iomgaa/mdpolish
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 1 Wiki Activity

feat: 增加项目无关的本地清洗评审器

Browse Source
This commit is contained in:
Bepr4
2026-08-28 19:06:10 +08:00
parent 10c026c7ad
commit 6c0dd5974b
82 changed files with 9335 additions and 48 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/test_review_parsing.py
+297
View File
@@ -0,0 +1,297 @@
from __future__ import annotations
from json import dumps, loads
from typing import cast
import pytest
from mdpolish import (
DocumentSnapshot,
Modifier,
Pipeline,
ProposedChange,
RunStatus,
TextEdit,
TextSpan,
)
from mdpolish.review import (
ReviewDetail,
ReviewDocument,
ReviewParseError,
build_review_document,
parse_json_report,
render_json_report,
)
def _replace_modifier(
needle: str,
replacement: str,
*,
modifier_id: str,
reason: str = "应用虚构测试替换",
) -> Modifier:
def propose(snapshot: DocumentSnapshot) -> tuple[ProposedChange, ...]:
start = snapshot.markdown.find(needle)
if start < 0:
return ()
return (
ProposedChange(
snapshot_sha256=snapshot.sha256,
reason=reason,
edits=(
TextEdit(
snapshot_sha256=snapshot.sha256,
span=TextSpan(start, start + len(needle)),
expected_text=needle,
replacement=replacement,
),
),
),
)
return Modifier(
modifier_id=modifier_id,
version="1.0.0",
parameters={"needle": needle, "replacement": replacement},
applicability="只处理虚构 JSON reader 测试标记。",
propose=propose,
)
def _noop_modifier(modifier_id: str = "test.noop") -> Modifier:
return Modifier(
modifier_id=modifier_id,
version="1.0.0",
parameters=(),
applicability="不修改任何文本。",
propose=lambda _snapshot: (),
)
def _success_review() -> ReviewDocument:
input_markdown = "\ufeff首行\r\n🙂Cafe\u0301\r末行"
result = Pipeline(
(
_noop_modifier(),
_replace_modifier("🙂Cafe\u0301", "完成", modifier_id="test.unicode"),
)
).transform(input_markdown)
assert result.status is RunStatus.SUCCESS
return build_review_document(input_markdown, result)
def _unstable_review() -> ReviewDocument:
residual_marker = "RESIDUAL_SECRET"
def residual(snapshot: DocumentSnapshot) -> tuple[ProposedChange, ...]:
start = snapshot.markdown.find(residual_marker)
if start < 0:
return ()
return (
ProposedChange(
snapshot_sha256=snapshot.sha256,
reason="残留虚构候选",
edits=(
TextEdit(
snapshot_sha256=snapshot.sha256,
span=TextSpan(start, start + len(residual_marker)),
expected_text=residual_marker,
replacement="resolved",
),
),
),
)
result = Pipeline(
(
Modifier(
modifier_id="test.residual",
version="1.0.0",
parameters=(),
applicability="只在最终复查产生候选。",
propose=residual,
),
_replace_modifier("start", residual_marker, modifier_id="test.producer"),
)
).transform("start")
assert result.status is RunStatus.UNSTABLE
return build_review_document("start", result)
def _error_reviews() -> tuple[ReviewDocument, ...]:
preflight_result = Pipeline(cast(tuple[Modifier, ...], (object(),))).transform("preflight")
def explode(_snapshot: DocumentSnapshot) -> tuple[ProposedChange, ...]:
raise RuntimeError("DIAGNOSTIC_SECRET")
transform_result = Pipeline(
(
Modifier(
modifier_id="test.transform-error",
version="1.0.0",
parameters=(),
applicability="只测试执行错误。",
propose=explode,
),
)
).transform("transform")
def explode_on_done(snapshot: DocumentSnapshot) -> tuple[ProposedChange, ...]:
if snapshot.markdown == "done":
raise RuntimeError("FINAL_DIAGNOSTIC_SECRET")
return ()
final_result = Pipeline(
(
Modifier(
modifier_id="test.final-error",
version="1.0.0",
parameters=(),
applicability="只测试最终复查错误。",
propose=explode_on_done,
),
_replace_modifier("start", "done", modifier_id="test.final-producer"),
)
).transform("start")
return (
build_review_document("preflight", preflight_result),
build_review_document("transform", transform_result),
build_review_document("start", final_result),
)
@pytest.mark.parametrize("detail", tuple(ReviewDetail))
def test_reader_accepts_every_official_detail(detail: ReviewDetail) -> None:
review = _success_review()
report = render_json_report(review, detail=detail)
parsed = parse_json_report(report, expected_detail=detail)
assert parsed == loads(report)
assert parsed["detail"] == detail.value
def test_reader_accepts_unstable_summary_without_materializing_residuals() -> None:
parsed = parse_json_report(render_json_report(_unstable_review(), detail="summary"))
assert parsed["status"] == "unstable"
assert "residual_proposals" not in parsed
@pytest.mark.parametrize("review", _error_reviews())
@pytest.mark.parametrize("detail", tuple(ReviewDetail))
def test_reader_accepts_each_failed_stage(review: ReviewDocument, detail: ReviewDetail) -> None:
parsed = parse_json_report(render_json_report(review, detail=detail))
assert parsed["status"] == "failed"
def test_reader_accepts_compatible_minor_and_unknown_object_field() -> None:
payload = loads(render_json_report(_success_review(), detail="full"))
payload["schema_version"] = "1.9"
payload["future_summary"] = {"available": True}
payload["hash_contract"]["future_hash_note"] = "compatible"
payload["coordinate_contract"]["future_coordinate_note"] = "compatible"
parsed = parse_json_report(dumps(payload, ensure_ascii=False))
assert parsed["schema_version"] == "1.9"
assert parsed["future_summary"] == {"available": True}
def test_reader_rejects_stage_after_that_was_not_produced_by_changes() -> None:
payload = loads(render_json_report(_success_review(), detail="full"))
stage = payload["stages"][1]
forged = "伪造结果"
forged_hash = DocumentSnapshot(forged).sha256
stage["after"] = {
"sha256": forged_hash,
"code_point_length": len(forged),
"markdown": forged,
}
stage["changes"][0]["after_sha256"] = forged_hash
payload["current"] = {
"sha256": forged_hash,
"code_point_length": len(forged),
"markdown": forged,
}
with pytest.raises(ReviewParseError, match="replayed changes do not produce"):
parse_json_report(dumps(payload, ensure_ascii=False), expected_detail="full")
def test_reader_rejects_changed_zero_change_stage() -> None:
payload = loads(render_json_report(_success_review(), detail="full"))
stage = payload["stages"][0]
forged = "改变了零修改阶段"
forged_hash = DocumentSnapshot(forged).sha256
stage["after"] = {
"sha256": forged_hash,
"code_point_length": len(forged),
"markdown": forged,
}
with pytest.raises(ReviewParseError, match="zero-change stage"):
parse_json_report(dumps(payload, ensure_ascii=False), expected_detail="full")
def test_reader_rejects_conflicting_change_batch() -> None:
payload = loads(render_json_report(_success_review(), detail="full"))
stage = payload["stages"][1]
duplicate = dict(stage["changes"][0])
duplicate["edit_index"] = 1
stage["changes"].append(duplicate)
stage["change_count"] = 2
payload["counts"]["change_count"] = 2
with pytest.raises(ReviewParseError, match="edit contract"):
parse_json_report(dumps(payload, ensure_ascii=False), expected_detail="full")
def test_reader_rejects_wrong_location_without_leaking_text() -> None:
payload = loads(render_json_report(_success_review(), detail="full"))
change = payload["stages"][1]["changes"][0]
change["location"] = {"line": 99, "column": 99}
with pytest.raises(ReviewParseError) as error:
parse_json_report(dumps(payload, ensure_ascii=False), expected_detail="full")
assert "🙂Cafe\u0301" not in str(error.value)
assert "location does not match" in str(error.value)
@pytest.mark.parametrize(
("report", "message"),
(
('{"schema_name":"a","schema_name":"b"}', "keys must be unique"),
('{"value":NaN}', "non-finite"),
('{"schema_name":"mdpolish.review","schema_version":"2.0"}', "schema major"),
(dumps({"schema_name": "\ud800"}), "valid Unicode"),
("not JSON", "malformed JSON"),
),
)
def test_reader_rejects_malformed_or_unsupported_json(report: str, message: str) -> None:
with pytest.raises(ReviewParseError, match=message):
parse_json_report(report)
def test_reader_rejects_detail_mismatch_and_invalid_expected_detail() -> None:
report = render_json_report(_success_review(), detail="summary")
with pytest.raises(ReviewParseError, match="does not match"):
parse_json_report(report, expected_detail="full")
with pytest.raises(ReviewParseError, match="expected_detail"):
parse_json_report(report, expected_detail="unknown")
def test_reader_enforces_body_exposure_boundaries() -> None:
summary = loads(render_json_report(_success_review(), detail="summary"))
summary["future"] = {"markdown": "SUMMARY_SECRET"}
with pytest.raises(ReviewParseError, match="body-bearing"):
parse_json_report(dumps(summary, ensure_ascii=False))
changes = loads(render_json_report(_success_review(), detail="changes"))
changes["future"] = {"markdown": "CHANGES_SECRET"}
with pytest.raises(ReviewParseError, match="full markdown"):
parse_json_report(dumps(changes, ensure_ascii=False))