bvandeusen/FabledScribe
1
0
Fork 0
Code Issues 1 Pull Requests 1 Actions Packages Projects Releases 53 Wiki Activity

docs(6205): knowledge comments state the reason, not the instance that taught it

Browse Source 
The ACL header states the share behaviour without its date and decision ids;
the facet table, door map, floor constant and facet-count comments state why
they are single sources instead of retelling how copies drifted; this
instance's counts (legacy plan-tasks, issue share) are removed.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
bvandeusen
2026-10-21 20:22:42 -04:00
co-authored by Claude Opus 5.5
parent d1b442c572
commit 21a409eac1
1 changed files with 42 additions and 52 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/scribe/services/knowledge.py
+42 -52
View File
@@ -1,8 +1,7 @@
"""Knowledge service — one query across every record kind Scribe holds.
ACL (rules #47/#78, decision note 2094): these queries were owner-only until
2026-07-25, which meant a record shared with you could be opened by id but never
*found*. They now honour shares — at two different widths:
ACL (services/access.py): these queries honour shares, so a record shared with
you can be *found* as well as opened by id — at two different widths:
- **searching** (a `q` the caller typed) uses the full read scope, so a record
shared directly with you is findable when you go looking for it. BOTH halves
@@ -32,7 +31,7 @@ logger = logging.getLogger(__name__)
_SNIPPET_LEN = 200
# --- the location filter (reverse lookup, #2083) ------------------------------
# --- the location filter (reverse lookup) ------------------------------------
#
# ONE predicate in two dialects: "this record carries a location matching every
# part asked for." The SQL form runs in the browse arm and the keyword arm,
@@ -79,7 +78,7 @@ def location_matches(data: dict | None, parts: dict[str, str]) -> bool:
_path_matches((loc.get(key) or "").strip(), want)
if key == "path"
# Repo names are recorded free-form ("Scribe" / "FabledScribe" /
# "fabledscribe") — case is never the distinguishing thing (#2874).
# "fabledscribe") — case is never the distinguishing thing.
else (loc.get(key) or "").strip().lower() == want.lower()
if key == "repo"
else (loc.get(key) or "").strip() == want
@@ -106,7 +105,7 @@ def location_jsonpath(parts: dict[str, str]) -> str:
prefix = json.dumps(want.rstrip("/") + "/")
filters.append(f"(@.path == {literal} || @.path starts with {prefix})")
elif key == "repo":
# Case-insensitive, anchored, regex-escaped (#2874) — mirrors the
# Case-insensitive, anchored, regex-escaped — mirrors the
# Python dialect's .lower() compare.
pattern = json.dumps("^" + re.escape(want) + "$")
filters.append(f'(@.repo like_regex {pattern} flag "i")')
@@ -120,7 +119,7 @@ def _location_clause(parts: dict[str, str]):
return Note.data.path_exists(location_jsonpath(parts))
# --- drift-check filter (#2086) ----------------------------------------------
# --- drift-check filter ------------------------------------------------------
# `verification` selects on the drift-check verdict stored in `data.verification`
# (see services/snippets.py). Statuses are the service's own constants; the two
# composite values are what the operator actually asks for.
@@ -154,7 +153,7 @@ def verification_matches(data: dict | None, value: str) -> bool:
if not status:
return want == "unverified"
expired = verdict.get("code_sha") != (data or {}).get("code_sha")
# A push touched the recorded location since the verdict (#2691): the repo
# A push touched the recorded location since the verdict: the repo
# moved under it, so it needs a look even though it hasn't failed.
invalidated = bool(verdict.get("invalidated_by"))
if want == "unverified":
@@ -171,7 +170,7 @@ def verification_matches(data: dict | None, value: str) -> bool:
_VERIFY_DRIFTED_JSONPATH = '$.verification ? (@.status != "ok")'
_VERIFY_EXPIRED_JSONPATH = "$ ? (@.verification.code_sha != @.code_sha)"
_VERIFY_ANY_JSONPATH = "$.verification"
# A push touched the recorded location since the verdict (#2691).
# A push touched the recorded location since the verdict.
_VERIFY_INVALIDATED_JSONPATH = "$.verification.invalidated_by"
@@ -189,7 +188,7 @@ def _verification_clause(value: str):
return Note.data.path_exists(_VERIFY_DRIFTED_JSONPATH)
if want == "attention":
# Everything worth looking at: a failing verdict, an expired one, OR
# one whose recorded location a push has since touched (#2691).
# one whose recorded location a push has since touched.
return or_(
Note.data.path_exists(_VERIFY_DRIFTED_JSONPATH),
and_(has_verdict, Note.data.path_exists(_VERIFY_EXPIRED_JSONPATH)),
@@ -221,10 +220,9 @@ def _note_to_item(note: Note, chunks: dict[int, dict] | None = None) -> dict:
any other — it shows the record's opening. `snippet_is` says which, on
every row, so the two never have to be told apart by guessing.
This was `(note.body or "")[:_SNIPPET_LEN]` on every row alike: the head of
the document, on the human's main search surface, with nothing marking the
cut. A record matched on its sixth paragraph was shown its first, which the
search had already scored lower (#4243).
The head of the body on every row alike (`body[:_SNIPPET_LEN]`) would show
a record matched on its sixth paragraph its first, which the search had
already scored lower, with nothing marking the cut.
"""
item: dict = {
"id": note.id,
@@ -244,7 +242,7 @@ def _note_to_item(note: Note, chunks: dict[int, dict] | None = None) -> dict:
"created_at": iso(note.created_at),
"updated_at": iso(note.updated_at),
}
# Drift verdict (#2086), when one has been recorded. Included here rather
# Drift verdict, when one has been recorded. Included here rather
# than decorated on by the snippet layer because `current` is derivable from
# `data` alone — the verdict's code_sha against the row's — so this needs no
# body parsing and stays a plain projection of the column. Omitted entirely
@@ -253,7 +251,7 @@ def _note_to_item(note: Note, chunks: dict[int, dict] | None = None) -> dict:
# Snippet language, same reasoning as the verdict below: a plain projection of
# the `data` mirror, no body parsing. Needed because a prior-art hit in a
# DIFFERENT language than the file being written is useful as the shape of a
# solution but must not be mistaken for code to paste (#2244) — and the caller
# solution but must not be mistaken for code to paste — and the caller
# can only say "different" if the language is on the item.
language = (note.data or {}).get("language") if note.data else None
if language:
@@ -267,8 +265,8 @@ def _note_to_item(note: Note, chunks: dict[int, dict] | None = None) -> dict:
trigger = (note.data or {}).get("when_to_apply") if note.data else None
if trigger:
item["when_to_apply"] = trigger
# A snippet's, for the same reason — and since milestone 427 the title no
# longer carries it, so without this a list shows names with no situation.
# A snippet's, for the same reason — the stored title does not carry it,
# so without this a list shows names with no situation.
usage = (note.data or {}).get("when_to_use") if note.data else None
if usage:
item["when_to_use"] = usage
@@ -283,7 +281,7 @@ def _note_to_item(note: Note, chunks: dict[int, dict] | None = None) -> dict:
"path": verdict.get("path"),
}
# Present only when a push has touched the recorded location since the
# verdict (#2691) — the "recheck me" marker, cleared by re-verifying.
# verdict — the "recheck me" marker, cleared by re-verifying.
if verdict.get("invalidated_by"):
item["verification"]["invalidated_by"] = verdict["invalidated_by"]
@@ -305,17 +303,15 @@ def _note_to_item(note: Note, chunks: dict[int, dict] | None = None) -> dict:
# why this is a table rather than a chain of ifs. Each entry is
# (is_task, the value pinned on that axis); None pins nothing, i.e. every task.
#
# It is a table because the alternative had already gone wrong. The predicate
# was written three times — a SQL if-chain, a Python if-chain over semantic
# candidates, and a ternary computing the `is_task` pre-filter — and the three
# only agreed by luck. Adding `issue` to the SQL arm alone (the obvious edit,
# and the one #3128 was about to make) would have set the pre-filter to
# is_task=False, handed the Python arm a candidate set containing no tasks at
# all, and returned an empty semantic half for the Issues facet forever, with
# nothing red anywhere. A new facet is now one row here.
# It is a table because the predicate has three dialects — SQL, a Python test
# over semantic candidates, and the `is_task` pre-filter — and three hand-kept
# copies agree only by luck. Adding `issue` to the SQL arm alone would set the
# pre-filter to is_task=False, hand the Python arm a candidate set containing
# no tasks at all, and return an empty semantic half for the Issues facet
# forever, with nothing red anywhere. A new facet is one row here.
#
# `plan` is retired (0066) but kept: 90 legacy plan-tasks exist and a facet
# they answer to costs one line. It simply has no chip in the UI any more.
# `plan` is retired but kept: legacy plan-tasks may exist, and a facet they
# answer to costs one line. It has no chip in the UI.
_FACETS: dict[str, tuple[bool, str | None]] = {
"task": (True, None),
"work": (True, "work"),
@@ -326,11 +322,10 @@ _FACETS: dict[str, tuple[bool, str | None]] = {
"process": (False, "process"),
"snippet": (False, "snippet"),
# A transferable insight, keyed by the situation it applies to rather than
# by its topic (milestone 385). It reaches the browse surface through this
# one entry: the door's validation, the counts and both dialects of the
# type filter are all generated from this table, which is the property
# #3161 asked for so that adding a kind is a single edit rather than four
# that must agree.
# by its topic. It reaches the browse surface through this one entry: the
# door's validation, the counts and both dialects of the type filter are
# all generated from this table, so adding a kind is a single edit rather
# than four that must agree.
"lesson": (False, "lesson"),
}
@@ -358,7 +353,7 @@ def facet_is_task(note_type: str | None) -> bool | None:
Used to narrow the semantic candidate set before it is fetched. Reads the
same table `_apply_type_filter` and `matches_facet` read, so the pre-filter
can no longer disagree with the predicate it is meant to anticipate.
cannot disagree with the predicate it is meant to anticipate.
"""
if not note_type:
return None
@@ -388,7 +383,7 @@ def search_filters_for(facet: str) -> dict:
narrow and no fetched row to test — it is passing filters INTO
`semantic_search_notes`. `_apply_type_filter` is the SQL dialect and
`matches_facet` the Python one; all three read `_FACETS`, which is what
keeps "adding a kind" a single edit (#3161).
keeps "adding a kind" a single edit.
Returns kwargs rather than a tuple so a caller splats it and cannot pair
`task_kind` with `is_task=False` by writing the positions out of order.
@@ -421,10 +416,9 @@ _CONTENT_TYPE_BROAD = {
def content_type_filters(content_type: str, extra: tuple = ()) -> dict:
"""The search kwargs a door's `content_type` parameter implies.
Lives here, beside `_FACETS`, because a door that keeps its own map is how
this went wrong: the agent's search offered two kinds and browse offered
nine, and a third copy in `/api/search` offered two more quietly still
(#4250). Derived, so adding a kind stays one edit (#3161).
Lives here, beside `_FACETS`, because a door that keeps its own map drifts:
one door offers two kinds while another offers nine. Derived, so adding a
kind stays one edit.
Raises on anything unrecognised. `_facet` alone would fall back to reading
an unknown string as a note_type, which matches no row — so a typo comes
@@ -611,18 +605,15 @@ async def _semantic_knowledge_search(
query=q,
report=_semantic_report,
limit=min(200, limit * 4),
# The shared interactive floor — this was a bare `0.3` while
# routes/search.py had the same number as a commented constant, the
# exact pair where one moves and the other doesn't (#2463).
# The shared interactive floor, one constant, so no copy of the
# number can move without the others.
threshold=INTERACTIVE_SEARCH_THRESHOLD,
is_task=is_task_filter,
project_id=project_id,
)
# The human's MAIN search surface, and it logged nothing — so
# retrieval_logs claimed the web's search was /api/search, a narrower
# path with (measured) zero frontend consumers. Distinct source, so
# The human's MAIN search surface, logged under its own source so
# threshold tuning can include or exclude human queries deliberately
# rather than by accident (#2463).
# rather than by accident.
from scribe.services.retrieval_telemetry import record_retrieval
record_retrieval(
user_id=user_id, source="browse_search", query=q,
@@ -698,11 +689,10 @@ async def get_knowledge_counts(user_id: int, tags: list[str] | None = None) -> d
stmt = stmt.where(Note.tags.contains([tag]))
return stmt
# One grouped query per typing axis. The task half used to be a count
# for 'task' plus a second count for 'plan', which is why 'issue' —
# 17% of every task here — had no number to show: each kind needed its
# own query and nobody added one. Grouping by task_kind counts every
# kind, including ones added later, for the same two round-trips.
# One grouped query per typing axis. A count per kind needs a new
# query for every kind and leaves a kind nobody added with no number;
# grouping by task_kind counts every kind, including ones added later,
# in two round-trips.
non_task = _scoped(
select(Note.note_type, func.count(Note.id))
.where(Note.status.is_(None))