You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

520 lines
20 KiB

10 years ago
10 years ago
10 years ago
  1. from __future__ import unicode_literals
  2. import io
  3. import os
  4. import subprocess
  5. import time
  6. from .common import AudioConversionError, PostProcessor
  7. from ..compat import (
  8. compat_subprocess_get_DEVNULL,
  9. )
  10. from ..utils import (
  11. encodeArgument,
  12. encodeFilename,
  13. get_exe_version,
  14. is_outdated_version,
  15. PostProcessingError,
  16. prepend_extension,
  17. shell_quote,
  18. subtitles_filename,
  19. dfxp2srt,
  20. ISO639Utils,
  21. )
  22. class FFmpegPostProcessorError(PostProcessingError):
  23. pass
  24. class FFmpegPostProcessor(PostProcessor):
  25. def __init__(self, downloader=None):
  26. PostProcessor.__init__(self, downloader)
  27. self._determine_executables()
  28. def check_version(self):
  29. if not self.available:
  30. raise FFmpegPostProcessorError('ffmpeg or avconv not found. Please install one.')
  31. required_version = '10-0' if self.basename == 'avconv' else '1.0'
  32. if is_outdated_version(
  33. self._versions[self.basename], required_version):
  34. warning = 'Your copy of %s is outdated, update %s to version %s or newer if you encounter any errors.' % (
  35. self.basename, self.basename, required_version)
  36. if self._downloader:
  37. self._downloader.report_warning(warning)
  38. @staticmethod
  39. def get_versions(downloader=None):
  40. return FFmpegPostProcessor(downloader)._versions
  41. def _determine_executables(self):
  42. programs = ['avprobe', 'avconv', 'ffmpeg', 'ffprobe']
  43. prefer_ffmpeg = self._downloader.params.get('prefer_ffmpeg', False)
  44. self.basename = None
  45. self.probe_basename = None
  46. self._paths = None
  47. self._versions = None
  48. if self._downloader:
  49. location = self._downloader.params.get('ffmpeg_location')
  50. if location is not None:
  51. if not os.path.exists(location):
  52. self._downloader.report_warning(
  53. 'ffmpeg-location %s does not exist! '
  54. 'Continuing without avconv/ffmpeg.' % (location))
  55. self._versions = {}
  56. return
  57. elif not os.path.isdir(location):
  58. basename = os.path.splitext(os.path.basename(location))[0]
  59. if basename not in programs:
  60. self._downloader.report_warning(
  61. 'Cannot identify executable %s, its basename should be one of %s. '
  62. 'Continuing without avconv/ffmpeg.' %
  63. (location, ', '.join(programs)))
  64. self._versions = {}
  65. return None
  66. location = os.path.dirname(os.path.abspath(location))
  67. if basename in ('ffmpeg', 'ffprobe'):
  68. prefer_ffmpeg = True
  69. self._paths = dict(
  70. (p, os.path.join(location, p)) for p in programs)
  71. self._versions = dict(
  72. (p, get_exe_version(self._paths[p], args=['-version']))
  73. for p in programs)
  74. if self._versions is None:
  75. self._versions = dict(
  76. (p, get_exe_version(p, args=['-version'])) for p in programs)
  77. self._paths = dict((p, p) for p in programs)
  78. if prefer_ffmpeg:
  79. prefs = ('ffmpeg', 'avconv')
  80. else:
  81. prefs = ('avconv', 'ffmpeg')
  82. for p in prefs:
  83. if self._versions[p]:
  84. self.basename = p
  85. break
  86. if prefer_ffmpeg:
  87. prefs = ('ffprobe', 'avprobe')
  88. else:
  89. prefs = ('avprobe', 'ffprobe')
  90. for p in prefs:
  91. if self._versions[p]:
  92. self.probe_basename = p
  93. break
  94. @property
  95. def available(self):
  96. return self.basename is not None
  97. @property
  98. def executable(self):
  99. return self._paths[self.basename]
  100. @property
  101. def probe_available(self):
  102. return self.probe_basename is not None
  103. @property
  104. def probe_executable(self):
  105. return self._paths[self.probe_basename]
  106. def run_ffmpeg_multiple_files(self, input_paths, out_path, opts):
  107. self.check_version()
  108. oldest_mtime = min(
  109. os.stat(encodeFilename(path)).st_mtime for path in input_paths)
  110. opts += self._configuration_args()
  111. files_cmd = []
  112. for path in input_paths:
  113. files_cmd.extend([encodeArgument('-i'), encodeFilename(path, True)])
  114. cmd = ([encodeFilename(self.executable, True), encodeArgument('-y')] +
  115. files_cmd +
  116. [encodeArgument(o) for o in opts] +
  117. [encodeFilename(self._ffmpeg_filename_argument(out_path), True)])
  118. if self._downloader.params.get('verbose', False):
  119. self._downloader.to_screen('[debug] ffmpeg command line: %s' % shell_quote(cmd))
  120. p = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, stdin=subprocess.PIPE)
  121. stdout, stderr = p.communicate()
  122. if p.returncode != 0:
  123. stderr = stderr.decode('utf-8', 'replace')
  124. msg = stderr.strip().split('\n')[-1]
  125. raise FFmpegPostProcessorError(msg)
  126. self.try_utime(out_path, oldest_mtime, oldest_mtime)
  127. def run_ffmpeg(self, path, out_path, opts):
  128. self.run_ffmpeg_multiple_files([path], out_path, opts)
  129. def _ffmpeg_filename_argument(self, fn):
  130. # ffmpeg broke --, see https://ffmpeg.org/trac/ffmpeg/ticket/2127 for details
  131. if fn.startswith('-'):
  132. return './' + fn
  133. return fn
  134. class FFmpegExtractAudioPP(FFmpegPostProcessor):
  135. def __init__(self, downloader=None, preferredcodec=None, preferredquality=None, nopostoverwrites=False):
  136. FFmpegPostProcessor.__init__(self, downloader)
  137. if preferredcodec is None:
  138. preferredcodec = 'best'
  139. self._preferredcodec = preferredcodec
  140. self._preferredquality = preferredquality
  141. self._nopostoverwrites = nopostoverwrites
  142. def get_audio_codec(self, path):
  143. if not self.probe_available:
  144. raise PostProcessingError('ffprobe or avprobe not found. Please install one.')
  145. try:
  146. cmd = [
  147. encodeFilename(self.probe_executable, True),
  148. encodeArgument('-show_streams'),
  149. encodeFilename(self._ffmpeg_filename_argument(path), True)]
  150. if self._downloader.params.get('verbose', False):
  151. self._downloader.to_screen('[debug] %s command line: %s' % (self.basename, shell_quote(cmd)))
  152. handle = subprocess.Popen(cmd, stderr=compat_subprocess_get_DEVNULL(), stdout=subprocess.PIPE, stdin=subprocess.PIPE)
  153. output = handle.communicate()[0]
  154. if handle.wait() != 0:
  155. return None
  156. except (IOError, OSError):
  157. return None
  158. audio_codec = None
  159. for line in output.decode('ascii', 'ignore').split('\n'):
  160. if line.startswith('codec_name='):
  161. audio_codec = line.split('=')[1].strip()
  162. elif line.strip() == 'codec_type=audio' and audio_codec is not None:
  163. return audio_codec
  164. return None
  165. def run_ffmpeg(self, path, out_path, codec, more_opts):
  166. if codec is None:
  167. acodec_opts = []
  168. else:
  169. acodec_opts = ['-acodec', codec]
  170. opts = ['-vn'] + acodec_opts + more_opts
  171. try:
  172. FFmpegPostProcessor.run_ffmpeg(self, path, out_path, opts)
  173. except FFmpegPostProcessorError as err:
  174. raise AudioConversionError(err.msg)
  175. def run(self, information):
  176. path = information['filepath']
  177. filecodec = self.get_audio_codec(path)
  178. if filecodec is None:
  179. raise PostProcessingError('WARNING: unable to obtain file audio codec with ffprobe')
  180. more_opts = []
  181. if self._preferredcodec == 'best' or self._preferredcodec == filecodec or (self._preferredcodec == 'm4a' and filecodec == 'aac'):
  182. if filecodec == 'aac' and self._preferredcodec in ['m4a', 'best']:
  183. # Lossless, but in another container
  184. acodec = 'copy'
  185. extension = 'm4a'
  186. more_opts = ['-bsf:a', 'aac_adtstoasc']
  187. elif filecodec in ['aac', 'mp3', 'vorbis', 'opus']:
  188. # Lossless if possible
  189. acodec = 'copy'
  190. extension = filecodec
  191. if filecodec == 'aac':
  192. more_opts = ['-f', 'adts']
  193. if filecodec == 'vorbis':
  194. extension = 'ogg'
  195. else:
  196. # MP3 otherwise.
  197. acodec = 'libmp3lame'
  198. extension = 'mp3'
  199. more_opts = []
  200. if self._preferredquality is not None:
  201. if int(self._preferredquality) < 10:
  202. more_opts += ['-q:a', self._preferredquality]
  203. else:
  204. more_opts += ['-b:a', self._preferredquality + 'k']
  205. else:
  206. # We convert the audio (lossy)
  207. acodec = {'mp3': 'libmp3lame', 'aac': 'aac', 'm4a': 'aac', 'opus': 'opus', 'vorbis': 'libvorbis', 'wav': None}[self._preferredcodec]
  208. extension = self._preferredcodec
  209. more_opts = []
  210. if self._preferredquality is not None:
  211. # The opus codec doesn't support the -aq option
  212. if int(self._preferredquality) < 10 and extension != 'opus':
  213. more_opts += ['-q:a', self._preferredquality]
  214. else:
  215. more_opts += ['-b:a', self._preferredquality + 'k']
  216. if self._preferredcodec == 'aac':
  217. more_opts += ['-f', 'adts']
  218. if self._preferredcodec == 'm4a':
  219. more_opts += ['-bsf:a', 'aac_adtstoasc']
  220. if self._preferredcodec == 'vorbis':
  221. extension = 'ogg'
  222. if self._preferredcodec == 'wav':
  223. extension = 'wav'
  224. more_opts += ['-f', 'wav']
  225. prefix, sep, ext = path.rpartition('.') # not os.path.splitext, since the latter does not work on unicode in all setups
  226. new_path = prefix + sep + extension
  227. # If we download foo.mp3 and convert it to... foo.mp3, then don't delete foo.mp3, silly.
  228. if (new_path == path or
  229. (self._nopostoverwrites and os.path.exists(encodeFilename(new_path)))):
  230. self._downloader.to_screen('[ffmpeg] Post-process file %s exists, skipping' % new_path)
  231. return [], information
  232. try:
  233. self._downloader.to_screen('[' + self.basename + '] Destination: ' + new_path)
  234. self.run_ffmpeg(path, new_path, acodec, more_opts)
  235. except AudioConversionError as e:
  236. raise PostProcessingError(
  237. 'audio conversion failed: ' + e.msg)
  238. except Exception:
  239. raise PostProcessingError('error running ' + self.basename)
  240. # Try to update the date time for extracted audio file.
  241. if information.get('filetime') is not None:
  242. self.try_utime(
  243. new_path, time.time(), information['filetime'],
  244. errnote='Cannot update utime of audio file')
  245. information['filepath'] = new_path
  246. information['ext'] = extension
  247. return [path], information
  248. class FFmpegVideoConvertorPP(FFmpegPostProcessor):
  249. def __init__(self, downloader=None, preferedformat=None):
  250. super(FFmpegVideoConvertorPP, self).__init__(downloader)
  251. self._preferedformat = preferedformat
  252. def run(self, information):
  253. path = information['filepath']
  254. if information['ext'] == self._preferedformat:
  255. self._downloader.to_screen('[ffmpeg] Not converting video file %s - already is in target format %s' % (path, self._preferedformat))
  256. return [], information
  257. options = []
  258. if self._preferedformat == 'avi':
  259. options.extend(['-c:v', 'libxvid', '-vtag', 'XVID'])
  260. prefix, sep, ext = path.rpartition('.')
  261. outpath = prefix + sep + self._preferedformat
  262. self._downloader.to_screen('[' + 'ffmpeg' + '] Converting video from %s to %s, Destination: ' % (information['ext'], self._preferedformat) + outpath)
  263. self.run_ffmpeg(path, outpath, options)
  264. information['filepath'] = outpath
  265. information['format'] = self._preferedformat
  266. information['ext'] = self._preferedformat
  267. return [path], information
  268. class FFmpegEmbedSubtitlePP(FFmpegPostProcessor):
  269. def run(self, information):
  270. if information['ext'] not in ['mp4', 'mkv']:
  271. self._downloader.to_screen('[ffmpeg] Subtitles can only be embedded in mp4 or mkv files')
  272. return [], information
  273. subtitles = information.get('requested_subtitles')
  274. if not subtitles:
  275. self._downloader.to_screen('[ffmpeg] There aren\'t any subtitles to embed')
  276. return [], information
  277. sub_langs = list(subtitles.keys())
  278. filename = information['filepath']
  279. sub_filenames = [subtitles_filename(filename, lang, sub_info['ext']) for lang, sub_info in subtitles.items()]
  280. input_files = [filename] + sub_filenames
  281. opts = [
  282. '-map', '0',
  283. '-c', 'copy',
  284. # Don't copy the existing subtitles, we may be running the
  285. # postprocessor a second time
  286. '-map', '-0:s',
  287. ]
  288. if information['ext'] == 'mp4':
  289. opts += ['-c:s', 'mov_text']
  290. for (i, lang) in enumerate(sub_langs):
  291. opts.extend(['-map', '%d:0' % (i + 1)])
  292. lang_code = ISO639Utils.short2long(lang)
  293. if lang_code is not None:
  294. opts.extend(['-metadata:s:s:%d' % i, 'language=%s' % lang_code])
  295. temp_filename = prepend_extension(filename, 'temp')
  296. self._downloader.to_screen('[ffmpeg] Embedding subtitles in \'%s\'' % filename)
  297. self.run_ffmpeg_multiple_files(input_files, temp_filename, opts)
  298. os.remove(encodeFilename(filename))
  299. os.rename(encodeFilename(temp_filename), encodeFilename(filename))
  300. return sub_filenames, information
  301. class FFmpegMetadataPP(FFmpegPostProcessor):
  302. def run(self, info):
  303. metadata = {}
  304. if info.get('title') is not None:
  305. metadata['title'] = info['title']
  306. if info.get('upload_date') is not None:
  307. metadata['date'] = info['upload_date']
  308. if info.get('artist') is not None:
  309. metadata['artist'] = info['artist']
  310. elif info.get('uploader') is not None:
  311. metadata['artist'] = info['uploader']
  312. elif info.get('uploader_id') is not None:
  313. metadata['artist'] = info['uploader_id']
  314. if info.get('description') is not None:
  315. metadata['description'] = info['description']
  316. metadata['comment'] = info['description']
  317. if info.get('webpage_url') is not None:
  318. metadata['purl'] = info['webpage_url']
  319. if info.get('album') is not None:
  320. metadata['album'] = info['album']
  321. if not metadata:
  322. self._downloader.to_screen('[ffmpeg] There isn\'t any metadata to add')
  323. return [], info
  324. filename = info['filepath']
  325. temp_filename = prepend_extension(filename, 'temp')
  326. if info['ext'] == 'm4a':
  327. options = ['-vn', '-acodec', 'copy']
  328. else:
  329. options = ['-c', 'copy']
  330. for (name, value) in metadata.items():
  331. options.extend(['-metadata', '%s=%s' % (name, value)])
  332. self._downloader.to_screen('[ffmpeg] Adding metadata to \'%s\'' % filename)
  333. self.run_ffmpeg(filename, temp_filename, options)
  334. os.remove(encodeFilename(filename))
  335. os.rename(encodeFilename(temp_filename), encodeFilename(filename))
  336. return [], info
  337. class FFmpegMergerPP(FFmpegPostProcessor):
  338. def run(self, info):
  339. filename = info['filepath']
  340. temp_filename = prepend_extension(filename, 'temp')
  341. args = ['-c', 'copy', '-map', '0:v:0', '-map', '1:a:0']
  342. self._downloader.to_screen('[ffmpeg] Merging formats into "%s"' % filename)
  343. self.run_ffmpeg_multiple_files(info['__files_to_merge'], temp_filename, args)
  344. os.rename(encodeFilename(temp_filename), encodeFilename(filename))
  345. return info['__files_to_merge'], info
  346. def can_merge(self):
  347. # TODO: figure out merge-capable ffmpeg version
  348. if self.basename != 'avconv':
  349. return True
  350. required_version = '10-0'
  351. if is_outdated_version(
  352. self._versions[self.basename], required_version):
  353. warning = ('Your copy of %s is outdated and unable to properly mux separate video and audio files, '
  354. 'youtube-dl will download single file media. '
  355. 'Update %s to version %s or newer to fix this.') % (
  356. self.basename, self.basename, required_version)
  357. if self._downloader:
  358. self._downloader.report_warning(warning)
  359. return False
  360. return True
  361. class FFmpegFixupStretchedPP(FFmpegPostProcessor):
  362. def run(self, info):
  363. stretched_ratio = info.get('stretched_ratio')
  364. if stretched_ratio is None or stretched_ratio == 1:
  365. return [], info
  366. filename = info['filepath']
  367. temp_filename = prepend_extension(filename, 'temp')
  368. options = ['-c', 'copy', '-aspect', '%f' % stretched_ratio]
  369. self._downloader.to_screen('[ffmpeg] Fixing aspect ratio in "%s"' % filename)
  370. self.run_ffmpeg(filename, temp_filename, options)
  371. os.remove(encodeFilename(filename))
  372. os.rename(encodeFilename(temp_filename), encodeFilename(filename))
  373. return [], info
  374. class FFmpegFixupM4aPP(FFmpegPostProcessor):
  375. def run(self, info):
  376. if info.get('container') != 'm4a_dash':
  377. return [], info
  378. filename = info['filepath']
  379. temp_filename = prepend_extension(filename, 'temp')
  380. options = ['-c', 'copy', '-f', 'mp4']
  381. self._downloader.to_screen('[ffmpeg] Correcting container in "%s"' % filename)
  382. self.run_ffmpeg(filename, temp_filename, options)
  383. os.remove(encodeFilename(filename))
  384. os.rename(encodeFilename(temp_filename), encodeFilename(filename))
  385. return [], info
  386. class FFmpegSubtitlesConvertorPP(FFmpegPostProcessor):
  387. def __init__(self, downloader=None, format=None):
  388. super(FFmpegSubtitlesConvertorPP, self).__init__(downloader)
  389. self.format = format
  390. def run(self, info):
  391. subs = info.get('requested_subtitles')
  392. filename = info['filepath']
  393. new_ext = self.format
  394. new_format = new_ext
  395. if new_format == 'vtt':
  396. new_format = 'webvtt'
  397. if subs is None:
  398. self._downloader.to_screen('[ffmpeg] There aren\'t any subtitles to convert')
  399. return [], info
  400. self._downloader.to_screen('[ffmpeg] Converting subtitles')
  401. for lang, sub in subs.items():
  402. ext = sub['ext']
  403. if ext == new_ext:
  404. self._downloader.to_screen(
  405. '[ffmpeg] Subtitle file for %s is already in the requested'
  406. 'format' % new_ext)
  407. continue
  408. new_file = subtitles_filename(filename, lang, new_ext)
  409. if ext == 'dfxp' or ext == 'ttml':
  410. self._downloader.report_warning(
  411. 'You have requested to convert dfxp (TTML) subtitles into another format, '
  412. 'which results in style information loss')
  413. dfxp_file = subtitles_filename(filename, lang, ext)
  414. srt_file = subtitles_filename(filename, lang, 'srt')
  415. with io.open(dfxp_file, 'rt', encoding='utf-8') as f:
  416. srt_data = dfxp2srt(f.read())
  417. with io.open(srt_file, 'wt', encoding='utf-8') as f:
  418. f.write(srt_data)
  419. ext = 'srt'
  420. subs[lang] = {
  421. 'ext': 'srt',
  422. 'data': srt_data
  423. }
  424. if new_ext == 'srt':
  425. continue
  426. self.run_ffmpeg(
  427. subtitles_filename(filename, lang, ext),
  428. new_file, ['-f', new_format])
  429. with io.open(new_file, 'rt', encoding='utf-8') as f:
  430. subs[lang] = {
  431. 'ext': ext,
  432. 'data': f.read(),
  433. }
  434. return [], info