> None:
+    """The phase-148 lazy-img pin re-runs green: the strip builder is
+    UNMODIFIED — every page ``<img>`` keeps ``loading = "lazy"`` (the
+    additive default view never fetches the strip)."""
+    strip = _fn(_read(DOCUMENT_JS), "docPagesStrip")
+    assert 'img.loading = "lazy";' in strip
+
+
+def test_view_toggle_css() -> None:
+    """styles.css: the ``.doc-view-toggle`` group (the chat-column
+    reading width, the pill pair adjacent) + the ``.doc-view-btn``
+    ghost pill (≥44px target, ink-soft text — AA) + the PRESSED state
+    (``[aria-pressed="true"]`` — brand fill + font-weight 700: text +
+    style, never color alone, B5) + the CSS-driven view swap
+    (``.doc-views .doc-pages`` hidden by default; ``[data-view=
+    "pages"]`` hides the Summary panel / .doc-md / .doc-raw /
+    .doc-image and shows the strip as the grid). Provenance comment
+    above the block (house style)."""
+    css = _read(STYLES_CSS)
+    toggle_i = css.find(".doc-view-toggle {")
+    assert toggle_i != -1, "styles.css must style .doc-view-toggle"
+    comment_start = css.rfind("/*", 0, toggle_i)
+    assert -1 < comment_start < toggle_i and "phase 163" in css[comment_start:toggle_i]
+    btn_i = css.find(".doc-view-btn {")
+    pressed_i = css.find('.doc-view-btn[aria-pressed="true"]')
+    assert toggle_i < btn_i < pressed_i
+    btn_rule = css[btn_i : css.find("}", btn_i)]
+    for decl in ("min-height: 44px", "color: var(--ink-soft)", "border: 1px solid var(--line)"):
+        assert decl in btn_rule
+    pressed_rule = css[pressed_i : css.find("}", pressed_i)]
+    assert "background: var(--brand)" in pressed_rule
+    assert "font-weight: 700" in pressed_rule, (
+        "the pressed state is text weight + fill — never color alone (B5)"
+    )
+    # The CSS-driven swap: hidden strip by default; pages hides the
+    # text view and shows the strip.
+    assert '.doc-views .doc-pages { display: none; }' in css
+    for sel in (".doc-summary", ".doc-md", ".doc-raw", ".doc-image"):
+        assert f'.doc-views[data-view="pages"] {sel}' in css
+    assert '.doc-views[data-view="pages"] .doc-pages { display: grid; }' in css
+
+
+# ---------------------------------------------------------------------------
+# Node behavioral runs — the real renderDocument over a minimal fake DOM
+# ---------------------------------------------------------------------------
+
+_FAKE_DOM = """
+function makeEl(tag) {
+  return {
+    tag, className: "", textContent: "", attrs: {}, children: [], handlers: {},
+    setAttribute(k, v) { this.attrs[k] = String(v); },
+    getAttribute(k) { return k in this.attrs ? this.attrs[k] : null; },
+    appendChild(c) { this.children.push(c); return c; },
+    append(...cs) { this.children.push(...cs); },
+    replaceChildren(...cs) { this.children = cs; },
+    addEventListener(ev, fn) { this.handlers[ev] = fn; },
+    click() { this.handlers.click(); },
+  };
+}
+globalThis.document = { createElement: (t) => makeEl(t) };
+function fmtDate(x) { return "DATE:" + x; }
+function metaBadge(cls, text, title) {
+  const s = makeEl("span");
+  s.className = cls;
+  s.textContent = text;
+  if (title !== undefined) s.setAttribute("title", title);
+  return s;
+}
+function copyPathButton() { return makeEl("button"); }
+function originAnchor() { return makeEl("a"); }
+async function docAdminReady() { return false; }
+function wireDateEdit() {}
+function wireSummaryEdit() {}
+function renderMarkdown(s) { return s; }
+function docImageBlock() { return makeEl("div"); }
+function describe(el) {
+  return {
+    tag: el.tag,
+    className: el.className,
+    attrs: el.attrs,
+    text: el.textContent,
+    children: el.children.map(describe),
+  };
+}
+"""
+
+
+def _render_script() -> str:
+    js = _read(DOCUMENT_JS)
+    parts = [
+        _fn(js, n)
+        for n in ("renderDocument", "docViewToggle", "viewToggleButton", "docPagesStrip")
+    ]
+    return _FAKE_DOM + "\n".join(parts)
+
+
+@pytest.mark.skipif(not HAVE_NODE, reason="node not available")
+def test_additive_doc_renders_text_views_and_toggle() -> None:
+    """Behavioral (node, fake DOM): an ADDITIVE doc renders the TEXT
+    doc exactly as a non-qualifying one — Summary panel + content —
+    inside ``.doc-views`` (data-view="text"), with the toggle group
+    FIRST in the content slot: role=group, aria-label="Document
+    view", the two ``doc-view-btn`` buttons labeled Text/Pages, Text
+    aria-pressed="true" (LOCKED A6), and the strip MOUNTED inside the
+    wrapper (figures + lazy imgs) after the text view."""
+    out = _run_node(
+        _render_script()
+        + """
+const doc = {
+  title: "Letumdei", path: "book.pdf", source: "Library", format: "pdf",
+  summary: "A book.", content: "BODY TEXT", created_at: "c", indexed_at: "i",
+  chunks: 9, page_urls: ["/api/documents/x/pages/1", "/api/documents/x/pages/2"],
+  page_strip_mode: "additive",
+};
+const contentEl = makeEl("div");
+renderDocument(doc, { titleEl: makeEl("h1"), metaEl: makeEl("div"), contentEl });
+console.log(JSON.stringify(describe(contentEl)));
+"""
+    )
+    content = json.loads(out)
+    assert [c["className"] for c in content["children"]] == [
+        "doc-view-toggle",
+        "doc-views",
+    ], "the toggle group sits above the .doc-views wrapper"
+    toggle, views = content["children"]
+    assert toggle["attrs"]["role"] == "group"
+    assert toggle["attrs"]["aria-label"] == "Document view"
+    labels = [c["text"] for c in toggle["children"]]
+    assert labels == ["Text", "Pages"]
+    assert [c["attrs"]["aria-pressed"] for c in toggle["children"]] == ["true", "false"], (
+        "Text pressed by default (LOCKED A6)"
+    )
+    for c in toggle["children"]:
+        assert c["tag"] == "button"
+    assert views["attrs"]["data-view"] == "text"
+    assert [c["className"] for c in views["children"]] == [
+        "doc-summary",
+        "doc-raw",
+        "doc-pages",
+    ], "additive = Summary panel + text content + the mounted strip"
+    # pre.textContent — a text node, never parsed as HTML
+    assert views["children"][1]["text"] == "BODY TEXT"
+    strip = views["children"][2]
+    assert len(strip["children"]) == 2, "one figure per page_urls entry"
+
+
+@pytest.mark.skipif(not HAVE_NODE, reason="node not available")
+def test_primary_and_absent_modes_keep_the_strip_only_dom() -> None:
+    """Behavioral: mode "primary" AND an absent mode (the pre-163
+    stamped row — the live Letumdei pre-sync shape) keep the phase-148
+    DOM byte-identically: the content slot holds EXACTLY the
+    ``.doc-pages`` strip — no wrapper, no toggle, no Summary panel,
+    no text block."""
+    out = _run_node(
+        _render_script()
+        + """
+const results = [];
+for (const mode of ["primary", undefined]) {
+  const doc = {
+    title: "Manual", path: "m.pdf", source: "Library", format: "pdf",
+    summary: "A manual.", content: "SHOULD NOT SHOW", created_at: "c",
+    indexed_at: "i", chunks: 4,
+    page_urls: ["/api/documents/y/pages/1"],
+  };
+  if (mode !== undefined) doc.page_strip_mode = mode;
+  const contentEl = makeEl("div");
+  renderDocument(doc, { titleEl: makeEl("h1"), metaEl: makeEl("div"), contentEl });
+  results.push(describe(contentEl));
+}
+console.log(JSON.stringify(results));
+"""
+    )
+    primary, absent = json.loads(out)
+    for content in (primary, absent):
+        assert len(content["children"]) == 1
+        assert content["children"][0]["className"] == "doc-pages"
+        classes = json.dumps(content)
+        assert "doc-view-toggle" not in classes
+        assert "doc-views" not in classes
+        assert "doc-summary" not in classes
+        assert "doc-raw" not in classes, "the strip replaces the text block (primary class)"
+
+
+@pytest.mark.skipif(not HAVE_NODE, reason="node not available")
+def test_text_doc_dom_is_untouched() -> None:
+    """Behavioral: a doc with no page_urls is UNTOUCHED — Summary
+    panel + content appended straight to the content slot, no
+    wrapper, no toggle (the byte-identity property for every
+    non-PDF)."""
+    out = _run_node(
+        _render_script()
+        + """
+const doc = {
+  title: "Note", path: "n.pdf", source: "Library", format: "pdf",
+  summary: "S.", content: "TEXT", created_at: "c", indexed_at: "i", chunks: 1,
+};
+const contentEl = makeEl("div");
+renderDocument(doc, { titleEl: makeEl("h1"), metaEl: makeEl("div"), contentEl });
+console.log(JSON.stringify(describe(contentEl)));
+"""
+    )
+    content = json.loads(out)
+    assert [c["className"] for c in content["children"]] == ["doc-summary", "doc-raw"]
+
+
+@pytest.mark.skipif(not HAVE_NODE, reason="node not available")
+def test_toggle_flips_view_without_rerender_or_refetch() -> None:
+    """Behavioral: the toggle handler mutates ONLY the two
+    aria-pressed states + the wrapper's data-view. Clicking Pages
+    flips data-view to "pages"; clicking back to Text reverses; the
+    strip node STAYS MOUNTED the whole time (same object, no
+    re-append — the flip-flop fetches nothing) and the wrapper's
+    children never change count."""
+    out = _run_node(
+        _render_script()
+        + """
+const doc = {
+  title: "Book", path: "b.pdf", source: "Library", format: "pdf",
+  summary: "S.", content: "TEXT", created_at: "c", indexed_at: "i", chunks: 2,
+  page_urls: ["/api/documents/z/pages/1", "/api/documents/z/pages/2"],
+  page_strip_mode: "additive",
+};
+const contentEl = makeEl("div");
+renderDocument(doc, { titleEl: makeEl("h1"), metaEl: makeEl("div"), contentEl });
+const toggle = contentEl.children[0];
+const views = contentEl.children[1];
+const [textBtn, pagesBtn] = toggle.children;
+const stripBefore = views.children[2];
+const countBefore = views.children.length;
+pagesBtn.click();
+const afterPages = {
+  view: views.getAttribute("data-view"),
+  pressed: toggle.children.map((b) => b.getAttribute("aria-pressed")),
+};
+textBtn.click();
+const afterText = {
+  view: views.getAttribute("data-view"),
+  pressed: toggle.children.map((b) => b.getAttribute("aria-pressed")),
+  stripSame: views.children[2] === stripBefore,
+  countSame: views.children.length === countBefore,
+};
+console.log(JSON.stringify({ afterPages, afterText }));
+"""
+    )
+    res = json.loads(out)
+    assert res["afterPages"] == {"view": "pages", "pressed": ["false", "true"]}
+    assert res["afterText"] == {
+        "view": "text",
+        "pressed": ["true", "false"],
+        "stripSame": True,
+        "countSame": True,
+    }, "the strip node persists across flips — no re-render, no refetch"

tests/unit/test_wide_column_css.py

@@ -13,12 +13,13 @@ without a browser:
 * NO ``@media (min-width: 1500px)`` block anywhere in the file — the
   phase-58 wide-desktop doubling is deleted in full (the token is
   72rem at every viewport);
-* the six reading-column selectors — ``.chat-shell``,
+* the reading-column selectors — ``.chat-shell``,
   ``.shared-shell``, ``.doc-md``, ``.doc-image`` (phase 122, task 04),
-  ``.doc-pages`` (phase 148, task 03 — the page strip), and
+  ``.doc-pages`` (phase 148, task 03 — the page strip),
-  ``.doc-summary:has(+ .doc-md)`` — each capped with
+  ``.doc-view-toggle`` (phase 163, task 04 — the Text/Pages toggle
-  ``max-width: var(--chat-column)`` and NOTHING else in the file uses
+  above the additive views), and ``.doc-summary:has(+ .doc-md)`` —
-  the token for a max-width (exactly six rules);
+  each capped with ``max-width: var(--chat-column)`` and NOTHING else
+  in the file uses the token for a max-width (exactly seven rules);
 * the flipped negative pin — ZERO literal ``max-width: 46rem`` rules
   remain (the phase-27/59/91 form-column caps are retired): the form
   shells (``.tuning-shell``, ``.theme-shell``, ``.doc-edit-shell``)
@@ -121,8 +122,10 @@ def test_the_reading_columns_use_the_token() -> None:
     task 04 — the viewer's image block rides the SAME reading column
     as its .doc-md sibling), .doc-pages (phase 148, task 03 — the
     viewer's page strip rides the SAME reading column in the shared
-    #doc-content slot) and .doc-summary:has(+ .doc-md) each cap with
+    #doc-content slot), .doc-view-toggle (phase 163, task 04 — the
-    max-width: var(--chat-column) — and exactly those six rules use
+    Text/Pages toggle sits above the additive views, aligned with
+    them) and .doc-summary:has(+ .doc-md) each cap with
+    max-width: var(--chat-column) — and exactly those seven rules use
     the token for a max-width (no other selector)."""
     css = _css()
     for selector in (
@@ -131,12 +134,13 @@ def test_the_reading_columns_use_the_token() -> None:
         ".doc-md",
         ".doc-image",
         ".doc-pages",
+        ".doc-view-toggle",
         ".doc-summary:has(+ .doc-md)",
     ):
         assert "max-width: var(--chat-column)" in _rule_block(css, selector), (
             f"{selector} must cap with max-width: var(--chat-column)"
         )
-    assert css.count("max-width: var(--chat-column)") == 6, (
+    assert css.count("max-width: var(--chat-column)") == 7, (
         "exactly the reading-column selectors use the token"
     )