Experimental/brain-of-reese
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
3095c4c5772645e19f50d82b9c681d21f285b357
brain-of-reese/tests/e2e/test_stale_saved_chats.py
T

558 lines
24 KiB
Python
Raw Blame History

"""Phase 53 E2E (Playwright): invalidate saved chats on sources sync.
TODO.md L4 (owner 2026-08-30): "Make sure the saved chats are invalidated
if the docs are synced, that way it generates a new answer with new data".
Run in isolation (DB must be up: ``podman compose up -d db``):
uv run pytest tests/e2e/test_stale_saved_chats.py -v --no-cov
The full user-visible invalidation loop under test:
* **Auto-save (fresh, phase 55)** — as admin: ask (the mock answers
deterministically) — the conversation auto-saves at the save points
(the Save pill is gone); ``GET /api/chats`` (admin cookie) reports
the row with ``stale: false``; opening the FRESH row at
``/?chat=<id>`` shows NO banner;
* **KB change** — the test process (which shares the app's environment)
bumps the ``sources_meta`` seed row through ``bump_sources_version``
over a short ``SessionLocal()`` — the exact helper BOTH real sync
paths (the Sync button, ``scripts/import_docs``) call change-gated.
The bump GATES are pinned by this phase's integration tests (task 02)
and the Sync button's end-to-end clone/import path by
``test_sync_button.py``; the E2E proves the user-visible loop, not
git plumbing (the phase's recorded ASSUMPTION);
* **Stale surfaced** — ``/history.html``: the row now carries the rose
Stale pill (``aria-label``d for screen readers); ``GET /api/chats``
and ``GET /api/chats/<id>`` carry ``stale: true`` (computed
server-side — the client never computes staleness);
* **Regenerate** — opening the row (``/?chat=<id>``, the same URL the
History table links) reveals ``#stale-banner`` with the Regenerate
button; clicking it re-streams the last answer IN PLACE against the
new index (the phase-49 redo-in-place: the old bubble leaves the DOM,
the question is not duplicated), and the handler then auto re-saves
the linked row — the server re-stamps ``sources_version`` → the
banner clears, ``GET /api/chats/<id>`` reports ``stale: false`` with
the last brain message being the fresh answer, and the History pill
is gone;
* **Guard** — a stale row whose conversation has NO brain record
(user-only) reveals the banner text WITHOUT the Regenerate button
(``retryLastTurn`` has nothing to redo);
* **Anonymous** — sharing the (now fresh) chat and opening
``/shared/<token>`` without a session renders the phase-51 snapshot
with NO staleness surface — neither in the DOM nor on the public
``SharedChatOut`` wire (the snapshot is frozen by design).
DB isolation: the shared e2e Postgres keeps ``saved_chats`` rows across
suites, so every test here uses a DISTINCTIVE question text (its
auto-title is therefore unique), never asserts on absolute row counts,
and deletes the rows it creates in a ``finally`` (admin cookie). The KB
tables are truncated + re-seeded the house way (deterministic mock
embeddings); ``saved_chats`` and ``sources_meta`` are never touched by
the reset — the version is monotonic by design, and rows stamped
against an older generation are simply stale (that is the point).
Determinism: the mock quotes the asked question into its grounded
answer (ending in the ``Deterministic mock answer for E2E`` marker),
so the regenerated answer is textually identical to the stale one —
OLD-vs-fresh bubble identity is proved the phase-49 way, with a
test-only ``data-retry-marker`` attribute set on the old wrap before
the click.
"""
from __future__ import annotations
import asyncio
import time
import uuid
from pathlib import Path
from threading import Thread
from typing import Any
import httpx
from playwright.sync_api import Browser, BrowserContext, Page, expect
from sqlalchemy import text
from app.config import Settings
from app.db import SessionLocal
from app.models import SavedChat
from app.rag.importer import ImportSummary, import_sources
from app.rag.llm import LLMClient
from app.rag.sources_meta import bump_sources_version, current_sources_version
from e2e.auth_helpers import login
REPO = Path(__file__).resolve().parents[2]
FIXTURES = REPO / "tests" / "fixtures" / "docs"
MOCK_ANSWER_MARKER = "Deterministic mock answer for E2E"
#: The banner's line (frontend/index.html, task 05).
BANNER_TEXT = "The sources have been updated since this chat was saved."
#: The live-region outcome of a successful Regenerate (app.js, task 05).
REGEN_STATUS = "Regenerated — the answer now reflects the current sources."
#: The fresh row's Stale cell (history.js, task 04).
FRESH_STALE_CELL = "—"
#: The stale row's Stale cell aria-label (history.js, task 04).
STALE_CELL_ARIA = "Stale — sources have changed since this chat was saved"
#: The typing indicator is itself a .msg.brain — exclude its bubble.
ANSWER = ".msg.brain .bubble:not(.typing)"
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 — deterministic
mock answers), then optionally re-import fixtures. ``saved_chats``
and ``sources_meta`` are deliberately NOT touched: rows persist
across suites (every test cleans up after itself) and the version
is monotonic (the invalidation marker)."""
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 _bump_sources_version() -> int:
"""The test-only stand-in for a KB-changing sync (task 02): the SAME
``bump_sources_version`` helper both real sync paths call
change-gated, over a short session (bump flushes, the caller
commits). The bump GATES — changed/unchanged/`--limit`/failed — are
pinned by this phase's integration tests, not by the E2E."""
with SessionLocal() as db:
v = bump_sources_version(db)
db.commit()
return v
def _current_version() -> int:
with SessionLocal() as db:
return current_sources_version(db)
def _row_stamp(chat_id: str) -> int:
"""The saved row's ``sources_version`` stamp (helper read — the
same pattern as the phase's integration tests)."""
with SessionLocal() as db:
row = db.get(SavedChat, uuid.UUID(chat_id))
assert row is not None
return row.sources_version
def _ask(page: Page, question: str) -> None:
"""Send one turn and wait until the grounded answer has fully
landed (the ``done`` event restored the Send button)."""
page.fill("#message-input", question)
page.click("#send-btn")
expect(page.locator(".msg.user .bubble").last).to_contain_text(question)
expect(page.locator(ANSWER).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 _wait_saved_row(
app_url: str,
cookies: dict[str, str],
title: str,
messages: int = 2,
) -> dict[str, Any]:
"""Wait for the auto-saved row (phase 55: auto-saves are SILENT —
A2 — so there is no status line to wait on). The upsert is
fire-and-forget from the UI's point of view, so poll the admin
list until the row with the conversation's auto-title appears with
the expected message count."""
deadline = time.monotonic() + 15
last: dict[str, Any] | None = None
while time.monotonic() < deadline:
last = _find_row(_chats(app_url, cookies), title)
if last is not None and last["message_count"] >= messages:
return last
time.sleep(0.2)
raise AssertionError(f"no auto-saved row for {title!r} (last: {last!r})")
def _admin_cookies(page: Page) -> dict[str, str]:
"""The signed session cookies the browser holds after a form login —
used to call the admin API with plain httpx (the test's API side
sees exactly what the signed-in browser sees)."""
return {
c["name"]: c["value"]
for c in page.context.cookies()
if "name" in c and "value" in c
}
def _chats(app_url: str, cookies: dict[str, str]) -> list[dict[str, Any]]:
r = httpx.get(f"{app_url}/api/chats", timeout=10, cookies=cookies)
assert r.status_code == 200
return r.json()["chats"]
def _chat(app_url: str, cookies: dict[str, str], chat_id: str) -> dict[str, Any]:
r = httpx.get(f"{app_url}/api/chats/{chat_id}", timeout=10, cookies=cookies)
assert r.status_code == 200
return r.json()
def _auto_title(question: str) -> str:
"""The phase-50 auto-title convention: the first question,
whitespace-collapsed, capped at 120 chars."""
return " ".join(question.split())[:120]
def _find_row(
rows: list[dict[str, Any]], title: str
) -> dict[str, Any] | None:
return next((c for c in rows if c["title"] == title), None)
def _delete_chat(app_url: str, cookies: dict[str, str], chat_id: str) -> None:
"""Best-effort row cleanup (a 404 — already deleted — is fine)."""
httpx.delete(f"{app_url}/api/chats/{chat_id}", timeout=10, cookies=cookies)
def _history_row(page: Page, chat_id: str) -> Any:
"""The History table row for this chat (located through the Open
link — ``/?chat=<id>`` — the same URL the table links)."""
return page.locator(
"#history-tbody tr", has=page.locator(f"a[href='/?chat={chat_id}']")
)
def _assert_no_error_banner(page: Page) -> None:
"""Regenerate settles through the normal done path — never the red
role=alert error banner (the KB-offline banner is a separate,
health-driven state the db_ready fixture keeps away)."""
banner = page.locator("#kb-banner")
expect(banner).to_be_hidden()
expect(banner).not_to_have_class("is-error")
def _tag_last_brain_wrap(page: Page, marker: str) -> None:
"""Tag the rendered wrap of the LAST brain bubble (settled state —
no typing indicator present) so the test can prove the OLD element
leaves the DOM: the mock answers are byte-stable, so the redo of
the same question is textually indistinguishable from the original."""
page.evaluate(
"""(marker) => {
const wraps = document.querySelectorAll("#messages > .msg.brain");
wraps[wraps.length - 1].setAttribute("data-retry-marker", marker);
}""",
marker,
)
def _regenerate_in_place(page: Page, marker: str) -> None:
"""Click the banner's Regenerate and wait for the FULL phase-49
redo-in-place + auto re-save to settle: the turn goes in-flight
(the button IS Stop), the old bubble leaves the DOM before the
first fresh token, the fresh answer lands, and the handler's
post-turn PUT (the server re-stamps ``sources_version``) clears the
banner and lands the outcome on the live region."""
_tag_last_brain_wrap(page, marker)
regen = page.locator("#stale-regenerate")
expect(regen).to_be_enabled()
regen.click()
# In flight: the redo owns the Send/Stop control…
expect(page.locator("#send-label")).to_have_text("Stop", timeout=5_000)
# …and the double-click guard holds (one regenerate at a time).
expect(regen).to_be_disabled()
# Redo in place: the OLD wrap is already gone (the phase-49
# contract — removal precedes the rerun).
expect(page.locator(f"[data-retry-marker='{marker}']")).to_have_count(0)
# The fresh answer streams into its place and the turn settles.
expect(page.locator(ANSWER).last).to_contain_text(MOCK_ANSWER_MARKER, timeout=30_000)
expect(page.locator("#send-label")).to_have_text("Send", timeout=30_000)
_assert_no_error_banner(page)
# Post-turn: the linked row was re-saved (the server re-stamped
# it), so the banner cleared and the outcome is announced.
expect(page.locator("#stale-banner")).to_be_hidden(timeout=15_000)
expect(page.locator("#send-status")).to_have_text(REGEN_STATUS)
# ---------------------------------------------------------------------------
# 1. The full loop: save (fresh) → KB change → stale surfaced → Regenerate
# → fresh again (UI + API agree at every step)
# ---------------------------------------------------------------------------
def test_full_invalidation_loop(
page: Page, app_url: str, mock_llm: int, db_ready: None
) -> None:
_reset_db(mock_llm, seed=True)
page.set_default_timeout(30_000)
login(page, app_url, next="/")
expect(page).to_have_url(app_url + "/", timeout=30_000)
q = "How is my Kubernetes cluster set up? (stale-loop)"
_ask(page, q)
# --- Auto-save (fresh, phase 55 — no Save pill): the API agrees,
# and a fresh open shows NO banner.
cookies = _admin_cookies(page)
row = _wait_saved_row(app_url, cookies, _auto_title(q))
assert row["message_count"] == 2
assert row["stale"] is False, "a just-saved chat is fresh, not stale"
chat_id: str = row["id"]
# The DB agrees: Save stamped the row with the CURRENT generation
# (helper read — the integration suite pins the same contract).
assert _row_stamp(chat_id) == _current_version(), (
"Save must stamp sources_version with the current generation"
)
try:
# A fresh row opened at /?chat=<id> shows NO stale banner.
page.goto(app_url + f"/?chat={chat_id}")
expect(page.locator(".msg.user .bubble")).to_have_count(1)
expect(page.locator(".msg.user .bubble")).to_contain_text(q)
expect(page.locator(ANSWER)).to_have_count(1)
expect(page.locator("#stale-banner")).to_be_hidden()
expect(page.locator("#stale-regenerate")).to_be_hidden()
# --- KB change: the version bumps (the test-only stand-in for a
# KB-changing sync — see the module docstring / task ASSUMPTION).
bumped = _bump_sources_version()
assert bumped > 0
# The API agrees the row is stale NOW (server-computed — both
# the detail and the list shapes carry the flag):
detail = _chat(app_url, cookies, chat_id)
assert detail["stale"] is True
mine = _find_row(_chats(app_url, cookies), _auto_title(q))
assert mine is not None and mine["stale"] is True
# --- History: the row carries the rose Stale pill (and its
# aria-label, so the marker is conveyed without the visual).
page.goto(app_url + "/history.html")
tr = _history_row(page, chat_id)
expect(tr.locator("a.history-title-link")).to_be_visible(timeout=15_000)
pill = tr.locator(".stale-pill")
expect(pill).to_have_count(1)
expect(pill).to_have_text("Stale")
expect(tr.locator(".history-stale-cell")).to_have_attribute(
"aria-label", STALE_CELL_ARIA
)
# --- Open the row (the SAME URL the History table links): the
# banner reveals, with the Regenerate button.
tr.locator("a.history-title-link").click()
banner = page.locator("#stale-banner")
expect(banner).to_be_visible(timeout=15_000)
expect(banner).to_have_attribute("role", "status")
expect(banner).to_contain_text(BANNER_TEXT)
expect(page.locator("#stale-regenerate")).to_have_text("Regenerate")
# The conversation restored in full (one question, one answer).
expect(page.locator(".msg.user .bubble")).to_have_count(1)
expect(page.locator(ANSWER)).to_have_count(1)
# --- Regenerate: the fresh answer streams in place against the
# new index, and the handler re-saves the linked row.
_regenerate_in_place(page, "stale-old")
# Redo mechanics: the question was NOT duplicated, the fresh
# answer stands alone in the last bubble (it quotes the
# question — the mock is deterministic).
expect(page.locator(".msg.user .bubble")).to_have_count(1)
expect(page.locator(ANSWER)).to_have_count(1)
expect(page.locator(ANSWER).last).to_contain_text(q)
expect(page.locator("#stale-regenerate")).to_be_enabled()
# --- The API agrees the row is fresh again, with the fresh
# answer as its last brain message (the re-save re-stamped the
# row — the DB agrees).
detail = _chat(app_url, cookies, chat_id)
assert detail["stale"] is False, "the re-saved row must be fresh"
msgs = detail["messages"]
assert [m["who"] for m in msgs] == ["user", "brain"], (
"the re-save must not duplicate the question"
)
assert msgs[0]["text"] == q
assert MOCK_ANSWER_MARKER in msgs[-1]["text"]
assert q in msgs[-1]["text"], "the fresh answer quotes the re-asked question"
assert _row_stamp(chat_id) == _current_version(), (
"the Regenerate re-save must re-stamp sources_version"
)
# --- History: the Stale pill is gone (the em-dash returns).
page.goto(app_url + "/history.html")
tr = _history_row(page, chat_id)
expect(tr.locator("a.history-title-link")).to_be_visible(timeout=15_000)
expect(tr.locator(".stale-pill")).to_have_count(0)
expect(tr.locator(".history-stale-cell")).to_have_text(FRESH_STALE_CELL)
finally:
_delete_chat(app_url, cookies, chat_id)
# ---------------------------------------------------------------------------
# 2. Guard: a stale chat with NO brain record reveals the banner text
# WITHOUT the Regenerate button (nothing to regenerate)
# ---------------------------------------------------------------------------
def test_stale_chat_without_brain_answer_is_text_only(
page: Page, app_url: str, db_ready: None
) -> None:
page.set_default_timeout(30_000)
js_errors: list[str] = []
page.on("pageerror", lambda e: js_errors.append(str(e)))
login(page, app_url, next="/")
expect(page).to_have_url(app_url + "/", timeout=30_000)
cookies = _admin_cookies(page)
# A user-only saved conversation — the same wire shape the Save pill
# posts (auto-title from the first user message).
q = "How is my Kubernetes cluster set up? (stale-nobrain)"
r = httpx.post(
f"{app_url}/api/chats",
timeout=10,
cookies=cookies,
json={"messages": [{"who": "user", "text": q}]},
)
assert r.status_code == 201, r.text
chat_id = r.json()["id"]
assert r.json()["stale"] is False # freshly stamped on create
try:
_bump_sources_version()
assert _chat(app_url, cookies, chat_id)["stale"] is True
# Opening the stale user-only row: the banner reveals…
page.goto(app_url + f"/?chat={chat_id}")
banner = page.locator("#stale-banner")
expect(banner).to_be_visible(timeout=15_000)
expect(banner).to_contain_text(BANNER_TEXT)
# …but the Regenerate button is REMOVED — with no brain record
# there is nothing to regenerate (retryLastTurn is never called).
expect(page.locator("#stale-regenerate")).to_have_count(0)
# The conversation restored: the user question, no brain bubble.
expect(page.locator(".msg.user .bubble")).to_have_count(1)
expect(page.locator(".msg.user .bubble")).to_contain_text(q)
expect(page.locator(ANSWER)).to_have_count(0)
assert not js_errors, f"the text-only banner must not throw: {js_errors}"
finally:
_delete_chat(app_url, cookies, chat_id)
# ---------------------------------------------------------------------------
# 3. Anonymous: the shared (now fresh) chat renders at /shared/<token>
# with NO staleness surface — neither in the DOM nor on the public
# SharedChatOut wire (phase 51 unchanged)
# ---------------------------------------------------------------------------
def test_anonymous_shared_snapshot_has_no_staleness_surface(
page: Page,
browser: Browser,
app_url: str,
mock_llm: int,
db_ready: None,
) -> None:
_reset_db(mock_llm, seed=True)
page.set_default_timeout(30_000)
login(page, app_url, next="/")
expect(page).to_have_url(app_url + "/", timeout=30_000)
q = "How is my Kubernetes cluster set up? (stale-share)"
_ask(page, q)
cookies = _admin_cookies(page)
row = _wait_saved_row(app_url, cookies, _auto_title(q)) # auto-saved (phase 55)
chat_id: str = row["id"]
anon_ctx: BrowserContext | None = None
try:
# Make the chat stale, then bring it back CURRENT through the
# full Regenerate loop — the chat shared below is fresh.
_bump_sources_version()
page.goto(app_url + f"/?chat={chat_id}")
expect(page.locator("#stale-banner")).to_be_visible(timeout=15_000)
_regenerate_in_place(page, "stale-share-old")
assert _chat(app_url, cookies, chat_id)["stale"] is False
# Share from the History row's Share column (Create link → the
# cell re-renders to Copy + Unshare).
page.goto(app_url + "/history.html")
tr = _history_row(page, chat_id)
create = tr.locator("button.history-share-create")
expect(create).to_be_visible(timeout=15_000)
create.click()
expect(tr.locator("button.history-share-copy")).to_be_visible(timeout=15_000)
share_url = _chat(app_url, cookies, chat_id).get("share_url")
assert share_url is not None, "the Create link must have shared the row"
token = share_url.removeprefix("/shared/")
# --- The FRESH anonymous context (no session): the phase-51
# snapshot renders in full…
anon_ctx = browser.new_context()
anon = anon_ctx.new_page()
anon.set_default_timeout(30_000)
anon.goto(app_url + share_url)
expect(anon.locator("#shared-title")).to_have_text(_auto_title(q))
expect(anon.locator(".msg.user .bubble")).to_have_count(1)
expect(anon.locator(".msg.user .bubble")).to_contain_text(q)
expect(anon.locator(".msg.brain .bubble").first).to_contain_text(
MOCK_ANSWER_MARKER, timeout=30_000
)
# …with NO staleness surface anywhere: none of the phase-53
# elements exist on the shared page, and the banner's line never
# appears (the snapshot is frozen by design — phase 51). (The
# check is on the UI strings, not the bare word: the question
# text is quoted into the mock's answer and is data, not a
# staleness surface.)
stale_selectors = (
"#stale-banner, #stale-regenerate, "
".stale-banner, .stale-pill, .stale-regenerate"
)
expect(anon.locator(stale_selectors)).to_have_count(0)
body_text = anon.locator("body").inner_text().lower()
assert "sources have been updated since this chat was saved" not in body_text, (
"the shared snapshot must carry no staleness surface"
)
anon_ctx.close()
anon_ctx = None
# …and the public wire agrees: the anonymous read carries
# title + messages ONLY (no stale flag, no staleness surface).
r = httpx.get(f"{app_url}/api/shared/{token}", timeout=10)
assert r.status_code == 200
body = r.json()
assert set(body) == {"title", "messages"}, (
f"the public snapshot must stay minimal: {set(body)}"
)
assert "stale" not in body
assert body["title"] == _auto_title(q)
assert MOCK_ANSWER_MARKER in body["messages"][-1]["text"]
finally:
if anon_ctx is not None:
anon_ctx.close()
_delete_chat(app_url, cookies, chat_id)
Reference in New Issue View Git Blame Copy Permalink