]>
Commit | Line | Data |
---|---|---|
1 | from __future__ import annotations | |
2 | ||
3 | # Allow direct execution | |
4 | import os | |
5 | import sys | |
6 | ||
7 | sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__)))) | |
8 | ||
9 | import enum | |
10 | import itertools | |
11 | import json | |
12 | import logging | |
13 | import re | |
14 | from collections import defaultdict | |
15 | from dataclasses import dataclass | |
16 | from functools import lru_cache | |
17 | from pathlib import Path | |
18 | ||
19 | from devscripts.utils import read_file, run_process, write_file | |
20 | ||
21 | BASE_URL = 'https://github.com' | |
22 | LOCATION_PATH = Path(__file__).parent | |
23 | HASH_LENGTH = 7 | |
24 | ||
25 | logger = logging.getLogger(__name__) | |
26 | ||
27 | ||
28 | class CommitGroup(enum.Enum): | |
29 | PRIORITY = 'Important' | |
30 | CORE = 'Core' | |
31 | EXTRACTOR = 'Extractor' | |
32 | DOWNLOADER = 'Downloader' | |
33 | POSTPROCESSOR = 'Postprocessor' | |
34 | NETWORKING = 'Networking' | |
35 | MISC = 'Misc.' | |
36 | ||
37 | @classmethod | |
38 | @lru_cache | |
39 | def subgroup_lookup(cls): | |
40 | return { | |
41 | name: group | |
42 | for group, names in { | |
43 | cls.MISC: { | |
44 | 'build', | |
45 | 'ci', | |
46 | 'cleanup', | |
47 | 'devscripts', | |
48 | 'docs', | |
49 | 'test', | |
50 | }, | |
51 | cls.NETWORKING: { | |
52 | 'rh', | |
53 | }, | |
54 | }.items() | |
55 | for name in names | |
56 | } | |
57 | ||
58 | @classmethod | |
59 | @lru_cache | |
60 | def group_lookup(cls): | |
61 | result = { | |
62 | 'fd': cls.DOWNLOADER, | |
63 | 'ie': cls.EXTRACTOR, | |
64 | 'pp': cls.POSTPROCESSOR, | |
65 | 'upstream': cls.CORE, | |
66 | } | |
67 | result.update({item.name.lower(): item for item in iter(cls)}) | |
68 | return result | |
69 | ||
70 | @classmethod | |
71 | def get(cls, value: str) -> tuple[CommitGroup | None, str | None]: | |
72 | group, _, subgroup = (group.strip().lower() for group in value.partition('/')) | |
73 | ||
74 | result = cls.group_lookup().get(group) | |
75 | if not result: | |
76 | if subgroup: | |
77 | return None, value | |
78 | subgroup = group | |
79 | result = cls.subgroup_lookup().get(subgroup) | |
80 | ||
81 | return result, subgroup or None | |
82 | ||
83 | ||
84 | @dataclass | |
85 | class Commit: | |
86 | hash: str | None | |
87 | short: str | |
88 | authors: list[str] | |
89 | ||
90 | def __str__(self): | |
91 | result = f'{self.short!r}' | |
92 | ||
93 | if self.hash: | |
94 | result += f' ({self.hash[:HASH_LENGTH]})' | |
95 | ||
96 | if self.authors: | |
97 | authors = ', '.join(self.authors) | |
98 | result += f' by {authors}' | |
99 | ||
100 | return result | |
101 | ||
102 | ||
103 | @dataclass | |
104 | class CommitInfo: | |
105 | details: str | None | |
106 | sub_details: tuple[str, ...] | |
107 | message: str | |
108 | issues: list[str] | |
109 | commit: Commit | |
110 | fixes: list[Commit] | |
111 | ||
112 | def key(self): | |
113 | return ((self.details or '').lower(), self.sub_details, self.message) | |
114 | ||
115 | ||
116 | def unique(items): | |
117 | return sorted({item.strip().lower(): item for item in items if item}.values()) | |
118 | ||
119 | ||
120 | class Changelog: | |
121 | MISC_RE = re.compile(r'(?:^|\b)(?:lint(?:ing)?|misc|format(?:ting)?|fixes)(?:\b|$)', re.IGNORECASE) | |
122 | ALWAYS_SHOWN = (CommitGroup.PRIORITY,) | |
123 | ||
124 | def __init__(self, groups, repo, collapsible=False): | |
125 | self._groups = groups | |
126 | self._repo = repo | |
127 | self._collapsible = collapsible | |
128 | ||
129 | def __str__(self): | |
130 | return '\n'.join(self._format_groups(self._groups)).replace('\t', ' ') | |
131 | ||
132 | def _format_groups(self, groups): | |
133 | first = True | |
134 | for item in CommitGroup: | |
135 | if self._collapsible and item not in self.ALWAYS_SHOWN and first: | |
136 | first = False | |
137 | yield '\n<details><summary><h3>Changelog</h3></summary>\n' | |
138 | ||
139 | group = groups[item] | |
140 | if group: | |
141 | yield self.format_module(item.value, group) | |
142 | ||
143 | if self._collapsible: | |
144 | yield '\n</details>' | |
145 | ||
146 | def format_module(self, name, group): | |
147 | result = f'\n#### {name} changes\n' if name else '\n' | |
148 | return result + '\n'.join(self._format_group(group)) | |
149 | ||
150 | def _format_group(self, group): | |
151 | sorted_group = sorted(group, key=CommitInfo.key) | |
152 | detail_groups = itertools.groupby(sorted_group, lambda item: (item.details or '').lower()) | |
153 | for _, items in detail_groups: | |
154 | items = list(items) | |
155 | details = items[0].details | |
156 | ||
157 | if details == 'cleanup': | |
158 | items = self._prepare_cleanup_misc_items(items) | |
159 | ||
160 | prefix = '-' | |
161 | if details: | |
162 | if len(items) == 1: | |
163 | prefix = f'- **{details}**:' | |
164 | else: | |
165 | yield f'- **{details}**' | |
166 | prefix = '\t-' | |
167 | ||
168 | sub_detail_groups = itertools.groupby(items, lambda item: tuple(map(str.lower, item.sub_details))) | |
169 | for sub_details, entries in sub_detail_groups: | |
170 | if not sub_details: | |
171 | for entry in entries: | |
172 | yield f'{prefix} {self.format_single_change(entry)}' | |
173 | continue | |
174 | ||
175 | entries = list(entries) | |
176 | sub_prefix = f'{prefix} {", ".join(entries[0].sub_details)}' | |
177 | if len(entries) == 1: | |
178 | yield f'{sub_prefix}: {self.format_single_change(entries[0])}' | |
179 | continue | |
180 | ||
181 | yield sub_prefix | |
182 | for entry in entries: | |
183 | yield f'\t{prefix} {self.format_single_change(entry)}' | |
184 | ||
185 | def _prepare_cleanup_misc_items(self, items): | |
186 | cleanup_misc_items = defaultdict(list) | |
187 | sorted_items = [] | |
188 | for item in items: | |
189 | if self.MISC_RE.search(item.message): | |
190 | cleanup_misc_items[tuple(item.commit.authors)].append(item) | |
191 | else: | |
192 | sorted_items.append(item) | |
193 | ||
194 | for commit_infos in cleanup_misc_items.values(): | |
195 | sorted_items.append(CommitInfo( | |
196 | 'cleanup', ('Miscellaneous',), ', '.join( | |
197 | self._format_message_link(None, info.commit.hash) | |
198 | for info in sorted(commit_infos, key=lambda item: item.commit.hash or '')), | |
199 | [], Commit(None, '', commit_infos[0].commit.authors), [])) | |
200 | ||
201 | return sorted_items | |
202 | ||
203 | def format_single_change(self, info: CommitInfo): | |
204 | message, sep, rest = info.message.partition('\n') | |
205 | if '[' not in message: | |
206 | # If the message doesn't already contain markdown links, try to add a link to the commit | |
207 | message = self._format_message_link(message, info.commit.hash) | |
208 | ||
209 | if info.issues: | |
210 | message = f'{message} ({self._format_issues(info.issues)})' | |
211 | ||
212 | if info.commit.authors: | |
213 | message = f'{message} by {self._format_authors(info.commit.authors)}' | |
214 | ||
215 | if info.fixes: | |
216 | fix_message = ', '.join(f'{self._format_message_link(None, fix.hash)}' for fix in info.fixes) | |
217 | ||
218 | authors = sorted({author for fix in info.fixes for author in fix.authors}, key=str.casefold) | |
219 | if authors != info.commit.authors: | |
220 | fix_message = f'{fix_message} by {self._format_authors(authors)}' | |
221 | ||
222 | message = f'{message} (With fixes in {fix_message})' | |
223 | ||
224 | return message if not sep else f'{message}{sep}{rest}' | |
225 | ||
226 | def _format_message_link(self, message, hash): | |
227 | assert message or hash, 'Improperly defined commit message or override' | |
228 | message = message if message else hash[:HASH_LENGTH] | |
229 | return f'[{message}]({self.repo_url}/commit/{hash})' if hash else message | |
230 | ||
231 | def _format_issues(self, issues): | |
232 | return ', '.join(f'[#{issue}]({self.repo_url}/issues/{issue})' for issue in issues) | |
233 | ||
234 | @staticmethod | |
235 | def _format_authors(authors): | |
236 | return ', '.join(f'[{author}]({BASE_URL}/{author})' for author in authors) | |
237 | ||
238 | @property | |
239 | def repo_url(self): | |
240 | return f'{BASE_URL}/{self._repo}' | |
241 | ||
242 | ||
243 | class CommitRange: | |
244 | COMMAND = 'git' | |
245 | COMMIT_SEPARATOR = '-----' | |
246 | ||
247 | AUTHOR_INDICATOR_RE = re.compile(r'Authored by:? ', re.IGNORECASE) | |
248 | MESSAGE_RE = re.compile(r''' | |
249 | (?:\[(?P<prefix>[^\]]+)\]\ )? | |
250 | (?:(?P<sub_details>`?[\w.-]+`?): )? | |
251 | (?P<message>.+?) | |
252 | (?:\ \((?P<issues>\#\d+(?:,\ \#\d+)*)\))? | |
253 | ''', re.VERBOSE | re.DOTALL) | |
254 | EXTRACTOR_INDICATOR_RE = re.compile(r'(?:Fix|Add)\s+Extractors?', re.IGNORECASE) | |
255 | REVERT_RE = re.compile(r'(?:\[[^\]]+\]\s+)?(?i:Revert)\s+([\da-f]{40})') | |
256 | FIXES_RE = re.compile(r'(?i:Fix(?:es)?(?:\s+bugs?)?(?:\s+in|\s+for)?|Revert|Improve)\s+([\da-f]{40})') | |
257 | UPSTREAM_MERGE_RE = re.compile(r'Update to ytdl-commit-([\da-f]+)') | |
258 | ||
259 | def __init__(self, start, end, default_author=None): | |
260 | self._start, self._end = start, end | |
261 | self._commits, self._fixes = self._get_commits_and_fixes(default_author) | |
262 | self._commits_added = [] | |
263 | ||
264 | def __iter__(self): | |
265 | return iter(itertools.chain(self._commits.values(), self._commits_added)) | |
266 | ||
267 | def __len__(self): | |
268 | return len(self._commits) + len(self._commits_added) | |
269 | ||
270 | def __contains__(self, commit): | |
271 | if isinstance(commit, Commit): | |
272 | if not commit.hash: | |
273 | return False | |
274 | commit = commit.hash | |
275 | ||
276 | return commit in self._commits | |
277 | ||
278 | def _get_commits_and_fixes(self, default_author): | |
279 | result = run_process( | |
280 | self.COMMAND, 'log', f'--format=%H%n%s%n%b%n{self.COMMIT_SEPARATOR}', | |
281 | f'{self._start}..{self._end}' if self._start else self._end).stdout | |
282 | ||
283 | commits, reverts = {}, {} | |
284 | fixes = defaultdict(list) | |
285 | lines = iter(result.splitlines(False)) | |
286 | for i, commit_hash in enumerate(lines): | |
287 | short = next(lines) | |
288 | skip = short.startswith('Release ') or short == '[version] update' | |
289 | ||
290 | authors = [default_author] if default_author else [] | |
291 | for line in iter(lambda: next(lines), self.COMMIT_SEPARATOR): | |
292 | match = self.AUTHOR_INDICATOR_RE.match(line) | |
293 | if match: | |
294 | authors = sorted(map(str.strip, line[match.end():].split(',')), key=str.casefold) | |
295 | ||
296 | commit = Commit(commit_hash, short, authors) | |
297 | if skip and (self._start or not i): | |
298 | logger.debug(f'Skipped commit: {commit}') | |
299 | continue | |
300 | elif skip: | |
301 | logger.debug(f'Reached Release commit, breaking: {commit}') | |
302 | break | |
303 | ||
304 | revert_match = self.REVERT_RE.fullmatch(commit.short) | |
305 | if revert_match: | |
306 | reverts[revert_match.group(1)] = commit | |
307 | continue | |
308 | ||
309 | fix_match = self.FIXES_RE.search(commit.short) | |
310 | if fix_match: | |
311 | commitish = fix_match.group(1) | |
312 | fixes[commitish].append(commit) | |
313 | ||
314 | commits[commit.hash] = commit | |
315 | ||
316 | for commitish, revert_commit in reverts.items(): | |
317 | reverted = commits.pop(commitish, None) | |
318 | if reverted: | |
319 | logger.debug(f'{commitish} fully reverted {reverted}') | |
320 | else: | |
321 | commits[revert_commit.hash] = revert_commit | |
322 | ||
323 | for commitish, fix_commits in fixes.items(): | |
324 | if commitish in commits: | |
325 | hashes = ', '.join(commit.hash[:HASH_LENGTH] for commit in fix_commits) | |
326 | logger.info(f'Found fix(es) for {commitish[:HASH_LENGTH]}: {hashes}') | |
327 | for fix_commit in fix_commits: | |
328 | del commits[fix_commit.hash] | |
329 | else: | |
330 | logger.debug(f'Commit with fixes not in changes: {commitish[:HASH_LENGTH]}') | |
331 | ||
332 | return commits, fixes | |
333 | ||
334 | def apply_overrides(self, overrides): | |
335 | for override in overrides: | |
336 | when = override.get('when') | |
337 | if when and when not in self and when != self._start: | |
338 | logger.debug(f'Ignored {when!r} override') | |
339 | continue | |
340 | ||
341 | override_hash = override.get('hash') or when | |
342 | if override['action'] == 'add': | |
343 | commit = Commit(override.get('hash'), override['short'], override.get('authors') or []) | |
344 | logger.info(f'ADD {commit}') | |
345 | self._commits_added.append(commit) | |
346 | ||
347 | elif override['action'] == 'remove': | |
348 | if override_hash in self._commits: | |
349 | logger.info(f'REMOVE {self._commits[override_hash]}') | |
350 | del self._commits[override_hash] | |
351 | ||
352 | elif override['action'] == 'change': | |
353 | if override_hash not in self._commits: | |
354 | continue | |
355 | commit = Commit(override_hash, override['short'], override.get('authors') or []) | |
356 | logger.info(f'CHANGE {self._commits[commit.hash]} -> {commit}') | |
357 | self._commits[commit.hash] = commit | |
358 | ||
359 | self._commits = {key: value for key, value in reversed(self._commits.items())} | |
360 | ||
361 | def groups(self): | |
362 | group_dict = defaultdict(list) | |
363 | for commit in self: | |
364 | upstream_re = self.UPSTREAM_MERGE_RE.search(commit.short) | |
365 | if upstream_re: | |
366 | commit.short = f'[upstream] Merged with youtube-dl {upstream_re.group(1)}' | |
367 | ||
368 | match = self.MESSAGE_RE.fullmatch(commit.short) | |
369 | if not match: | |
370 | logger.error(f'Error parsing short commit message: {commit.short!r}') | |
371 | continue | |
372 | ||
373 | prefix, sub_details_alt, message, issues = match.groups() | |
374 | issues = [issue.strip()[1:] for issue in issues.split(',')] if issues else [] | |
375 | ||
376 | if prefix: | |
377 | groups, details, sub_details = zip(*map(self.details_from_prefix, prefix.split(','))) | |
378 | group = next(iter(filter(None, groups)), None) | |
379 | details = ', '.join(unique(details)) | |
380 | sub_details = list(itertools.chain.from_iterable(sub_details)) | |
381 | else: | |
382 | group = CommitGroup.CORE | |
383 | details = None | |
384 | sub_details = [] | |
385 | ||
386 | if sub_details_alt: | |
387 | sub_details.append(sub_details_alt) | |
388 | sub_details = tuple(unique(sub_details)) | |
389 | ||
390 | if not group: | |
391 | if self.EXTRACTOR_INDICATOR_RE.search(commit.short): | |
392 | group = CommitGroup.EXTRACTOR | |
393 | logger.error(f'Assuming [ie] group for {commit.short!r}') | |
394 | else: | |
395 | group = CommitGroup.CORE | |
396 | ||
397 | commit_info = CommitInfo( | |
398 | details, sub_details, message.strip(), | |
399 | issues, commit, self._fixes[commit.hash]) | |
400 | ||
401 | logger.debug(f'Resolved {commit.short!r} to {commit_info!r}') | |
402 | group_dict[group].append(commit_info) | |
403 | ||
404 | return group_dict | |
405 | ||
406 | @staticmethod | |
407 | def details_from_prefix(prefix): | |
408 | if not prefix: | |
409 | return CommitGroup.CORE, None, () | |
410 | ||
411 | prefix, *sub_details = prefix.split(':') | |
412 | ||
413 | group, details = CommitGroup.get(prefix) | |
414 | if group is CommitGroup.PRIORITY and details: | |
415 | details = details.partition('/')[2].strip() | |
416 | ||
417 | if details and '/' in details: | |
418 | logger.error(f'Prefix is overnested, using first part: {prefix}') | |
419 | details = details.partition('/')[0].strip() | |
420 | ||
421 | if details == 'common': | |
422 | details = None | |
423 | elif group is CommitGroup.NETWORKING and details == 'rh': | |
424 | details = 'Request Handler' | |
425 | ||
426 | return group, details, sub_details | |
427 | ||
428 | ||
429 | def get_new_contributors(contributors_path, commits): | |
430 | contributors = set() | |
431 | if contributors_path.exists(): | |
432 | for line in read_file(contributors_path).splitlines(): | |
433 | author, _, _ = line.strip().partition(' (') | |
434 | authors = author.split('/') | |
435 | contributors.update(map(str.casefold, authors)) | |
436 | ||
437 | new_contributors = set() | |
438 | for commit in commits: | |
439 | for author in commit.authors: | |
440 | author_folded = author.casefold() | |
441 | if author_folded not in contributors: | |
442 | contributors.add(author_folded) | |
443 | new_contributors.add(author) | |
444 | ||
445 | return sorted(new_contributors, key=str.casefold) | |
446 | ||
447 | ||
448 | def create_changelog(args): | |
449 | logging.basicConfig( | |
450 | datefmt='%Y-%m-%d %H-%M-%S', format='{asctime} | {levelname:<8} | {message}', | |
451 | level=logging.WARNING - 10 * args.verbosity, style='{', stream=sys.stderr) | |
452 | ||
453 | commits = CommitRange(None, args.commitish, args.default_author) | |
454 | ||
455 | if not args.no_override: | |
456 | if args.override_path.exists(): | |
457 | overrides = json.loads(read_file(args.override_path)) | |
458 | commits.apply_overrides(overrides) | |
459 | else: | |
460 | logger.warning(f'File {args.override_path.as_posix()} does not exist') | |
461 | ||
462 | logger.info(f'Loaded {len(commits)} commits') | |
463 | ||
464 | new_contributors = get_new_contributors(args.contributors_path, commits) | |
465 | if new_contributors: | |
466 | if args.contributors: | |
467 | write_file(args.contributors_path, '\n'.join(new_contributors) + '\n', mode='a') | |
468 | logger.info(f'New contributors: {", ".join(new_contributors)}') | |
469 | ||
470 | return Changelog(commits.groups(), args.repo, args.collapsible) | |
471 | ||
472 | ||
473 | def create_parser(): | |
474 | import argparse | |
475 | ||
476 | parser = argparse.ArgumentParser( | |
477 | description='Create a changelog markdown from a git commit range') | |
478 | parser.add_argument( | |
479 | 'commitish', default='HEAD', nargs='?', | |
480 | help='The commitish to create the range from (default: %(default)s)') | |
481 | parser.add_argument( | |
482 | '-v', '--verbosity', action='count', default=0, | |
483 | help='increase verbosity (can be used twice)') | |
484 | parser.add_argument( | |
485 | '-c', '--contributors', action='store_true', | |
486 | help='update CONTRIBUTORS file (default: %(default)s)') | |
487 | parser.add_argument( | |
488 | '--contributors-path', type=Path, default=LOCATION_PATH.parent / 'CONTRIBUTORS', | |
489 | help='path to the CONTRIBUTORS file') | |
490 | parser.add_argument( | |
491 | '--no-override', action='store_true', | |
492 | help='skip override json in commit generation (default: %(default)s)') | |
493 | parser.add_argument( | |
494 | '--override-path', type=Path, default=LOCATION_PATH / 'changelog_override.json', | |
495 | help='path to the changelog_override.json file') | |
496 | parser.add_argument( | |
497 | '--default-author', default='pukkandan', | |
498 | help='the author to use without a author indicator (default: %(default)s)') | |
499 | parser.add_argument( | |
500 | '--repo', default='yt-dlp/yt-dlp', | |
501 | help='the github repository to use for the operations (default: %(default)s)') | |
502 | parser.add_argument( | |
503 | '--collapsible', action='store_true', | |
504 | help='make changelog collapsible (default: %(default)s)') | |
505 | ||
506 | return parser | |
507 | ||
508 | ||
509 | if __name__ == '__main__': | |
510 | print(create_changelog(create_parser().parse_args())) |