feat: 增加项目无关的本地清洗评审器

2026-08-28 19:06:10 +08:00
parent 10c026c7ad
commit 6c0dd5974b
82 changed files with 9335 additions and 48 deletions
@@ -0,0 +1,303 @@
+from __future__ import annotations
+
+import http.client
+import json
+import threading
+from pathlib import Path
+from typing import cast
+
+import pytest
+
+from mdpolish import DocumentSnapshot, Modifier, Pipeline, ProposedChange, TextEdit, TextSpan
+from mdpolish.review import build_review_document, render_json_report
+from mdpolish.reviewer import ReviewDataError, ReviewRepository, create_server, main
+
+
+def _object(value: object) -> dict[str, object]:
+    assert isinstance(value, dict)
+    return cast(dict[str, object], value)
+
+
+def _array(value: object) -> list[object]:
+    assert isinstance(value, list)
+    return cast(list[object], value)
+
+
+def _full_review_json(source: str = "😀 old\r\nlast line\r\n") -> str:
+    """生成包含 emoji、CRLF、一次修改和一个零修改阶段的虚构 full JSON。"""
+
+    def replace_old(snapshot: DocumentSnapshot) -> tuple[ProposedChange, ...]:
+        start = snapshot.markdown.find("old")
+        if start < 0:
+            return ()
+        return (
+            ProposedChange(
+                snapshot_sha256=snapshot.sha256,
+                reason="替换虚构测试词",
+                edits=(
+                    TextEdit(
+                        snapshot_sha256=snapshot.sha256,
+                        span=TextSpan(start, start + 3),
+                        expected_text="old",
+                        replacement="new",
+                    ),
+                ),
+            ),
+        )
+
+    pipeline = Pipeline(
+        (
+            Modifier(
+                modifier_id="test.replace",
+                version="1.0.0",
+                parameters={"mode": "exact"},
+                applicability="只替换虚构测试词。",
+                propose=replace_old,
+            ),
+            Modifier(
+                modifier_id="test.zero",
+                version="1.0.0",
+                parameters={},
+                applicability="用于验证零修改阶段。",
+                propose=lambda _snapshot: (),
+            ),
+        )
+    )
+    result = pipeline.transform(source)
+    review = build_review_document(source, result)
+    return render_json_report(review, detail="full") + "\n"
+
+
+def _review_directory(tmp_path: Path, content: str | None = None) -> Path:
+    review_dir = tmp_path / "reviews"
+    review_dir.mkdir(parents=True)
+    (review_dir / "paper.review.json").write_text(
+        _full_review_json() if content is None else content,
+        encoding="utf-8",
+        newline="",
+    )
+    return review_dir
+
+
+def test_repository_exposes_collection_document_and_modifier_stages(tmp_path: Path) -> None:
+    repository = ReviewRepository(_review_directory(tmp_path))
+
+    collection = repository.collection_response()
+    document = repository.document("paper").document_response()
+    first_stage = repository.document("paper").stage_response(0)
+    zero_stage = repository.document("paper").stage_response(1)
+
+    assert collection["collection"] == {"label": "reviews", "status": "success"}
+    assert collection["summary"] == {
+        "document_count": 1,
+        "success_count": 1,
+        "failed_count": 0,
+        "unstable_count": 0,
+        "change_count": 1,
+    }
+    assert "input_markdown" not in json.dumps(collection)
+    assert document["input_markdown"] == "😀 old\r\nlast line\r\n"
+    assert document["current_markdown"] == "😀 new\r\nlast line\r\n"
+    modifiers = _array(document["modifiers"])
+    assert [_object(item)["modifier_id"] for item in modifiers] == ["test.replace", "test.zero"]
+    changes = _array(document["changes"])
+    # Python 码点 2 位于 emoji 后; CodeMirror UTF-16 下标因此是 3。
+    assert _object(changes[0])["span"] == {"start": 2, "end": 5}
+    assert _object(changes[0])["editor_range"] == {"start": 3, "end": 6}
+    assert first_stage["before_markdown"] == "😀 old\r\nlast line\r\n"
+    assert first_stage["after_markdown"] == "😀 new\r\nlast line\r\n"
+    assert zero_stage["before_markdown"] == zero_stage["after_markdown"]
+    assert zero_stage["changes"] == []
+
+
+def test_repository_handles_unicode_empty_text_and_file_labels(tmp_path: Path) -> None:
+    source = "中e\u0301😀\r\nold\rend"
+    repository = ReviewRepository(_review_directory(tmp_path, _full_review_json(source)))
+    document = repository.document("paper").document_response()
+    changes = _array(document["changes"])
+
+    assert _object(changes[0])["location"] == {"line": 2, "column": 1}
+    assert _object(changes[0])["span"] == {"start": 6, "end": 9}
+    assert _object(changes[0])["editor_range"] == {"start": 7, "end": 10}
+    assert document["current_markdown"] == "中e\u0301😀\r\nnew\rend"
+    assert _object(document["document"])["source_label"] == "paper"
+
+    empty_directory = tmp_path / "empty"
+    empty_directory.mkdir()
+    (empty_directory / "empty.review.json").write_text(
+        _full_review_json(""),
+        encoding="utf-8",
+        newline="",
+    )
+    empty_document = ReviewRepository(empty_directory).document("empty").document_response()
+    assert empty_document["input_markdown"] == ""
+    assert empty_document["current_markdown"] == ""
+
+
+@pytest.mark.parametrize(
+    "mutation",
+    (
+        lambda payload: payload.__setitem__("detail", "changes"),
+        lambda payload: payload["input"].__setitem__(
+            "markdown",
+            payload["input"]["markdown"].replace("old", "bad"),
+        ),
+        lambda payload: payload["stages"][0]["changes"][0]["span"].__setitem__("start", 0),
+    ),
+)
+def test_repository_rejects_untrusted_review_shapes(tmp_path: Path, mutation: object) -> None:
+    payload = json.loads(_full_review_json())
+    assert callable(mutation)
+    mutation(payload)
+    review_dir = _review_directory(tmp_path, json.dumps(payload, ensure_ascii=False) + "\n")
+
+    with pytest.raises(ReviewDataError) as raised:
+        ReviewRepository(review_dir)
+
+    assert raised.value.code == "invalid_review"
+    assert str(raised.value).startswith("paper.review.json：")
+
+
+def test_repository_accepts_compatible_schema_minor(tmp_path: Path) -> None:
+    payload = json.loads(_full_review_json())
+    payload["schema_version"] = "1.4"
+    payload["future_summary"] = {"available": True}
+
+    repository = ReviewRepository(
+        _review_directory(tmp_path, json.dumps(payload, ensure_ascii=False) + "\n")
+    )
+
+    assert repository.document("paper").status == "success"
+
+
+def test_repository_rejects_bom_invalid_utf8_and_symlinks(tmp_path: Path) -> None:
+    review_dir = _review_directory(tmp_path)
+    review_path = review_dir / "paper.review.json"
+    review_path.write_bytes(b"\xef\xbb\xbf" + review_path.read_bytes())
+    with pytest.raises(ReviewDataError, match="BOM"):
+        ReviewRepository(review_dir)
+
+    review_path.write_bytes(b'{"bad": "\xff"}')
+    with pytest.raises(ReviewDataError, match="UTF-8"):
+        ReviewRepository(review_dir)
+
+    review_path.unlink()
+    target = tmp_path / "outside.review.json"
+    target.write_text(_full_review_json(), encoding="utf-8", newline="")
+    review_path.symlink_to(target)
+    with pytest.raises(ReviewDataError) as raised:
+        ReviewRepository(review_dir)
+    assert raised.value.code == "unsafe_path"
+
+    review_path.unlink()
+    review_path.write_text(_full_review_json(), encoding="utf-8", newline="")
+    link = tmp_path / "review-link"
+    link.symlink_to(review_dir, target_is_directory=True)
+    with pytest.raises(ReviewDataError) as raised:
+        ReviewRepository(link)
+    assert raised.value.code == "unsafe_path"
+
+
+def test_repository_rejects_empty_directory_and_empty_label(tmp_path: Path) -> None:
+    empty = tmp_path / "empty"
+    empty.mkdir()
+    with pytest.raises(ReviewDataError) as raised:
+        ReviewRepository(empty)
+    assert raised.value.code == "empty_review_directory"
+
+    (empty / ".review.json").write_text(_full_review_json(), encoding="utf-8", newline="")
+    with pytest.raises(ReviewDataError, match="缺少文档标签"):
+        ReviewRepository(empty)
+
+
+def test_console_entry_reports_expected_startup_failure_without_traceback(tmp_path: Path) -> None:
+    empty = tmp_path / "empty"
+    empty.mkdir()
+
+    with pytest.raises(SystemExit) as raised:
+        main(["--review-dir", str(empty)])
+
+    assert str(raised.value) == "评审器启动失败：评审目录没有直属 full review JSON。"
+
+
+def test_http_server_is_loopback_read_only_and_does_not_leak_text(tmp_path: Path) -> None:
+    repository = ReviewRepository(_review_directory(tmp_path))
+    static_root = tmp_path / "static"
+    static_root.mkdir()
+    (static_root / "index.html").write_text("<!doctype html><title>review</title>", encoding="utf-8")
+    server = create_server(repository, static_root)
+    thread = threading.Thread(target=server.serve_forever, daemon=True)
+    thread.start()
+    host = cast(str, server.server_address[0])
+    port = server.server_address[1]
+
+    try:
+        assert host == "127.0.0.1"
+        connection = http.client.HTTPConnection(host, port)
+        connection.request("GET", "/api/v1/collection")
+        response = connection.getresponse()
+        body = response.read().decode("utf-8")
+        assert response.status == 200
+        assert response.getheader("Cache-Control") == "no-store"
+        assert response.getheader("Access-Control-Allow-Origin") is None
+        assert "😀 old" not in body
+
+        connection.request("POST", "/api/v1/collection")
+        response = connection.getresponse()
+        response.read()
+        assert response.status == 405
+
+        connection.request("GET", "/api/v1/documents/missing")
+        response = connection.getresponse()
+        response.read()
+        assert response.status == 404
+
+        connection.request("GET", "/api/v1/documents/paper/modifiers/99")
+        response = connection.getresponse()
+        response.read()
+        assert response.status == 404
+
+        connection.request("HEAD", "/")
+        response = connection.getresponse()
+        assert response.status == 200
+        assert response.read() == b""
+        assert response.getheader("Content-Security-Policy") is not None
+        assert response.getheader("X-Content-Type-Options") == "nosniff"
+
+        connection.putrequest("GET", "/api/v1/collection", skip_host=True)
+        connection.putheader("Host", "example.com")
+        connection.endheaders()
+        response = connection.getresponse()
+        response.read()
+        assert response.status == 403
+        connection.close()
+
+        connection = http.client.HTTPConnection(host, port)
+        connection.putrequest("GET", "/api/v1/collection")
+        connection.putheader("Origin", "http://example.com")
+        connection.endheaders()
+        response = connection.getresponse()
+        response.read()
+        assert response.status == 403
+        connection.close()
+    finally:
+        server.shutdown()
+        server.server_close()
+        thread.join(timeout=5)
+
+
+def test_server_rejects_missing_or_symlink_static_root(tmp_path: Path) -> None:
+    repository = ReviewRepository(_review_directory(tmp_path))
+    missing = tmp_path / "missing-static"
+    with pytest.raises(ReviewDataError) as raised:
+        create_server(repository, missing)
+    assert raised.value.code == "missing_build"
+
+    real = tmp_path / "real-static"
+    real.mkdir()
+    (real / "index.html").write_text("ok", encoding="utf-8")
+    link = tmp_path / "static-link"
+    link.symlink_to(real, target_is_directory=True)
+    with pytest.raises(ReviewDataError) as raised:
+        create_server(repository, link)
+    assert raised.value.code == "missing_build"