Files
brain-of-reese/tests/e2e/test_steering.py
T
ducoterra ffa919b8bf fix(chat): keep in-flight answers alive across in-app view switches
Root cause (owner repro, verified in a real browser 2026-09-06): the
five navbar views (Chat, RAG, Sources, Tuning, History) were separate
HTML documents, so a navbar click was a REAL cross-document navigation
— the chat page unloaded, the in-flight SSE fetch was aborted, and the
phase-48 teardown (app/api/chat.py `finally`, "chat: turn cancelled")
stopped the model. Observed: send question -> click RAG mid-stream ->
click Chat -> the answer never finished: no `query_log` row, and on
return a dangling question with no brain record (the pre-token pagehide
partial persist skips because `acc` is empty).

Phase-48 LOCKED-DECISION REFINEMENT (owner-confirmed 2026-09-06,
flagged per AGENTS.md rule 3, not silently deviated): "real navigation
cancels the fetch" now means LEAVING THE APP — tab close,
external/other-document navigation, the Stop button. In-app navbar
switches are client-side view switches and no longer cancel.

Fix — Option A (SPA shell), chosen over B (Service Worker owns the
stream) and C (server-side turn registry + resume):
- frontend/index.html is the shell: ONE `<main id="main">` holds the
  five `<section class="view">` blocks; hidden views carry BOTH
  `hidden` and `inert` (WCAG — no focus/keyboard traversal). The
  shared header, the single `doc-modal-*` skeleton, and the
  `#app-version` footer each exist exactly once; the per-view copies
  from the four folded pages are dropped.
- New frontend/assets/router.js (vanilla module — no framework, no
  bundler, No-CDN rule intact): lazy-imports a view module on FIRST
  show only (mount-once, hide-forever — the chat view's in-flight SSE
  reader persists across switches; that persistence IS the fix);
  intercepts same-shell navbar links with preventDefault +
  history.pushState (never a document load); handles popstate; single
  writer of `.nav-link` active state (is-active + aria-current),
  document.title, and the per-view meta description (values carried
  over from the old pages' heads, brand-resolved at write time).
- Each folded page's JS becomes `export async function mount(root)` —
  root-scoped queries; `initSharedHeader()` dropped (the header boots
  once in the shell via the chat module; the admin flag comes from the
  same cached `fetchIsAdmin()` promise — zero extra requests).
- app/main.py: a small list-driven route factory serves the shell for
  /tuning.html, /sources.html, /git-sources.html, /history.html —
  registered AFTER the API routers and BEFORE the static catch-all
  (routes-first). The phase-33 caching middleware applies no-cache +
  `?v=` rewriting unchanged; app/core/caching.py needed NO change
  (the view paths did not change — pinned by the integration tests).
- The four old view .html files are DELETED (one source of truth);
  deep links to the old URLs keep working (the router picks the view
  from the pathname); `/?chat=<id>` is unaffected; the Containerfile
  bundles router.js (inlining the lazy view modules) and drops the
  folded page files.
- app/schemas.py: HistoryTurn.text cap 4000 -> 32000 — the shell
  keeps long saved answers in the chat, and the old cap (stricter than
  the 24_000-char total history budget) 422-rejected any second turn
  in such a chat (found by the phase-42 E2E suite on the shell).

Boundaries: login.html, shared.html, doc-edit.html, document.html
REMAIN separate documents (flow pages, not navbar tabs); a mid-stream
navigation to doc-edit/document.html still cancels per phase 48
(follow-up candidate, out of scope). The SSE API is unchanged. Real
departures still cancel the turn — phase 48 intact (pinned by
tests/e2e/test_stop_generation.py, unchanged, and by the new suite's
real-departure control).

Tests:
- Phase-20 suite REWRITTEN to the new semantics
  (tests/e2e/test_sources_midstream_bug.py): a navbar switch no longer
  cancels — the stream survives the switch and the FULL answer
  settles; the pagehide partial persist REMAINS for real departures
  (the partial's exact shape — first streamed chunk prefix, no done
  metadata — is still pinned there).
- NEW story suite tests/e2e/test_nav_switch_keeps_stream.py (mock
  LLM): the owner repro (send -> RAG mid-stream -> Chat: window
  sentinel survives = same document, FULL answer, exactly one brain
  turn in bor.chat.v1, exactly one settled query_log row, auto-saved
  row matches) + the same mid-stream switch against the other three
  views + the real-departure-still-cancels control + the no-switch
  baseline.
- tests/unit/test_frontend_router.py: source-level pins of the router
  invariants (click interceptor targets ONLY same-shell view paths,
  pushState-only switches, mount-once guard, hidden+inert pair,
  single-writer active state/title); shell-route integration tests
  (each folded path serves the shell with no-cache + `?v=` body; a
  non-view path still 404s); the file-reading unit pins re-pointed at
  the shell (the four view files are gone — the shell is the source
  of truth).

Verification (this commit): full suite green — 1565 unit+integration
tests, app/ coverage 99% (>90% floor); ruff + pyright clean; the
phase's E2E suites green in isolation (house protocol, AGENTS.md rule
9). Owner repro verified in a real browser against the real LLM
(dev server :8010, headful Chromium): "tell me about everquest" ->
RAG mid-stream -> Chat — the answer completed with one brain bubble
and no error banner, `query_log` gained exactly one settled row
(deflected=True: the dev KB holds no EverQuest docs — the settle, not
the topic, is the proof), zero "chat: turn cancelled" lines for that
turn; the control (real navigation to /shared.html mid-stream) still
cancelled (no settled row, the cancel line logged, the partial
persisted on return). Screenshots: .agents/screenshots/76_manual_*.

Phase 76 (76_spa_nav_shell) complete — moved to
.agents/phases/complete/.
2026-09-06 06:31:31 -04:00

295 lines
12 KiB
Python

"""Phase 15 E2E (Playwright): tune how Brain answers (steering notes).
Phase 16 adaptation: tuning is admin-only — every test performs the real
form login (``e2e.auth_helpers.login``) before touching the tuning UI.
Story: ``.agents/user_stories/steering-notes.md``
Run in isolation (DB must be up: ``podman compose up -d db``):
uv run pytest tests/e2e/test_steering.py -v --no-cov
The steering loop: "Tune" under a completed answer → short instruction →
stored in Postgres (``steering_notes``) → injected into the system prompt
of every subsequent turn as the ``<tuning>`` section. The mock LLM
echoes the first tuning note into its answer
(`` (tuning: <first note line>)``), so prompt injection is observable in
the UI deterministically. Notes are listed newest-first on the Tuning
page (``/tuning.html``), where each can be deleted — the header "Tuning"
toggle was removed from the navbar at owner request (2026-08-28).
Test → story mapping (Playwright Mapping Rule):
1. ``test_tune_under_answer_persists_and_steers``
2. ``test_delete_note_stops_steering``
3. ``test_note_rendered_as_text_xss_safe``
4. ``test_tuning_panel_a11y``
"""
from __future__ import annotations
import asyncio
from pathlib import Path
from threading import Thread
from typing import Any
from playwright.sync_api import Page, expect
from sqlalchemy import select, text
from app.config import Settings
from app.db import SessionLocal
from app.models import SteeringNote
from app.rag.importer import ImportSummary, import_sources
from app.rag.llm import LLMClient
from e2e.auth_helpers import login
REPO = Path(__file__).resolve().parents[2]
FIXTURES = REPO / "tests" / "fixtures" / "docs"
QUESTION = "How is my Kubernetes cluster set up?"
MOCK_ANSWER_MARKER = "Deterministic mock answer for E2E"
NOTE = "STEEER-MARKER be concise"
XSS_NOTE = "<script>window.__xss = true; alert('xss')</script>"
#: The shell (index.html — served for BOTH / and /tuning.html since
#: phase 76 task 01, when the Tuning view folded into it) ships exactly
#: FOUR classic/module script tags: the phase-39 brand.js classic layer
#: + markdown.js + the chat module (app.js) + the shell router
#: (router.js, which lazy-imports the tuning.js view module).
BASE_SCRIPT_COUNT = 4
async def _import_fixtures(mock_port: int) -> ImportSummary:
kwargs: dict[str, Any] = {"_env_file": None, "llm_base_url": f"http://127.0.0.1:{mock_port}/v1"}
settings = Settings(**kwargs) # pyright: ignore[reportCallIssue]
return await import_sources([FIXTURES], LLMClient(settings))
def _run_in_thread(coro: Any) -> Any:
"""Run a coroutine on a worker thread.
Playwright's sync API keeps an asyncio loop running on the test thread,
so ``asyncio.run`` cannot be called directly from a test body.
"""
box: dict[str, Any] = {}
def runner() -> None:
try:
box["value"] = asyncio.run(coro)
except BaseException as e: # noqa: BLE001 — re-raised on the test thread
box["error"] = e
t = Thread(target=runner)
t.start()
t.join()
if "error" in box:
raise box["error"]
return box["value"]
def _reset_db(mock_port: int, seed: bool) -> ImportSummary | None:
"""Truncate the KB (and query log + steering notes), re-import fixtures."""
with SessionLocal() as db:
db.execute(text("TRUNCATE chunks, documents, query_log, steering_notes"))
db.commit()
if not seed:
return None
return _run_in_thread(_import_fixtures(mock_port))
def _ask(page: Page, question: str) -> None:
"""Send one turn and wait until the grounded answer has fully landed."""
page.fill("#message-input", question)
page.click("#send-btn")
expect(page.locator(".msg.user .bubble").last).to_contain_text(question)
expect(page.locator(".msg.brain .bubble").last).to_contain_text(
MOCK_ANSWER_MARKER, timeout=30_000
)
expect(page.locator("#send-btn")).to_be_enabled()
expect(page.locator("#send-label")).to_have_text("Send")
def _tune_and_save(page: Page, note: str) -> None:
"""Tune the last completed brain bubble and save *note*."""
tune = page.locator(".msg.brain .tune-btn").last
expect(tune).to_be_visible()
tune.click()
form = page.locator(".msg.brain .tune-form").last
expect(form).to_be_visible()
form.locator("textarea").fill(note)
form.locator(".tune-save").click()
saved = page.locator(".msg.brain .tune-saved").last
expect(saved).to_contain_text("Saved — future answers will follow this.", timeout=15_000)
# ---------------------------------------------------------------------------
# 1. Tune under an answer → persisted → next answer carries the note
# ---------------------------------------------------------------------------
def test_tune_under_answer_persists_and_steers(
page: Page, app_url: str, mock_llm: int, db_ready: None
) -> None:
summary = _reset_db(mock_llm, seed=True)
assert summary is not None and summary.added == 13 # A9 formats (phase 47 added quadlet+j2)
page.set_default_timeout(30_000)
page.goto(app_url)
login(page, app_url, next="/") # phase 16: tuning is admin-only
_ask(page, QUESTION)
# The Tune control: ghost button in the answer's meta row, ≥44px.
tune = page.locator(".msg.brain .tune-btn").last
expect(tune).to_have_count(1)
expect(tune).to_have_attribute("type", "button")
box = tune.bounding_box()
assert box is not None and box["height"] >= 44
_tune_and_save(page, NOTE)
# Persisted in Postgres.
with SessionLocal() as db:
rows = db.scalars(select(SteeringNote)).all()
assert [r.note for r in rows] == [NOTE]
# The Tuning page (the steering-notes manager — the navbar toggle
# was removed at owner request, 2026-08-28) lists the note.
page.goto(app_url + "/tuning.html")
expect(page.locator("#tune-list .tuning-note")).to_have_count(1)
expect(page.locator("#tune-list .tuning-note-text")).to_have_text(NOTE)
# The NEXT answer carries the note — it reached the system prompt.
page.goto(app_url + "/")
expect(page.locator("#send-btn")).to_be_enabled()
_ask(page, QUESTION)
bubble = page.locator(".msg.brain .bubble").last
expect(bubble).to_contain_text(f"(tuning: {NOTE})")
# ---------------------------------------------------------------------------
# 2. Delete from the panel → count 0 → steering stops
# ---------------------------------------------------------------------------
def test_delete_note_stops_steering(
page: Page, app_url: str, mock_llm: int, db_ready: None
) -> None:
_reset_db(mock_llm, seed=True)
page.set_default_timeout(30_000)
page.goto(app_url)
login(page, app_url, next="/") # phase 16: tuning is admin-only
_ask(page, QUESTION)
_tune_and_save(page, NOTE)
# Steering is live: one more answer carries the marker.
_ask(page, QUESTION)
expect(page.locator(".msg.brain .bubble").last).to_contain_text(f"(tuning: {NOTE})")
# Delete the note from the Tuning page (the header panel is no
# longer reachable — the navbar toggle is gone).
page.goto(app_url + "/tuning.html")
expect(page.locator("#tune-list .tuning-note")).to_have_count(1)
page.locator(".tuning-delete").click()
expect(page.locator("#tune-list .tuning-note")).to_have_count(0, timeout=15_000)
expect(page.locator("#tune-empty")).to_be_visible()
expect(page.locator("#tune-announcer")).to_contain_text("deleted")
with SessionLocal() as db:
assert db.scalars(select(SteeringNote)).all() == []
# The next answer no longer carries the marker.
page.goto(app_url + "/")
expect(page.locator("#send-btn")).to_be_enabled()
_ask(page, QUESTION)
bubble = page.locator(".msg.brain .bubble").last
expect(bubble).to_contain_text(MOCK_ANSWER_MARKER)
expect(bubble).not_to_contain_text("STEEER-MARKER")
# ---------------------------------------------------------------------------
# 3. Notes render as text (XSS-safe)
# ---------------------------------------------------------------------------
def test_note_rendered_as_text_xss_safe(
page: Page, app_url: str, mock_llm: int, db_ready: None
) -> None:
_reset_db(mock_llm, seed=True)
page.set_default_timeout(30_000)
page.goto(app_url)
login(page, app_url, next="/") # phase 16: tuning is admin-only
dialogs: list[str] = []
def _handle_dialog(d) -> None:
dialogs.append(d.message)
d.dismiss()
page.on("dialog", _handle_dialog)
_ask(page, QUESTION)
_tune_and_save(page, XSS_NOTE)
# Tuning page: the payload is visible as LITERAL text…
page.goto(app_url + "/tuning.html")
expect(page.locator("#tune-list .tuning-note-text")).to_have_text(XSS_NOTE)
# …never as an executed element: no script tag in the list, the page
# still carries exactly its own three scripts, no dialog.
assert page.locator("#tune-list script").count() == 0
expect(page.locator("script")).to_have_count(BASE_SCRIPT_COUNT)
assert dialogs == [], f"the note must never execute as script: {dialogs}"
assert page.evaluate("() => window.__xss === undefined") is True
# ---------------------------------------------------------------------------
# 4. Tuning panel accessibility
# ---------------------------------------------------------------------------
def test_tuning_panel_a11y(page: Page, app_url: str, db_ready: None) -> None:
"""The steering surface's a11y now lives on the Tuning page (the
header toggle was removed from the navbar at owner request,
2026-08-28): the note list, the polite live region, and the labeled
per-note delete (≥44px)."""
_reset_db(mock_port=0, seed=False) # no KB seeding needed for the page a11y
page.set_default_timeout(30_000)
page.goto(app_url)
login(page, app_url, next="/tuning.html") # phase 16: the notes are admin-only
# The navbar no longer carries a steering toggle — absent for the
# admin too — and the header panel section still ships hidden.
assert page.locator("#steering-toggle").count() == 0
expect(page.locator("#steering-panel")).to_be_hidden()
# The page's own polite live region.
announcer = page.locator("#tune-announcer")
assert announcer.get_attribute("role") == "status"
assert announcer.get_attribute("aria-live") == "polite"
# Add a note (API), reload to refresh the list (tuning.js fetches
# on boot).
page.evaluate(
"""async () => {
const r = await fetch('/api/steering', {
method: 'POST',
headers: {'Content-Type': 'application/json'},
body: JSON.stringify({note: 'a11y note one'}),
});
if (!r.ok) throw new Error('steering POST failed: ' + r.status);
}"""
)
page.reload()
expect(page.locator("#sign-out-btn")).to_be_visible(timeout=15_000)
note_item = page.locator("#tune-list .tuning-note")
expect(note_item).to_have_count(1)
expect(note_item.locator(".tuning-note-text")).to_have_text("a11y note one")
# The per-note delete is a real, labeled button (≥44px target).
delete = page.locator(".tuning-delete")
expect(delete).to_have_attribute("type", "button")
assert (delete.get_attribute("aria-label") or "").startswith("Delete tuning note:")
box = delete.bounding_box()
assert box is not None and box["height"] >= 44
# Delete: list empties, the empty state shows, the live region
# announces it.
delete.click()
expect(page.locator("#tune-list .tuning-note")).to_have_count(0, timeout=15_000)
expect(page.locator("#tune-empty")).to_be_visible()
expect(announcer).to_contain_text("deleted")