Files
FabledScribe/tests/test_rule_via_lesson.py
T
bvandeusenandClaude Opus 5.5 d2fa723373
CI & Build / Python lint (push) Successful in 3s
CI & Build / Plugin hooks (push) Successful in 14s
CI & Build / TypeScript typecheck (push) Successful in 53s
CI & Build / integration (push) Successful in 1m1s
CI & Build / Python tests (push) Successful in 1m54s
CI & Build / Build & push image (push) Successful in 28s
refactor(retrieval): the rule builders compose one moment - _rule_moment runs the arm and its via-lesson step for all three (milestone 456 step 6, #4908)
build_prompt_rule_hint, build_tool_rule_hint and build_write_path_hint
each ran the rule arm and then the via-lesson step by hand, the first two
passing the query between them through a _via_query key on the payload.
They now call one composer, _rule_moment, and the split helpers
(_prompt_rule_hint, _tool_rule_hint, _add_rules_via_lessons) and the side
channel are deleted. Output shapes are unchanged: the prompt builder
returns no checkpoint key, the tool builder always does, and
shown_rule_ids stays the direct band.

A structural test pins that only _rule_moment runs a rule arm or the
via-lesson step, and that the three builders are its only callers.

plugin_context.py: 2,418 -> 2,361 lines this step; 3,558 when the
milestone began.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-05 20:25:47 -04:00

182 lines
7.4 KiB
Python

"""A rule reached through its lessons (milestone 440, #4633).
The step runs after each of the three rule arms. These pin, with the database
and the embedder stubbed: that nothing is searched when no lesson carries a
confirmed link; that a matching linked lesson brings its rule in rule voice,
naming the lesson; that suppression is by RULE; that the slot holds one line;
and that the call is logged under its own source. That a SUGGESTED link never
expands — the soft link proving itself — is pinned against Postgres in
tests/test_integration_lesson_rule_links.py, where the state filter lives.
"""
from __future__ import annotations
from types import SimpleNamespace
from unittest.mock import AsyncMock, MagicMock, patch
import pytest
from scribe.services import plugin_context as pc
from scribe.services import rule_usage
from scribe.services.retrieval_registry import POINTS
from scribe.services.retrieval_pipeline import RuleMoment, RuleResult
# Bound before conftest's autouse stub replaces the module attribute.
_REAL = pc._rules_via_lessons
def _lesson(lid=41, title="An overrun run usually failed early"):
return SimpleNamespace(
id=lid, title=title, note_type="lesson", body="",
data={"what": title, "when_to_apply": "a CI run overran"},
)
def _rule(rid, title="Read the job log first", kind="rule"):
return SimpleNamespace(id=rid, title=title, kind=kind, when_to_apply="a run overran")
def _stubs(*, confirmed, found, by_lesson):
log, surfaced = MagicMock(), MagicMock()
stack = [
patch.object(pc.lesson_rules_svc, "confirmed_lessons", AsyncMock(return_value=confirmed)),
patch.object(pc.lesson_rules_svc, "confirmed_rules_in_scope", AsyncMock(return_value=by_lesson)),
patch.object(pc, "get_autoinject_config", AsyncMock(return_value={"threshold": 0.5})),
patch.object(pc, "semantic_search_notes", AsyncMock(return_value=found)),
patch.object(pc, "record_retrieval", log),
patch.object(pc, "record_rule_surfaced", surfaced),
]
return stack, log, surfaced
async def _run(stack, **kw):
for p in stack:
p.start()
try:
return await _REAL(1, "the CI run is still going", project_id=2,
skip=kw.get("skip", set()), held=set(), where="to this request")
finally:
for p in stack:
p.stop()
@pytest.mark.asyncio
async def test_no_confirmed_link_means_no_search_and_no_row():
search = AsyncMock()
stack, log, _ = _stubs(confirmed=set(), found=[], by_lesson={})
stack[3] = patch.object(pc, "semantic_search_notes", search)
assert await _run(stack) == ([], [])
search.assert_not_awaited()
log.assert_not_called()
@pytest.mark.asyncio
async def test_a_matching_linked_lesson_brings_its_rule_in_rule_voice():
stack, log, surfaced = _stubs(
confirmed={41}, found=[(0.71, _lesson())], by_lesson={41: [_rule(7)]},
)
lines, ids = await _run(stack)
assert ids == [7]
assert lines[0].startswith("Standing rule")
assert "Reached through lesson #41" in lines[0]
assert log.call_args.kwargs["source"] == "rule_via_lesson"
assert surfaced.call_args.kwargs == {"user_id": 1, "rule_ids": [7], "source": "rule_via_lesson"}
@pytest.mark.asyncio
async def test_a_lesson_without_a_confirmed_link_carries_nothing():
"""The search found a lesson, but it is not in the confirmed set — a
suggested or unlinked lesson brings no rule."""
stack, log, surfaced = _stubs(
confirmed={99}, found=[(0.9, _lesson(41))], by_lesson={},
)
assert await _run(stack) == ([], [])
surfaced.assert_not_called()
assert log.call_args.kwargs["results"] == []
@pytest.mark.asyncio
async def test_suppression_is_by_rule_whichever_lesson_reached_it():
stack, log, surfaced = _stubs(
confirmed={41}, found=[(0.71, _lesson())], by_lesson={41: [_rule(7)]},
)
assert await _run(stack, skip={7}) == ([], [])
assert log.call_args.kwargs["suppressed"] == 1
surfaced.assert_not_called()
@pytest.mark.asyncio
async def test_the_slot_holds_one_line():
stack, _log, _ = _stubs(
confirmed={41, 42},
found=[(0.8, _lesson(41)), (0.7, _lesson(42, "Another"))],
by_lesson={41: [_rule(7), _rule(8)], 42: [_rule(9)]},
)
lines, ids = await _run(stack)
assert ids == [7] and len(lines) == pc.VIA_LESSON_LIMIT == 1
@pytest.mark.asyncio
async def test_a_failure_brings_no_rule_and_raises_nothing():
stack, _log, _ = _stubs(confirmed={41}, found=[], by_lesson={})
stack[0] = patch.object(pc.lesson_rules_svc, "confirmed_lessons",
AsyncMock(side_effect=RuntimeError("db down")))
assert await _run(stack) == ([], [])
@pytest.mark.asyncio
async def test_the_step_follows_the_arm_on_its_query_and_skips_what_it_named():
"""One composer for all three rule builders (milestone 456 step 6): the
via-lesson step searches with the query the arm searched with, skips the
session's ledger and every rule the arm's band named, and its lines and
fresh ids join the arm's — while `shown_rule_ids` stays the direct band."""
step = AsyncMock(return_value=(["Standing rule … Reached through lesson #41"], [7]))
direct = RuleResult(lines=["direct line"], rule_ids=[5], shown_rule_ids=[5, 6])
moment = RuleMoment(user_id=1, query="q", project_id=2, where="here",
exclude=frozenset({3, 6}))
with patch.object(pc, "_rules_via_lessons", step), \
patch.object(pc.rp, "run_rule_arm", AsyncMock(return_value=direct)):
out = await pc._rule_moment(pc.rp.PROMPT_RULE, moment, floor=0.5, budget=3)
assert step.await_args.args[1] == "q"
assert step.await_args.kwargs["skip"] == {3, 5, 6}
assert out.rule_ids == [5, 7]
assert out.lines == ["direct line", "Standing rule … Reached through lesson #41"]
assert out.shown_rule_ids == [5, 6]
@pytest.mark.asyncio
async def test_a_blank_prompt_runs_neither_the_arm_nor_the_step():
step = AsyncMock(return_value=([], []))
with patch.object(pc, "_rules_via_lessons", step):
out = await pc.build_prompt_rule_hint(1, " ")
assert out == {"context": "", "rule_ids": []}
step.assert_not_awaited()
def test_only_the_composer_runs_a_rule_arm_or_the_via_lesson_step():
"""Structural (rule 167): every rule builder goes through `_rule_moment`,
so a builder that runs the arm or the via-lesson step by hand again is the
copy this step removed coming back — and a count is what lets it fail."""
import ast
from pathlib import Path
tree = ast.parse(Path("src/scribe/services/plugin_context.py").read_text())
callers: dict[str, set[str]] = {}
for fn in ast.walk(tree):
if not isinstance(fn, (ast.FunctionDef, ast.AsyncFunctionDef)):
continue
for call in ast.walk(fn):
if isinstance(call, ast.Call):
name = getattr(call.func, "attr", None) or getattr(call.func, "id", None)
if name in ("run_rule_arm", "_rules_via_lessons", "_rule_moment"):
callers.setdefault(name, set()).add(fn.name)
assert callers["run_rule_arm"] == {"_rule_moment"}
assert callers["_rules_via_lessons"] == {"_rule_moment"}
assert callers["_rule_moment"] == {
"build_prompt_rule_hint", "build_tool_rule_hint", "build_write_path_hint",
}
def test_the_source_is_ranked_and_registered():
assert "rule_via_lesson" in rule_usage.RANKED_SOURCES
assert "rule_via_lesson" in POINTS