imacat/pop-fem-audit
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 1 Wiki Activity

Build the arbitration input with a per-ID extras merge

Browse Source 
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
This commit is contained in:
imacat
2026-08-17 22:38:33 +08:00
co-authored by Claude Opus 5
parent 1714257c89
commit 3bae24b930
4 changed files with 291 additions and 42 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tools/tests/test_compare_codings.py
+30 -7
View File
@@ -79,11 +79,20 @@ class TestCompareCodings(unittest.TestCase):
def __read_disagreements(self) -> dict[str, Any]:
"""Read the disagreement JSON file.
:return: The parsed disagreements.
:return: The parsed disagreements, as written, with every
song's keywords in its "disagreements" wrapper.
"""
return json.loads(
self.__disagreements_json.read_text(encoding="utf-8"))
def __read_keywords(self, song_id: str) -> dict[str, Any]:
"""Read one song's disagreed keywords from the file.
:param song_id: The song ID.
:return: The keywords of that song, unwrapped.
"""
return self.__read_disagreements()[song_id]["disagreements"]
def __read_disagreement_text(self) -> str:
"""Read the disagreement JSON file verbatim.
@@ -103,7 +112,23 @@ class TestCompareCodings(unittest.TestCase):
status, _ = self.__run_compare()
self.assertEqual(status, 0)
self.assertEqual(self.__read_disagreements(), {
"song-1": {"only-1": ["q2"], "only-2": ["q4"]}})
"song-1": {"disagreements": {
"only-1": ["q2"], "only-2": ["q4"]}}})
def test_keywords_wrapped_for_merging(self) -> None:
"""Test that each song's value is the object merged into
the arbitration input, holding "disagreements" alone."""
self.__write_codings(self.__run1, {
"song-1": {"only-1": ["q"]}})
self.__write_codings(self.__run2, {"song-1": {}})
status: int
status, _ = self.__run_compare()
self.assertEqual(status, 0)
self.assertEqual(
list(self.__read_disagreements()["song-1"].keys()),
["disagreements"])
self.assertEqual(
self.__read_keywords("song-1"), {"only-1": ["q"]})
def test_quotes_do_not_take_part_in_comparison(self) -> None:
"""Test that identical key sets with different quotes
@@ -162,7 +187,7 @@ class TestCompareCodings(unittest.TestCase):
status, _ = self.__run_compare()
self.assertEqual(status, 0)
self.assertEqual(
list(self.__read_disagreements()["song-1"].keys()),
list(self.__read_keywords("song-1").keys()),
["alpha", "beta", "mu", "zeta"])
def test_summary_line_counts(self) -> None:
@@ -195,8 +220,7 @@ class TestCompareCodings(unittest.TestCase):
status, _ = self.__run_compare()
self.assertEqual(status, 0)
self.assertEqual(
self.__read_disagreements(),
{"song-1": {"only-1": [quote]}})
self.__read_keywords("song-1"), {"only-1": [quote]})
def test_non_ascii_written_verbatim(self) -> None:
"""Test that the output is UTF-8 with the non-ASCII text
@@ -362,8 +386,7 @@ class TestCompareCodings(unittest.TestCase):
status, _ = self.__run_compare()
self.assertEqual(status, 0)
self.assertEqual(
self.__read_disagreements(),
{"song-1": {"only-1": ["q"]}})
self.__read_keywords("song-1"), {"only-1": ["q"]})
def test_compare_returns_agreed_keyword_names(self) -> None:
"""Test that the comparison function returns the agreed
tools/tests/test_export_llm_input.py
+128 -1
View File
@@ -76,16 +76,21 @@ class TestExportLlmInput(unittest.TestCase):
session.close()
def __run_export(
self, extras: Path | None = None) -> tuple[int, str]:
self, extras: Path | None = None,
extras_per_id: Path | None = None) -> tuple[int, str]:
"""Run the exporter with the standard error captured.
:param extras: The extras JSON file, or None for none.
:param extras_per_id: The per-ID extras JSON file, or None
for none.
:return: A tuple of the exit status and the standard
error.
"""
argv: list[str] = [str(self.__output)]
if extras is not None:
argv += ["--extras", str(extras)]
if extras_per_id is not None:
argv += ["--extras-per-id", str(extras_per_id)]
stderr: io.StringIO = io.StringIO()
with redirect_stderr(stderr):
status: int = export_llm_input.main(argv)
@@ -101,6 +106,16 @@ class TestExportLlmInput(unittest.TestCase):
path.write_text(text, encoding="utf-8")
return path
def __write_extras_per_id(self, text: str) -> Path:
"""Write a per-ID extras file with the given raw text.
:param text: The raw file content.
:return: The path of the written per-ID extras file.
"""
path: Path = self.__dir / "extras-per-id.json"
path.write_text(text, encoding="utf-8")
return path
@staticmethod
def __read_records(path: Path) -> list[dict[str, str]]:
"""Read the JSONL records of a file.
@@ -228,3 +243,115 @@ class TestExportLlmInput(unittest.TestCase):
self.assertEqual(status, 1)
self.assertIn("error:", stderr)
self.assertFalse(self.__output.exists())
def test_extras_per_id_merges_each_song_its_own(self) -> None:
"""Test that with per-ID extras, each record's content is
a JSON object of the lyrics followed by that song's own
keys in their file order."""
self.__seed([
("Hello", "Adele", "hello lyrics\n"),
("Umbrella", "Rihanna", "umbrella lyrics\n")])
per_id: Path = self.__write_extras_per_id(
'{"song-1": {"b": 2, "a": 1},'
' "song-2": {"disagreements": {"k": ["q"]}}}')
status: int
stderr: str
status, stderr = self.__run_export(None, per_id)
self.assertEqual(status, 0)
records: list[dict[str, str]] = self.__read_records(
self.__output)
self.assertEqual(len(records), 2)
first: dict[str, Any] = json.loads(records[0]["content"])
self.assertEqual(
list(first.keys()), ["lyrics", "b", "a"])
self.assertEqual(first["lyrics"], "hello lyrics\n")
self.assertEqual(first["b"], 2)
self.assertEqual(first["a"], 1)
second: dict[str, Any] = json.loads(records[1]["content"])
self.assertEqual(
list(second.keys()), ["lyrics", "disagreements"])
self.assertEqual(second["lyrics"], "umbrella lyrics\n")
self.assertEqual(
second["disagreements"], {"k": ["q"]})
def test_extras_per_id_restricts_the_export(self) -> None:
"""Test that the songs the per-ID extras do not name are
not exported."""
self.__seed([
("Hello", "Adele", "hello lyrics\n"),
("Umbrella", "Rihanna", "umbrella lyrics\n"),
("Halo", "Beyonce", "halo lyrics\n")])
per_id: Path = self.__write_extras_per_id(
'{"song-3": {"a": 1}, "song-1": {"a": 2}}')
status: int
stderr: str
status, stderr = self.__run_export(None, per_id)
self.assertEqual(status, 0)
records: list[dict[str, str]] = self.__read_records(
self.__output)
self.assertEqual(
[x["id"] for x in records], ["song-1", "song-3"])
self.assertIn("done: 2 songs exported", stderr)
def test_extras_per_id_unknown_id_fails(self) -> None:
"""Test that a per-ID extras file naming a song the
working store does not have is rejected."""
self.__seed([("Hello", "Adele", "hello lyrics\n")])
per_id: Path = self.__write_extras_per_id(
'{"song-1": {"a": 1}, "song-9": {"a": 2}}')
status: int
stderr: str
status, stderr = self.__run_export(None, per_id)
self.assertEqual(status, 1)
self.assertIn("song-9", stderr)
self.assertFalse(self.__output.exists())
def test_extras_per_id_after_the_shared_extras(self) -> None:
"""Test that with both extras options, a record's keys are
the lyrics, the shared extras, then that song's own."""
self.__seed([
("Hello", "Adele", "hello lyrics\n"),
("Umbrella", "Rihanna", "umbrella lyrics\n")])
extras: Path = self.__write_extras('{"shared": "s"}')
per_id: Path = self.__write_extras_per_id(
'{"song-2": {"own": "o"}}')
status: int
stderr: str
status, stderr = self.__run_export(extras, per_id)
self.assertEqual(status, 0)
records: list[dict[str, str]] = self.__read_records(
self.__output)
self.assertEqual([x["id"] for x in records], ["song-2"])
content: dict[str, Any] = json.loads(
records[0]["content"])
self.assertEqual(
list(content.keys()), ["lyrics", "shared", "own"])
self.assertEqual(content["lyrics"], "umbrella lyrics\n")
self.assertEqual(content["shared"], "s")
self.assertEqual(content["own"], "o")
def test_extras_per_id_non_object_value_fails(self) -> None:
"""Test that a per-ID extras file whose song does not have
a JSON object is rejected."""
self.__seed([("Hello", "Adele", "hello lyrics\n")])
per_id: Path = self.__write_extras_per_id(
'{"song-1": [1, 2]}')
status: int
stderr: str
status, stderr = self.__run_export(None, per_id)
self.assertEqual(status, 1)
self.assertIn("error:", stderr)
self.assertFalse(self.__output.exists())
def test_extras_per_id_with_lyrics_key_fails(self) -> None:
"""Test that a per-ID extras file whose song carries a
"lyrics" key is rejected."""
self.__seed([("Hello", "Adele", "hello lyrics\n")])
per_id: Path = self.__write_extras_per_id(
'{"song-1": {"lyrics": "not allowed"}}')
status: int
stderr: str
status, stderr = self.__run_export(None, per_id)
self.assertEqual(status, 1)
self.assertIn("error:", stderr)
self.assertFalse(self.__output.exists())