Repository navigation
Expand file tree
/
Copy pathconformance.py
More file actions
1134 lines (965 loc) · 36.8 KB
/
Copy pathconformance.py
File metadata and controls
1134 lines (965 loc) · 36.8 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
# /// script
# requires-python = ">=3.12"
# dependencies = []
#
# [tool.ty.rules]
# blanket-ignore-comment = "warn"
# missing-type-argument = "warn"
# possibly-unresolved-reference = "warn"
# unsound-return-statement = "warn"
# unsound-yield = "warn"
# unsupported-dynamic-base = "warn"
# division-by-zero = "warn"
#
# [tool.uv]
# no-build = true
# exclude-newer = "P7D"
# ///
"""
Run typing conformance tests and compare results between two ty versions.
ty versions can be supplied as `uvx ty` or `uvx ty@version`
for a specific version. This requires `uv` to be installed
and available on the system PATH.
If CONFORMANCE_SUITE_COMMIT is set, the hash will be used to create
links to the corresponding line in the conformance repository for each
diagnostic. Otherwise, it will default to `main'.
Examples:
# Compare two specific ty versions
%(prog)s --old-ty uvx [email protected] --new-ty uvx [email protected]
# Use local ty builds
%(prog)s --old-ty ./target/debug/ty-old --new-ty ./target/debug/ty-new
# Custom test directory
%(prog)s --target-path custom/tests --old-ty uvx [email protected] --new-ty uvx [email protected]
# Show a diff with local paths to the test directory instead of table of links
%(prog)s --old-ty uvx [email protected] --new-ty uvx [email protected] --format diff
"""
from __future__ import annotations
import argparse
import json
import os
import re
import subprocess
import sys
import tomllib
from collections import defaultdict
from collections.abc import Sequence, Set as AbstractSet
from dataclasses import dataclass
from enum import StrEnum, auto
from itertools import chain, groupby
from pathlib import Path
from textwrap import dedent
from typing import Any, Literal, Self, assert_never
# The conformance tests include 4 types of errors:
# 1. Required errors (E): The type checker must raise an error on this line
# 2. Optional errors (E?): The type checker may raise an error on this line
# 3. Tagged errors (E[tag]): The type checker must raise at most one error
# on a set of lines with a matching tag
# 4. Tagged multi-errors (E[tag+]): The type checker should raise one or
# more errors on a set of lines with a matching tag
CONFORMANCE_ERROR_PATTERN = re.compile(
r"""
\#\s*E # "# E" begins each error
(?P<optional>\?)? # Optional '?' (E?) indicates that an error is optional
(?: # An optional tag for errors that may appear on multiple lines at most once
\[
(?P<tag>[^+\]]+) # identifier
(?P<multi>\+)? # '+' indicates that an error may occur more than once on tagged lines
\]
)?
(?:
\s*:\s*(?P<description>.*) # optional description
)?
""",
re.VERBOSE,
)
CONFORMANCE_SUITE_COMMIT = os.environ.get("CONFORMANCE_SUITE_COMMIT", "main")
CONFORMANCE_DIR_WITH_README = (
f"https://github.com/python/typing/blob/{CONFORMANCE_SUITE_COMMIT}/conformance/"
)
CONFORMANCE_URL = CONFORMANCE_DIR_WITH_README + "tests/{filename}#L{line}"
GITHUB_HEADER = [
"<table>",
"<tr>",
"<th>Test case</th>",
"<th>Diff</th>",
"</tr>",
]
GITHUB_FOOTER = ["</table>"]
SUMMARY_NOTE = """
Each test case represents one expected error annotation or a group of annotations
sharing a tag. Counts are per test case, not per diagnostic — multiple diagnostics
on the same line count as one. Required annotations (`E`) are true positives when
ty flags the expected location and false negatives when it does not. Optional
annotations (`E?`) are true positives when flagged but true negatives (not false
negatives) when not. Tagged annotations (`E[tag]`) require ty to flag exactly one
of the tagged lines; tagged multi-annotations (`E[tag+]`) allow any number up to
the tag count. Flagging unexpected locations counts as a false positive.
"""
# Priority order for section headings: improvements first, regressions last.
TITLE_PRIORITY: dict[str, int] = {
"True positives added": 0,
"False positives removed": 1,
"True positives changed": 2,
"False positives changed": 3,
"False positives added": 4,
"True positives removed": 5,
"Optional Diagnostics Added": 6,
"Optional Diagnostics Removed": 7,
"Optional Diagnostics Changed": 8,
}
class Source(StrEnum):
OLD = auto()
NEW = auto()
class Classification(StrEnum):
TRUE_POSITIVE = auto()
FALSE_POSITIVE = auto()
TRUE_NEGATIVE = auto()
FALSE_NEGATIVE = auto()
def into_title(self, *, verb: Literal["added", "removed", "changed"]) -> str:
match self:
case Classification.TRUE_POSITIVE:
return f"True positives {verb}"
case Classification.FALSE_POSITIVE:
return f"False positives {verb}"
case Classification.TRUE_NEGATIVE:
return f"True negatives {verb}"
case Classification.FALSE_NEGATIVE:
return f"False negatives {verb}"
class Change(StrEnum):
ADDED = auto()
REMOVED = auto()
CHANGED = auto()
UNCHANGED = auto()
def into_title(self) -> str:
match self:
case Change.ADDED:
return "Optional Diagnostics Added"
case Change.REMOVED:
return "Optional Diagnostics Removed"
case Change.CHANGED:
return "Optional Diagnostics Changed"
case Change.UNCHANGED:
return "Optional Diagnostics Unchanged"
@dataclass(kw_only=True, slots=True)
class Position:
line: int
column: int
@dataclass(kw_only=True, slots=True)
class Positions:
begin: Position
end: Position
@dataclass(kw_only=True, slots=True)
class Location:
path: Path
positions: Positions
def as_link(self) -> str:
file = self.path.name
link = CONFORMANCE_URL.format(filename=file, line=self.positions.begin.line)
return f"[{file}:{self.positions.begin.line}:{self.positions.begin.column}]({link})"
@dataclass(kw_only=True, slots=True)
class TyDiagnostic:
"""A diagnostic emitted by a ty version (old or new) during a conformance run."""
check_name: str
description: str
severity: str
location: Location
source: Source
def __post_init__(self) -> None:
# Remove check name prefix from description
self.description = self.description.replace(f"{self.check_name}: ", "")
def __str__(self) -> str:
return (
f"{self.location.path}:{self.location.positions.begin.line}:"
f"{self.location.positions.begin.column}: "
f"{self.severity_for_display}[{self.check_name}] {self.description}"
)
@classmethod
def from_gitlab_output(
cls,
dct: dict[str, Any],
source: Source,
) -> Self:
return cls(
check_name=dct["check_name"],
description=dct["description"],
severity=dct["severity"],
location=Location(
path=Path(dct["location"]["path"]).resolve(),
positions=Positions(
begin=Position(
line=dct["location"]["positions"]["begin"]["line"],
column=dct["location"]["positions"]["begin"]["column"],
),
end=Position(
line=dct["location"]["positions"]["end"]["line"],
column=dct["location"]["positions"]["end"]["column"],
),
),
),
source=source,
)
@property
def severity_for_display(self) -> str:
return {
"major": "error",
"minor": "warning",
}.get(self.severity, "unknown")
@dataclass(kw_only=True, slots=True)
class ExpectedError:
"""An error annotation parsed from a conformance test file (e.g. ``# E: ...``)."""
description: str
location: Location
optional: bool
# tag identifying an error that can occur on multiple lines
tag: str | None
# True if one or more errors can occur on lines with the same tag
multi: bool
@property
def key(self) -> str:
"""Key to group expected errors by path and beginning line or path and tag."""
return (
f"{self.location.path.as_posix()}:{self.location.positions.begin.line}"
if self.tag is None
else f"{self.location.path.as_posix()}:{self.tag}"
)
def diagnostics_are_equivalent(a: list[TyDiagnostic], b: list[TyDiagnostic]) -> bool:
"""Compare two diagnostic lists for equality, ignoring the ``source`` field."""
def fingerprint(d: TyDiagnostic) -> tuple[str, str, str, str, int, int]:
return (
d.check_name,
d.description,
d.severity,
str(d.location.path),
d.location.positions.begin.line,
d.location.positions.begin.column,
)
return sorted(map(fingerprint, a)) == sorted(map(fingerprint, b))
@dataclass(kw_only=True, slots=True)
class TestCase:
key: str
old: list[TyDiagnostic]
new: list[TyDiagnostic]
expected: list[ExpectedError]
@property
def change(self) -> Change:
if self.new and not self.old:
return Change.ADDED
elif self.old and not self.new:
return Change.REMOVED
elif (
self.old and self.new and not diagnostics_are_equivalent(self.old, self.new)
):
return Change.CHANGED
else:
return Change.UNCHANGED
@property
def optional(self) -> bool:
return bool(self.expected) and all(e.optional for e in self.expected)
@property
def multi(self) -> bool:
return bool(self.expected) and all(e.multi for e in self.expected)
@property
def path(self) -> Path:
"""Return the source file path for this test case."""
for diags in (self.new, self.old, self.expected):
if diags:
return diags[0].location.path
raise ValueError(f"No diagnostics in test case {self.key}")
def diagnostics_by_source(self, source: Source) -> list[TyDiagnostic]:
return self.old if source == Source.OLD else self.new
def classify(self, source: Source) -> Classification:
diagnostics = self.diagnostics_by_source(source)
if diagnostics:
if self.optional:
return Classification.TRUE_POSITIVE
if self.expected:
distinct_lines = len(
{d.location.positions.begin.line for d in diagnostics}
)
expected_max = len(self.expected) if self.multi else 1
if 1 <= distinct_lines <= expected_max:
return Classification.TRUE_POSITIVE
else:
return Classification.FALSE_POSITIVE
else:
return Classification.FALSE_POSITIVE
elif self.expected:
if self.optional:
return Classification.TRUE_NEGATIVE
return Classification.FALSE_NEGATIVE
else:
return Classification.TRUE_NEGATIVE
def render_html_diff_row(tc: TestCase, *, source: Source | None) -> list[str]:
"""Render a single HTML <tr> with a test-case link and a markdown diff block."""
all_diags = tc.old + tc.new if source is None else tc.diagnostics_by_source(source)
if all_diags:
min_line = min(d.location.positions.begin.line for d in all_diags)
max_line = max(d.location.positions.begin.line for d in all_diags)
filename = all_diags[0].location.path.name
if min_line == max_line:
url = CONFORMANCE_URL.format(filename=filename, line=min_line)
display = f"{filename}:{min_line}"
else:
url = (
f"{CONFORMANCE_DIR_WITH_README}tests/{filename}#L{min_line}-L{max_line}"
)
display = f"{filename}:{min_line}:{max_line}"
location = f"[{display}]({url})"
else:
location = tc.key
diff_lines = []
if source is None:
for d in tc.old:
diff_lines.append(
f"-{d.severity_for_display}[{d.check_name}] {d.description}"
)
for d in tc.new:
diff_lines.append(
f"+{d.severity_for_display}[{d.check_name}] {d.description}"
)
else:
sign = "-" if source == Source.OLD else "+"
for d in tc.diagnostics_by_source(source):
diff_lines.append(
f"{sign}{d.severity_for_display}[{d.check_name}] {d.description}"
)
return [
"",
"<tr>",
"<td>",
"",
location,
"",
"</td>",
"",
"<td>",
"",
"```diff",
*diff_lines,
"```",
"",
"</td>",
"</tr>",
]
def render_diff_row(diagnostics: list[TyDiagnostic], *, removed: bool = False) -> str:
sign = "-" if removed else "+"
return "\n".join(f"{sign} {d}" for d in diagnostics)
@dataclass(kw_only=True, slots=True)
class Statistics:
true_positives: int = 0
false_positives: int = 0
false_negatives: int = 0
total_diagnostics: int = 0
@property
def precision(self) -> float:
if self.true_positives + self.false_positives > 0:
return self.true_positives / (self.true_positives + self.false_positives)
return 0.0
@property
def recall(self) -> float:
if self.true_positives + self.false_negatives > 0:
return self.true_positives / (self.true_positives + self.false_negatives)
else:
return 0.0
@dataclass(kw_only=True, slots=True)
class DiagnosticEntry:
"""A test case bound to the section title and source side it should be rendered under."""
title: str
test_case: TestCase
# None means show both old (removed) and new (added) in a single "changed" section.
source: Source | None
@dataclass(kw_only=True, slots=True)
class FileStats:
path: Path
old: Statistics
new: Statistics
@property
def old_passes(self) -> bool:
return self.old.false_positives == 0 and self.old.false_negatives == 0
@property
def new_passes(self) -> bool:
return self.new.false_positives == 0 and self.new.false_negatives == 0
@property
def total_change(self) -> int:
return (
abs(self.new.true_positives - self.old.true_positives)
+ abs(self.new.false_positives - self.old.false_positives)
+ abs(self.new.false_negatives - self.old.false_negatives)
)
def collect_expected_diagnostics(test_files: Sequence[Path]) -> list[ExpectedError]:
errors: list[ExpectedError] = []
for file in test_files:
for idx, line in enumerate(file.read_text().splitlines(), 1):
if match := re.search(CONFORMANCE_ERROR_PATTERN, line):
prefix = line[: match.start()].lstrip()
if not prefix or prefix.startswith("#"):
continue
errors.append(
ExpectedError(
description=(match.group("description") or "Missing"),
location=Location(
path=file,
positions=Positions(
begin=Position(
line=idx,
column=match.start(),
),
end=Position(
line=idx,
column=match.end(),
),
),
),
optional=match.group("optional") is not None,
tag=(
f"{file.name}:{match.group('tag')}"
if match.group("tag")
else None
),
multi=match.group("multi") is not None,
)
)
assert errors, "Failed to discover any expected diagnostics!"
return errors
def collect_ty_diagnostics(
ty_path: list[str],
source: Source,
test_files: Sequence[Path],
python_version: str = "3.12",
extra_search_paths: Sequence[Path] = (),
) -> list[TyDiagnostic]:
extra_search_path_args = [
f"--extra-search-path={path}" for path in extra_search_paths
]
process = subprocess.run(
[
*ty_path,
"check",
f"--python-version={python_version}",
"--output-format=gitlab",
"--ignore=assert-type-unspellable-subtype",
"--error=ambiguous-protocol-member",
"--error=invalid-enum-member-annotation",
"--error=invalid-legacy-positional-parameter",
"--error=mismatched-type-name",
"--error=invalid-named-tuple-override",
"--error=deprecated",
"--error=redundant-final-classvar",
"--exit-zero",
*extra_search_path_args,
*map(str, test_files),
],
capture_output=True,
text=True,
check=True,
timeout=15,
)
return [
TyDiagnostic.from_gitlab_output(dct, source=source)
for dct in json.loads(process.stdout)
if dct["severity"] == "major"
]
def group_diagnostics_by_key(
old: list[TyDiagnostic],
new: list[TyDiagnostic],
expected: list[ExpectedError],
) -> list[TestCase]:
# Build a lookup from (filename, line) to tag so ty diagnostics on a tagged
# line can be grouped with all other expected errors sharing that tag.
tagged_lines: dict[tuple[str, int], str] = {
(e.location.path.name, e.location.positions.begin.line): e.tag
for e in expected
if e.tag is not None
}
def ty_key(diag: TyDiagnostic) -> str:
tag = tagged_lines.get(
(diag.location.path.name, diag.location.positions.begin.line)
)
return (
f"{diag.location.path.as_posix()}:{tag}"
if tag is not None
else f"{diag.location.path.as_posix()}:{diag.location.positions.begin.line}"
)
old_by_key: defaultdict[str, list[TyDiagnostic]] = defaultdict(list)
new_by_key: defaultdict[str, list[TyDiagnostic]] = defaultdict(list)
expected_by_key: defaultdict[str, list[ExpectedError]] = defaultdict(list)
for diag in old:
old_by_key[ty_key(diag)].append(diag)
for diag in new:
new_by_key[ty_key(diag)].append(diag)
for err in expected:
expected_by_key[err.key].append(err)
all_keys = sorted(old_by_key.keys() | new_by_key.keys() | expected_by_key.keys())
return [
TestCase(
key=key,
old=old_by_key[key],
new=new_by_key[key],
expected=expected_by_key[key],
)
for key in all_keys
]
def compute_stats(test_cases: list[TestCase], source: Source) -> Statistics:
stats = Statistics()
for tc in test_cases:
match tc.classify(source):
case Classification.TRUE_POSITIVE:
stats.true_positives += 1
case Classification.FALSE_POSITIVE:
stats.false_positives += 1
case Classification.FALSE_NEGATIVE:
stats.false_negatives += 1
case Classification.TRUE_NEGATIVE:
pass
stats.total_diagnostics += len(tc.diagnostics_by_source(source))
return stats
def collect_diagnostic_entries(test_cases: list[TestCase]) -> list[DiagnosticEntry]:
"""Classify each changed test case and assign it to a titled section."""
entries: list[DiagnosticEntry] = []
for tc in test_cases:
change = tc.change
if change == Change.UNCHANGED:
continue
if tc.optional:
if change == Change.ADDED:
entries.append(
DiagnosticEntry(
title=Change.ADDED.into_title(), test_case=tc, source=Source.NEW
)
)
elif change == Change.REMOVED:
entries.append(
DiagnosticEntry(
title=Change.REMOVED.into_title(),
test_case=tc,
source=Source.OLD,
)
)
elif change == Change.CHANGED:
entries.append(
DiagnosticEntry(
title=Change.CHANGED.into_title(), test_case=tc, source=None
)
)
else:
if change == Change.ADDED:
new_class = tc.classify(Source.NEW)
entries.append(
DiagnosticEntry(
title=new_class.into_title(verb="added"),
test_case=tc,
source=Source.NEW,
)
)
elif change == Change.REMOVED:
old_class = tc.classify(Source.OLD)
entries.append(
DiagnosticEntry(
title=old_class.into_title(verb="removed"),
test_case=tc,
source=Source.OLD,
)
)
elif change == Change.CHANGED:
old_class = tc.classify(Source.OLD)
new_class = tc.classify(Source.NEW)
if old_class == new_class:
# Same classification but different diagnostics: one "changed" section.
entries.append(
DiagnosticEntry(
title=new_class.into_title(verb="changed"),
test_case=tc,
source=None,
)
)
else:
# Classification changed: split into separate removed/added sections.
entries.append(
DiagnosticEntry(
title=old_class.into_title(verb="removed"),
test_case=tc,
source=Source.OLD,
)
)
entries.append(
DiagnosticEntry(
title=new_class.into_title(verb="added"),
test_case=tc,
source=Source.NEW,
)
)
entries.sort(
key=lambda e: (TITLE_PRIORITY.get(e.title, 99), e.title, e.test_case.key)
)
return entries
def render_test_cases(
test_cases: list[TestCase],
*,
format: Literal["diff", "github"] = "diff",
) -> str:
entries = collect_diagnostic_entries(test_cases)
if not entries:
return ""
lines = []
for title, group in groupby(entries, key=lambda e: e.title):
group_list = list(group)
n = len(group_list)
lines.append(f"### {title} ({n})")
lines.extend(
[
"",
"<details>",
f"<summary>{n} {'diagnostic' if n == 1 else 'diagnostics'}</summary>",
"",
]
)
if format == "diff":
lines.append("```diff")
else:
lines.extend(GITHUB_HEADER)
for entry in group_list:
tc, source = entry.test_case, entry.source
if source is None:
if format == "diff":
lines.append(render_diff_row(tc.old, removed=True))
lines.append(render_diff_row(tc.new, removed=False))
else:
lines.extend(render_html_diff_row(tc, source=None))
else:
if format == "diff":
lines.append(
render_diff_row(
tc.diagnostics_by_source(source),
removed=source == Source.OLD,
)
)
else:
lines.extend(render_html_diff_row(tc, source=source))
if format == "diff":
lines.append("```")
else:
lines.extend(GITHUB_FOOTER)
lines.extend(["</details>", ""])
return "\n".join(lines)
def collect_file_stats(test_cases: list[TestCase]) -> list[FileStats]:
"""Compute per-file statistics from grouped test cases."""
path_to_cases: dict[Path, list[TestCase]] = {}
for tc in test_cases:
path_to_cases.setdefault(tc.path, []).append(tc)
return [
FileStats(
path=path,
old=compute_stats(cases, Source.OLD),
new=compute_stats(cases, Source.NEW),
)
for path, cases in path_to_cases.items()
]
def render_file_stats_table(file_stats: list[FileStats]) -> str:
"""Render a per-file breakdown showing only files whose TP/FP/FN counts changed."""
def fmt(old: int, new: int, *, greater_is_better: bool = True) -> str:
if old == new:
return str(new)
diff = new - old
improved = diff > 0 if greater_is_better else diff < 0
indicator = " ✅" if improved else " ❌"
return f"{new} ({diff:+}){indicator}"
# Collect totals across all files regardless of change.
old_totals = Statistics()
new_totals = Statistics()
passing = 0
total_files = len(file_stats)
for fs in file_stats:
old_totals.true_positives += fs.old.true_positives
old_totals.false_positives += fs.old.false_positives
old_totals.false_negatives += fs.old.false_negatives
new_totals.true_positives += fs.new.true_positives
new_totals.false_positives += fs.new.false_positives
new_totals.false_negatives += fs.new.false_negatives
passing += fs.new_passes
changed = [fs for fs in file_stats if fs.total_change > 0]
if not changed:
return ""
changed.sort(key=lambda fs: (-fs.total_change, fs.path.name))
rows = []
for fs in changed:
if fs.new_passes and not fs.old_passes:
status = "✅ Newly Passing 🎉"
elif fs.old_passes and not fs.new_passes:
status = "❌ Newly Failing ☹️"
elif fs.new_passes:
status = "✅ Still Passing"
else:
old_errors = fs.old.false_positives + fs.old.false_negatives
new_errors = fs.new.false_positives + fs.new.false_negatives
if new_errors < old_errors:
status = "📈 Improving"
elif new_errors > old_errors:
status = "📉 Regressing"
else:
status = "➡️ Neutral"
url = CONFORMANCE_DIR_WITH_README + f"tests/{fs.path.name}"
rows.append(
f"| [{fs.path.name}]({url})"
f" | {fmt(fs.old.true_positives, fs.new.true_positives, greater_is_better=True)}"
f" | {fmt(fs.old.false_positives, fs.new.false_positives, greater_is_better=False)}"
f" | {fmt(fs.old.false_negatives, fs.new.false_negatives, greater_is_better=False)}"
f" | {status} |"
)
totals_row = (
f"| **Total (all files)**"
f" | **{fmt(old_totals.true_positives, new_totals.true_positives, greater_is_better=True)}**"
f" | **{fmt(old_totals.false_positives, new_totals.false_positives, greater_is_better=False)}**"
f" | **{fmt(old_totals.false_negatives, new_totals.false_negatives, greater_is_better=False)}**"
f" | {passing}/{total_files} |"
)
lines = [
"### Test file breakdown",
"",
"<details>",
f"<summary>{len(changed)} file{'s' if len(changed) != 1 else ''} altered</summary>",
"",
"| File | True Positives | False Positives | False Negatives | Status |",
"|------|----|----|----|--------|",
*rows,
totals_row,
"",
"</details>",
"",
]
return "\n".join(lines)
def diff_format(
diff: float,
*,
greater_is_better: bool = True,
neutral: bool = False,
) -> str:
if diff == 0:
return ""
increased = diff > 0
good = " (✅)" if not neutral else ""
bad = " (❌)" if not neutral else ""
up = "⏫"
down = "⏬"
match (greater_is_better, increased):
case (True, True):
return f"{up}{good}"
case (False, True):
return f"{up}{bad}"
case (True, False):
return f"{down}{bad}"
case (False, False):
return f"{down}{good}"
case _:
assert_never((greater_is_better, increased))
def render_summary(
test_cases: list[TestCase],
file_stats: list[FileStats],
*,
force_summary_table: bool,
) -> str:
def format_metric(diff: float, old: float, new: float) -> str:
if diff > 0:
return f"increased from <b>{old:.2%}</b> to <b>{new:.2%}</b>"
if diff < 0:
return f"decreased from <b>{old:.2%}</b> to <b>{new:.2%}</b>"
return f"held steady at <b>{old:.2%}</b>"
def format_int_metric(diff: int, old: int, new: int, total: int) -> str:
if diff > 0:
return f"improved from <b>{old}/{total}</b> to <b>{new}/{total}</b>"
if diff < 0:
return f"regressed from <b>{old}/{total}</b> to <b>{new}/{total}</b>"
return f"held steady at <b>{old}/{total}</b>"
old = compute_stats(test_cases, Source.OLD)
new = compute_stats(test_cases, Source.NEW)
old_files_passing = sum(fs.old_passes for fs in file_stats)
new_files_passing = sum(fs.new_passes for fs in file_stats)
total_files = len(file_stats)
files_passing_change = new_files_passing - old_files_passing
assert new.true_positives > 0, (
"Expected ty to have at least one true positive.\n"
f"Sample of grouped diagnostics: {test_cases[:5]}"
)
precision_change = new.precision - old.precision
recall_change = new.recall - old.recall
true_pos_change = new.true_positives - old.true_positives
false_pos_change = new.false_positives - old.false_positives
false_neg_change = new.false_negatives - old.false_negatives
total_change = new.total_diagnostics - old.total_diagnostics
summary_paragraph = (
f"The percentage of diagnostics emitted that were expected errors "
f"{format_metric(precision_change, old.precision, new.precision)}. "
f"The percentage of expected errors that received a diagnostic "
f"{format_metric(recall_change, old.recall, new.recall)}. "
f"The number of fully passing files "
f"{format_int_metric(files_passing_change, old_files_passing, new_files_passing, total_files)}."
)
base_header = f"[Typing conformance results]({CONFORMANCE_DIR_WITH_README})"
if not force_summary_table and all(
diag.change is Change.UNCHANGED for diag in test_cases
):
return dedent(
f"""
## {base_header}
### No changes detected ✅
<details>
<summary>Current numbers</summary>
<br>
{summary_paragraph}
</details>
"""
)
true_pos_diff = diff_format(true_pos_change, greater_is_better=True)
false_pos_diff = diff_format(false_pos_change, greater_is_better=False)
false_neg_diff = diff_format(false_neg_change, greater_is_better=False)
precision_diff = diff_format(precision_change, greater_is_better=True)
recall_diff = diff_format(recall_change, greater_is_better=True)
total_diff = diff_format(total_change, neutral=True)
passing_diff = diff_format(files_passing_change, greater_is_better=True)
if (
(precision_change > 0 and recall_change >= 0 and files_passing_change >= 0)
or (recall_change > 0 and precision_change >= 0 and files_passing_change >= 0)
or (files_passing_change > 0 and precision_change >= 0 and recall_change >= 0)
):
header = f"{base_header} improved 🎉"
elif (
(precision_change < 0 and recall_change <= 0 and files_passing_change <= 0)
or (recall_change < 0 and precision_change <= 0 and files_passing_change <= 0)
or (files_passing_change < 0 and precision_change <= 0 and recall_change <= 0)
):
header = f"{base_header} regressed ❌"
else:
header = base_header
summary_note = " ".join(SUMMARY_NOTE.split())
return dedent(
f"""
## {header}
{summary_paragraph}
### Summary
<details>
<summary>How are test cases classified?</summary>
<br>
{summary_note}
</details>
| Metric | Old | New | Diff | Outcome |
|--------|-----|-----|------|---------|
| True Positives | {old.true_positives} | {new.true_positives} | {true_pos_change:+} | {true_pos_diff} |
| False Positives | {old.false_positives} | {new.false_positives} | {false_pos_change:+} | {false_pos_diff} |
| False Negatives | {old.false_negatives} | {new.false_negatives} | {false_neg_change:+} | {false_neg_diff} |
| Total Diagnostics | {old.total_diagnostics} | {new.total_diagnostics} | {total_change:+} | {total_diff} |
| Precision | {old.precision:.2%} | {new.precision:.2%} | {precision_change:+.2%} | {precision_diff} |
| Recall | {old.recall:.2%} | {new.recall:.2%} | {recall_change:+.2%} | {recall_diff} |
| Passing Files | {old_files_passing}/{total_files} | {new_files_passing}/{total_files} | {files_passing_change:+} | {passing_diff} |
"""
)
def get_test_groups(root_dir: Path) -> AbstractSet[str]:
"""Adapted from typing/conformance/test_groups.py."""
# Read the TOML file that defines the test groups. Each test