fix(embeddings): embed in the service, so every caller gets it (#2056)
CI & Build / Python lint (push) Successful in 3s
CI & Build / Plugin hooks (push) Successful in 8s
CI & Build / integration (push) Successful in 15s
CI & Build / TypeScript typecheck (push) Successful in 33s
CI & Build / Python tests (push) Successful in 57s
CI & Build / Build & push image (push) Successful in 29s
CI & Build / Python lint (push) Successful in 3s
CI & Build / Plugin hooks (push) Successful in 8s
CI & Build / integration (push) Successful in 15s
CI & Build / TypeScript typecheck (push) Successful in 33s
CI & Build / Python tests (push) Successful in 57s
CI & Build / Build & push image (push) Successful in 29s
A note or task created through MCP was not semantically searchable until the next restart's backfill ran. Embedding fired at the five REST route handlers and nowhere else; the MCP tools call the service directly, so they skipped it. The shape of this bug is the reason to care: it is invisible on an instance that redeploys constantly (this one does, per rule 46) and permanent on one that doesn't. Rule 115 — the product has to stand up for the install that restarts twice a year, not just for the one that restarts hourly. Moved to services/notes.embed_note(), called from create_note and update_note, and deleted from all five routes. Every caller — REST, MCP, recurrence, snippets — now gets it by construction rather than by remembering. Two things fall out of having one implementation instead of six: - It uses note.user_id, the OWNER. The routes were inconsistent: some passed the caller's uid, some the owner's. On a shared record the caller's id mints a second embedding row that nothing reads. - services/snippets.py's _embed_snippet existed only because snippets are created via MCP and the routes couldn't cover them. Every one of its four call sites goes through notes_svc, so the helper and its four calls are gone, along with the eight test patches that existed to neutralise it. RuntimeError (no running loop — unit tests, scripts) and any indexing failure are both swallowed: a write that succeeded must not be failed by its index. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01UaYUaouG9jjhATyuxCKrQs
This commit is contained in:
@@ -0,0 +1,64 @@
|
||||
"""services/notes.py — the inline embedding moved here from the routes (#2056).
|
||||
|
||||
Why it moved: embedding was triggered at each REST route and nowhere else, so a
|
||||
record created through MCP stayed out of semantic search and auto-inject until
|
||||
the next restart's backfill. Putting it in the service means every caller — REST,
|
||||
MCP, recurrence, snippets — gets it by construction rather than by remembering.
|
||||
|
||||
These test the helper directly. The point of the change is that there is now ONE
|
||||
place to test.
|
||||
"""
|
||||
from scribe.services import notes as notes_svc
|
||||
|
||||
# --- inline embedding (#2056) -----------------------------------------------
|
||||
|
||||
def test_embed_note_uses_the_OWNER_not_the_caller():
|
||||
"""LOAD-BEARING for shared records. An embedding belongs to the record; a
|
||||
collaborator editing a shared note must refresh the owner's row rather than
|
||||
mint a second one under their own id. The routes this replaced passed the
|
||||
caller's uid on one path and the owner's on another — exactly the kind of
|
||||
inconsistency that moving it to one place removes."""
|
||||
from unittest.mock import MagicMock, patch
|
||||
|
||||
note = MagicMock(id=5, user_id=42, title="T", body="B")
|
||||
with patch("scribe.services.embeddings.upsert_note_embedding") as upsert, \
|
||||
patch("asyncio.create_task") as create_task:
|
||||
notes_svc.embed_note(note)
|
||||
|
||||
assert create_task.called
|
||||
upsert.assert_called_once()
|
||||
assert upsert.call_args.args[0] == 5
|
||||
assert upsert.call_args.args[1] == 42 # owner, never the caller
|
||||
assert upsert.call_args.args[2] == "T\nB"
|
||||
|
||||
|
||||
def test_embed_note_skips_a_record_with_no_text():
|
||||
"""An empty embedding is worse than none — it is a row that matches nothing
|
||||
and hides the fact that the record was never indexed."""
|
||||
from unittest.mock import MagicMock, patch
|
||||
|
||||
note = MagicMock(id=5, user_id=42, title="", body="")
|
||||
with patch("asyncio.create_task") as create_task:
|
||||
notes_svc.embed_note(note)
|
||||
assert not create_task.called
|
||||
|
||||
|
||||
def test_embed_note_without_a_running_loop_is_not_an_error():
|
||||
"""Unit tests and scripts call create_note with no event loop. That must be
|
||||
an ordinary case: a write that succeeded cannot be failed by its index
|
||||
refresh."""
|
||||
from unittest.mock import MagicMock, patch
|
||||
|
||||
note = MagicMock(id=5, user_id=42, title="T", body="B")
|
||||
with patch("asyncio.create_task", side_effect=RuntimeError("no running loop")):
|
||||
notes_svc.embed_note(note) # must not raise
|
||||
|
||||
|
||||
def test_embed_note_swallows_an_indexing_failure():
|
||||
"""Same reason, wider net: the embedding model being unavailable must not
|
||||
turn a successful save into a 500."""
|
||||
from unittest.mock import MagicMock, patch
|
||||
|
||||
note = MagicMock(id=5, user_id=42, title="T", body="B")
|
||||
with patch("asyncio.create_task", side_effect=ValueError("model gone")):
|
||||
notes_svc.embed_note(note) # must not raise
|
||||
@@ -40,8 +40,7 @@ async def test_editor_share_can_update_and_writes_as_the_owner():
|
||||
with patch.object(svc, "get_snippet", AsyncMock(return_value=note)), \
|
||||
patch("scribe.services.access.can_write_note", AsyncMock(return_value=True)), \
|
||||
patch.object(svc.notes_svc, "update_note",
|
||||
AsyncMock(return_value=updated)) as mock_update, \
|
||||
patch.object(svc, "_embed_snippet", MagicMock()):
|
||||
AsyncMock(return_value=updated)) as mock_update:
|
||||
got = await svc.update_snippet(7, 1, name="formatDuration")
|
||||
|
||||
assert got is updated
|
||||
@@ -107,8 +106,7 @@ async def test_merge_skips_sources_owned_by_someone_else():
|
||||
with patch.object(svc, "get_snippet", AsyncMock(side_effect=fake_get)), \
|
||||
patch("scribe.services.access.can_write_note", AsyncMock(return_value=True)), \
|
||||
patch.object(svc.notes_svc, "update_note", AsyncMock(return_value=target)), \
|
||||
patch("scribe.services.trash.delete", AsyncMock(return_value=object())), \
|
||||
patch.object(svc, "_embed_snippet", MagicMock()):
|
||||
patch("scribe.services.trash.delete", AsyncMock(return_value=object())):
|
||||
_note, merged_ids = await svc.merge_snippets(7, 1, [2, 3])
|
||||
|
||||
assert merged_ids == [2]
|
||||
|
||||
@@ -39,8 +39,7 @@ async def _run_merge(target, sources, source_ids):
|
||||
patch("scribe.services.access.can_write_note", AsyncMock(return_value=True)), \
|
||||
patch.object(s.notes_svc, "update_note",
|
||||
AsyncMock(return_value=target)) as mock_update, \
|
||||
patch("scribe.services.trash.delete", AsyncMock(return_value=object())), \
|
||||
patch.object(s, "_embed_snippet", MagicMock()):
|
||||
patch("scribe.services.trash.delete", AsyncMock(return_value=object())):
|
||||
await s.merge_snippets(7, target.id, source_ids)
|
||||
return mock_update.await_args.kwargs
|
||||
|
||||
@@ -88,8 +87,7 @@ async def test_an_ordinary_edit_carries_provenance_forward():
|
||||
with patch.object(s, "get_snippet", AsyncMock(return_value=note)), \
|
||||
patch("scribe.services.access.can_write_note", AsyncMock(return_value=True)), \
|
||||
patch.object(s.notes_svc, "update_note",
|
||||
AsyncMock(return_value=note)) as mock_update, \
|
||||
patch.object(s, "_embed_snippet", MagicMock()):
|
||||
AsyncMock(return_value=note)) as mock_update:
|
||||
await s.update_snippet(7, 1, signature="f(ms) -> string")
|
||||
|
||||
kwargs = mock_update.await_args.kwargs
|
||||
@@ -106,8 +104,7 @@ async def test_a_pre_0070_row_keeps_provenance_through_the_body():
|
||||
with patch.object(s, "get_snippet", AsyncMock(return_value=note)), \
|
||||
patch("scribe.services.access.can_write_note", AsyncMock(return_value=True)), \
|
||||
patch.object(s.notes_svc, "update_note",
|
||||
AsyncMock(return_value=note)) as mock_update, \
|
||||
patch.object(s, "_embed_snippet", MagicMock()):
|
||||
AsyncMock(return_value=note)) as mock_update:
|
||||
await s.update_snippet(7, 1, when_to_use="humanize a ms count")
|
||||
|
||||
assert "**Merged from:** #5" in mock_update.await_args.kwargs["body"]
|
||||
|
||||
@@ -44,7 +44,6 @@ async def _run_unmerge(survivor, *, source_alive=None, restore=1):
|
||||
patch("scribe.services.trash.restore_entity", AsyncMock(return_value=restore)),
|
||||
patch.object(s.notes_svc, "update_note",
|
||||
AsyncMock(return_value=survivor)) as upd,
|
||||
patch.object(s, "_embed_snippet", MagicMock()),
|
||||
):
|
||||
await s.unmerge_snippet(7, 1, 2)
|
||||
return upd.await_args.kwargs
|
||||
@@ -127,7 +126,6 @@ async def test_a_purged_source_is_refused_and_the_survivor_is_untouched():
|
||||
patch("scribe.services.access.can_write_note", AsyncMock(return_value=True)),
|
||||
patch("scribe.services.trash.restore_entity", AsyncMock(return_value=None)),
|
||||
patch.object(s.notes_svc, "update_note", AsyncMock()) as upd,
|
||||
patch.object(s, "_embed_snippet", MagicMock()),
|
||||
):
|
||||
with pytest.raises(s.UnmergeError, match="purged"):
|
||||
await s.unmerge_snippet(7, 1, 2)
|
||||
@@ -150,7 +148,6 @@ async def test_an_entry_without_attribution_is_refused_not_guessed():
|
||||
patch.object(s, "get_snippet", AsyncMock(side_effect=fake_get)),
|
||||
patch("scribe.services.access.can_write_note", AsyncMock(return_value=True)),
|
||||
patch.object(s.notes_svc, "update_note", AsyncMock()) as upd,
|
||||
patch.object(s, "_embed_snippet", MagicMock()),
|
||||
):
|
||||
with pytest.raises(s.UnmergeError, match="provenance"):
|
||||
await s.unmerge_snippet(7, 1, 2)
|
||||
|
||||
Reference in New Issue
Block a user