]>
Commit | Line | Data |
---|---|---|
3aa578ca PH |
1 | from __future__ import unicode_literals |
2 | ||
e9fade72 | 3 | import io |
7dde84f3 | 4 | import itertools |
496c1923 PH |
5 | import os |
6 | import subprocess | |
496c1923 | 7 | import time |
fa2a36d9 | 8 | import re |
06167fbb | 9 | import json |
496c1923 | 10 | |
496c1923 PH |
11 | from .common import AudioConversionError, PostProcessor |
12 | ||
84601bb7 | 13 | from ..compat import compat_str, compat_numeric_types |
8c25f81b | 14 | from ..utils import ( |
7a340e0d | 15 | dfxp2srt, |
f07b74fc | 16 | encodeArgument, |
496c1923 | 17 | encodeFilename, |
95807118 | 18 | get_exe_version, |
48844745 | 19 | is_outdated_version, |
7a340e0d NA |
20 | ISO639Utils, |
21 | orderedSet, | |
496c1923 PH |
22 | PostProcessingError, |
23 | prepend_extension, | |
f5b1bca9 | 24 | process_communicate_or_kill, |
06167fbb | 25 | replace_extension, |
7a340e0d | 26 | shell_quote, |
324ad820 | 27 | traverse_obj, |
6606817a | 28 | variadic, |
496c1923 PH |
29 | ) |
30 | ||
31 | ||
a755f825 | 32 | EXT_TO_OUT_FORMATS = { |
21bfcd3d PH |
33 | 'aac': 'adts', |
34 | 'flac': 'flac', | |
35 | 'm4a': 'ipod', | |
36 | 'mka': 'matroska', | |
37 | 'mkv': 'matroska', | |
38 | 'mpg': 'mpeg', | |
39 | 'ogv': 'ogg', | |
40 | 'ts': 'mpegts', | |
41 | 'wma': 'asf', | |
42 | 'wmv': 'asf', | |
43 | } | |
44 | ACODECS = { | |
45 | 'mp3': 'libmp3lame', | |
46 | 'aac': 'aac', | |
47 | 'flac': 'flac', | |
48 | 'm4a': 'aac', | |
d2ae7e24 | 49 | 'opus': 'libopus', |
21bfcd3d PH |
50 | 'vorbis': 'libvorbis', |
51 | 'wav': None, | |
a755f825 | 52 | } |
53 | ||
54 | ||
496c1923 PH |
55 | class FFmpegPostProcessorError(PostProcessingError): |
56 | pass | |
57 | ||
d799b47b | 58 | |
496c1923 | 59 | class FFmpegPostProcessor(PostProcessor): |
d47aeb22 | 60 | def __init__(self, downloader=None): |
496c1923 | 61 | PostProcessor.__init__(self, downloader) |
73fac4e9 | 62 | self._determine_executables() |
496c1923 | 63 | |
48844745 | 64 | def check_version(self): |
f740fae2 | 65 | if not self.available: |
beb4b92a | 66 | raise FFmpegPostProcessorError('ffmpeg not found. Please install or provide the path using --ffmpeg-location') |
48844745 | 67 | |
65bf37ef | 68 | required_version = '10-0' if self.basename == 'avconv' else '1.0' |
48844745 | 69 | if is_outdated_version( |
73fac4e9 | 70 | self._versions[self.basename], required_version): |
3aa578ca | 71 | warning = 'Your copy of %s is outdated, update %s to version %s or newer if you encounter any errors.' % ( |
73fac4e9 | 72 | self.basename, self.basename, required_version) |
f446cc66 | 73 | self.report_warning(warning) |
48844745 | 74 | |
496c1923 | 75 | @staticmethod |
73fac4e9 PH |
76 | def get_versions(downloader=None): |
77 | return FFmpegPostProcessor(downloader)._versions | |
6271f1ca | 78 | |
73fac4e9 PH |
79 | def _determine_executables(self): |
80 | programs = ['avprobe', 'avconv', 'ffmpeg', 'ffprobe'] | |
d4a24f40 | 81 | prefer_ffmpeg = True |
73fac4e9 | 82 | |
a64646e4 RA |
83 | def get_ffmpeg_version(path): |
84 | ver = get_exe_version(path, args=['-version']) | |
85 | if ver: | |
86 | regexs = [ | |
cbdc688c | 87 | r'(?:\d+:)?([0-9.]+)-[0-9]+ubuntu[0-9.]+$', # Ubuntu, see [1] |
5caa531a | 88 | r'n([0-9.]+)$', # Arch Linux |
cbdc688c | 89 | # 1. http://www.ducea.com/2006/06/17/ubuntu-package-version-naming-explanation/ |
a64646e4 RA |
90 | ] |
91 | for regex in regexs: | |
92 | mobj = re.match(regex, ver) | |
93 | if mobj: | |
94 | ver = mobj.group(1) | |
95 | return ver | |
96 | ||
73fac4e9 PH |
97 | self.basename = None |
98 | self.probe_basename = None | |
99 | ||
100 | self._paths = None | |
101 | self._versions = None | |
102 | if self._downloader: | |
f446cc66 | 103 | prefer_ffmpeg = self.get_param('prefer_ffmpeg', True) |
104 | location = self.get_param('ffmpeg_location') | |
73fac4e9 PH |
105 | if location is not None: |
106 | if not os.path.exists(location): | |
f446cc66 | 107 | self.report_warning( |
73fac4e9 | 108 | 'ffmpeg-location %s does not exist! ' |
e4172ac9 | 109 | 'Continuing without ffmpeg.' % (location)) |
73fac4e9 PH |
110 | self._versions = {} |
111 | return | |
8c0ae192 | 112 | elif os.path.isdir(location): |
113 | dirname, basename = location, None | |
114 | else: | |
73fac4e9 | 115 | basename = os.path.splitext(os.path.basename(location))[0] |
89efdc15 | 116 | basename = next((p for p in programs if basename.startswith(p)), 'ffmpeg') |
117 | dirname = os.path.dirname(os.path.abspath(location)) | |
73fac4e9 PH |
118 | if basename in ('ffmpeg', 'ffprobe'): |
119 | prefer_ffmpeg = True | |
120 | ||
121 | self._paths = dict( | |
89efdc15 | 122 | (p, os.path.join(dirname, p)) for p in programs) |
8c0ae192 | 123 | if basename: |
124 | self._paths[basename] = location | |
73fac4e9 | 125 | self._versions = dict( |
a64646e4 | 126 | (p, get_ffmpeg_version(self._paths[p])) for p in programs) |
73fac4e9 PH |
127 | if self._versions is None: |
128 | self._versions = dict( | |
a64646e4 | 129 | (p, get_ffmpeg_version(p)) for p in programs) |
73fac4e9 PH |
130 | self._paths = dict((p, p) for p in programs) |
131 | ||
d4a24f40 | 132 | if prefer_ffmpeg is False: |
d28b5171 | 133 | prefs = ('avconv', 'ffmpeg') |
d4a24f40 S |
134 | else: |
135 | prefs = ('ffmpeg', 'avconv') | |
d28b5171 PH |
136 | for p in prefs: |
137 | if self._versions[p]: | |
73fac4e9 PH |
138 | self.basename = p |
139 | break | |
76b1bd67 | 140 | |
d4a24f40 | 141 | if prefer_ffmpeg is False: |
1a253e13 | 142 | prefs = ('avprobe', 'ffprobe') |
d4a24f40 S |
143 | else: |
144 | prefs = ('ffprobe', 'avprobe') | |
1a253e13 PH |
145 | for p in prefs: |
146 | if self._versions[p]: | |
73fac4e9 PH |
147 | self.probe_basename = p |
148 | break | |
149 | ||
f740fae2 | 150 | @property |
73fac4e9 PH |
151 | def available(self): |
152 | return self.basename is not None | |
1a253e13 | 153 | |
73fac4e9 PH |
154 | @property |
155 | def executable(self): | |
156 | return self._paths[self.basename] | |
157 | ||
3da4b313 JMF |
158 | @property |
159 | def probe_available(self): | |
160 | return self.probe_basename is not None | |
161 | ||
73fac4e9 PH |
162 | @property |
163 | def probe_executable(self): | |
164 | return self._paths[self.probe_basename] | |
76b1bd67 | 165 | |
30d9e209 | 166 | def get_audio_codec(self, path): |
eb35b163 | 167 | if not self.probe_available and not self.available: |
beb4b92a | 168 | raise PostProcessingError('ffprobe and ffmpeg not found. Please install or provide the path using --ffmpeg-location') |
30d9e209 | 169 | try: |
eb35b163 RA |
170 | if self.probe_available: |
171 | cmd = [ | |
172 | encodeFilename(self.probe_executable, True), | |
173 | encodeArgument('-show_streams')] | |
174 | else: | |
175 | cmd = [ | |
176 | encodeFilename(self.executable, True), | |
177 | encodeArgument('-i')] | |
178 | cmd.append(encodeFilename(self._ffmpeg_filename_argument(path), True)) | |
f446cc66 | 179 | self.write_debug('%s command line: %s' % (self.basename, shell_quote(cmd))) |
eb35b163 RA |
180 | handle = subprocess.Popen( |
181 | cmd, stderr=subprocess.PIPE, | |
182 | stdout=subprocess.PIPE, stdin=subprocess.PIPE) | |
f5b1bca9 | 183 | stdout_data, stderr_data = process_communicate_or_kill(handle) |
eb35b163 RA |
184 | expected_ret = 0 if self.probe_available else 1 |
185 | if handle.wait() != expected_ret: | |
30d9e209 RA |
186 | return None |
187 | except (IOError, OSError): | |
188 | return None | |
eb35b163 RA |
189 | output = (stdout_data if self.probe_available else stderr_data).decode('ascii', 'ignore') |
190 | if self.probe_available: | |
191 | audio_codec = None | |
192 | for line in output.split('\n'): | |
193 | if line.startswith('codec_name='): | |
194 | audio_codec = line.split('=')[1].strip() | |
195 | elif line.strip() == 'codec_type=audio' and audio_codec is not None: | |
196 | return audio_codec | |
197 | else: | |
198 | # Stream #FILE_INDEX:STREAM_INDEX[STREAM_ID](LANGUAGE): CODEC_TYPE: CODEC_NAME | |
199 | mobj = re.search( | |
200 | r'Stream\s*#\d+:\d+(?:\[0x[0-9a-f]+\])?(?:\([a-z]{3}\))?:\s*Audio:\s*([0-9a-z]+)', | |
201 | output) | |
202 | if mobj: | |
203 | return mobj.group(1) | |
30d9e209 RA |
204 | return None |
205 | ||
06167fbb | 206 | def get_metadata_object(self, path, opts=[]): |
207 | if self.probe_basename != 'ffprobe': | |
208 | if self.probe_available: | |
209 | self.report_warning('Only ffprobe is supported for metadata extraction') | |
beb4b92a | 210 | raise PostProcessingError('ffprobe not found. Please install or provide the path using --ffmpeg-location') |
06167fbb | 211 | self.check_version() |
212 | ||
213 | cmd = [ | |
214 | encodeFilename(self.probe_executable, True), | |
215 | encodeArgument('-hide_banner'), | |
216 | encodeArgument('-show_format'), | |
217 | encodeArgument('-show_streams'), | |
218 | encodeArgument('-print_format'), | |
219 | encodeArgument('json'), | |
220 | ] | |
221 | ||
222 | cmd += opts | |
223 | cmd.append(encodeFilename(self._ffmpeg_filename_argument(path), True)) | |
06869367 | 224 | self.write_debug('ffprobe command line: %s' % shell_quote(cmd)) |
06167fbb | 225 | p = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, stdin=subprocess.PIPE) |
226 | stdout, stderr = p.communicate() | |
227 | return json.loads(stdout.decode('utf-8', 'replace')) | |
228 | ||
229 | def get_stream_number(self, path, keys, value): | |
230 | streams = self.get_metadata_object(path)['streams'] | |
231 | num = next( | |
324ad820 | 232 | (i for i, stream in enumerate(streams) if traverse_obj(stream, keys, casesense=False) == value), |
06167fbb | 233 | None) |
234 | return num, len(streams) | |
235 | ||
00034c14 | 236 | def run_ffmpeg_multiple_files(self, input_paths, out_path, opts, **kwargs): |
e92caff5 | 237 | return self.real_run_ffmpeg( |
238 | [(path, []) for path in input_paths], | |
00034c14 | 239 | [(out_path, opts)], **kwargs) |
e92caff5 | 240 | |
00034c14 | 241 | def real_run_ffmpeg(self, input_path_opts, output_path_opts, *, expected_retcodes=(0,)): |
48844745 | 242 | self.check_version() |
496c1923 | 243 | |
52afb2ac | 244 | oldest_mtime = min( |
7dde84f3 | 245 | os.stat(encodeFilename(path)).st_mtime for path, _ in input_path_opts if path) |
43bc8890 | 246 | |
ce52c7c1 S |
247 | cmd = [encodeFilename(self.executable, True), encodeArgument('-y')] |
248 | # avconv does not have repeat option | |
249 | if self.basename == 'ffmpeg': | |
250 | cmd += [encodeArgument('-loglevel'), encodeArgument('repeat+info')] | |
5b1ecbb3 | 251 | |
e92caff5 | 252 | def make_args(file, args, name, number): |
253 | keys = ['_%s%d' % (name, number), '_%s' % name] | |
254 | if name == 'o' and number == 1: | |
255 | keys.append('') | |
256 | args += self._configuration_args(self.basename, keys) | |
257 | if name == 'i': | |
258 | args.append('-i') | |
5b1ecbb3 | 259 | return ( |
e92caff5 | 260 | [encodeArgument(arg) for arg in args] |
5b1ecbb3 | 261 | + [encodeFilename(self._ffmpeg_filename_argument(file), True)]) |
262 | ||
e92caff5 | 263 | for arg_type, path_opts in (('i', input_path_opts), ('o', output_path_opts)): |
7dde84f3 | 264 | cmd += itertools.chain.from_iterable( |
265 | make_args(path, list(opts), arg_type, i + 1) | |
266 | for i, (path, opts) in enumerate(path_opts) if path) | |
496c1923 | 267 | |
f446cc66 | 268 | self.write_debug('ffmpeg command line: %s' % shell_quote(cmd)) |
cffcbc02 | 269 | p = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, stdin=subprocess.PIPE) |
f5b1bca9 | 270 | stdout, stderr = process_communicate_or_kill(p) |
00034c14 | 271 | if p.returncode not in variadic(expected_retcodes): |
06167fbb | 272 | stderr = stderr.decode('utf-8', 'replace').strip() |
06869367 | 273 | if self.get_param('verbose', False): |
06167fbb | 274 | self.report_error(stderr) |
275 | raise FFmpegPostProcessorError(stderr.split('\n')[-1]) | |
e92caff5 | 276 | for out_path, _ in output_path_opts: |
7dde84f3 | 277 | if out_path: |
278 | self.try_utime(out_path, oldest_mtime, oldest_mtime) | |
06167fbb | 279 | return stderr.decode('utf-8', 'replace') |
cc55d088 | 280 | |
00034c14 | 281 | def run_ffmpeg(self, path, out_path, opts, **kwargs): |
282 | return self.run_ffmpeg_multiple_files([path], out_path, opts, **kwargs) | |
496c1923 | 283 | |
7a340e0d NA |
284 | @staticmethod |
285 | def _ffmpeg_filename_argument(fn): | |
8a7bbd16 JMF |
286 | # Always use 'file:' because the filename may contain ':' (ffmpeg |
287 | # interprets that as a protocol) or can start with '-' (-- is broken in | |
288 | # ffmpeg, see https://ffmpeg.org/trac/ffmpeg/ticket/2127 for details) | |
b9f2fdd3 | 289 | # Also leave '-' intact in order not to break streaming to stdout. |
06167fbb | 290 | if fn.startswith(('http://', 'https://')): |
291 | return fn | |
d868f43c | 292 | return 'file:' + fn if fn != '-' else fn |
496c1923 | 293 | |
7a340e0d NA |
294 | @staticmethod |
295 | def _quote_for_ffmpeg(string): | |
296 | # See https://ffmpeg.org/ffmpeg-utils.html#toc-Quoting-and-escaping | |
297 | # A sequence of '' produces '\'''\''; | |
298 | # final replace removes the empty '' between \' \'. | |
299 | string = string.replace("'", r"'\''").replace("'''", "'") | |
300 | # Handle potential ' at string boundaries. | |
301 | string = string[1:] if string[0] == "'" else "'" + string | |
302 | return string[:-1] if string[-1] == "'" else string + "'" | |
303 | ||
304 | def force_keyframes(self, filename, timestamps): | |
305 | timestamps = orderedSet(timestamps) | |
306 | if timestamps[0] == 0: | |
307 | timestamps = timestamps[1:] | |
308 | keyframe_file = prepend_extension(filename, 'keyframes.temp') | |
309 | self.to_screen(f'Re-encoding "{filename}" with appropriate keyframes') | |
310 | self.run_ffmpeg(filename, keyframe_file, ['-force_key_frames', ','.join( | |
311 | f'{t:.6f}' for t in timestamps)]) | |
312 | return keyframe_file | |
313 | ||
314 | def concat_files(self, in_files, out_file, concat_opts=None): | |
315 | """ | |
316 | Use concat demuxer to concatenate multiple files having identical streams. | |
317 | ||
318 | Only inpoint, outpoint, and duration concat options are supported. | |
319 | See https://ffmpeg.org/ffmpeg-formats.html#concat-1 for details | |
320 | """ | |
321 | concat_file = f'{out_file}.concat' | |
322 | self.write_debug(f'Writing concat spec to {concat_file}') | |
323 | with open(concat_file, 'wt', encoding='utf-8') as f: | |
324 | f.writelines(self._concat_spec(in_files, concat_opts)) | |
325 | ||
326 | out_flags = ['-c', 'copy'] | |
327 | if out_file.rpartition('.')[-1] in ('mp4', 'mov'): | |
328 | # For some reason, '-c copy' is not enough to copy subtitles | |
329 | out_flags.extend(['-c:s', 'mov_text', '-movflags', '+faststart']) | |
330 | ||
331 | try: | |
332 | self.real_run_ffmpeg( | |
333 | [(concat_file, ['-hide_banner', '-nostdin', '-f', 'concat', '-safe', '0'])], | |
334 | [(out_file, out_flags)]) | |
335 | finally: | |
336 | os.remove(concat_file) | |
337 | ||
338 | @classmethod | |
339 | def _concat_spec(cls, in_files, concat_opts=None): | |
340 | if concat_opts is None: | |
341 | concat_opts = [{}] * len(in_files) | |
342 | yield 'ffconcat version 1.0\n' | |
343 | for file, opts in zip(in_files, concat_opts): | |
344 | yield f'file {cls._quote_for_ffmpeg(cls._ffmpeg_filename_argument(file))}\n' | |
345 | # Iterate explicitly to yield the following directives in order, ignoring the rest. | |
346 | for directive in 'inpoint', 'outpoint', 'duration': | |
347 | if directive in opts: | |
348 | yield f'{directive} {opts[directive]}\n' | |
349 | ||
496c1923 PH |
350 | |
351 | class FFmpegExtractAudioPP(FFmpegPostProcessor): | |
81a23040 | 352 | COMMON_AUDIO_EXTS = ('wav', 'flac', 'm4a', 'aiff', 'mp3', 'ogg', 'mka', 'opus', 'wma') |
353 | SUPPORTED_EXTS = ('best', 'aac', 'flac', 'mp3', 'm4a', 'opus', 'vorbis', 'wav') | |
1de75fa1 | 354 | |
496c1923 PH |
355 | def __init__(self, downloader=None, preferredcodec=None, preferredquality=None, nopostoverwrites=False): |
356 | FFmpegPostProcessor.__init__(self, downloader) | |
81a23040 | 357 | self._preferredcodec = preferredcodec or 'best' |
496c1923 PH |
358 | self._preferredquality = preferredquality |
359 | self._nopostoverwrites = nopostoverwrites | |
360 | ||
496c1923 | 361 | def run_ffmpeg(self, path, out_path, codec, more_opts): |
496c1923 PH |
362 | if codec is None: |
363 | acodec_opts = [] | |
364 | else: | |
365 | acodec_opts = ['-acodec', codec] | |
366 | opts = ['-vn'] + acodec_opts + more_opts | |
367 | try: | |
368 | FFmpegPostProcessor.run_ffmpeg(self, path, out_path, opts) | |
369 | except FFmpegPostProcessorError as err: | |
370 | raise AudioConversionError(err.msg) | |
371 | ||
8326b00a | 372 | @PostProcessor._restrict_to(images=False) |
496c1923 PH |
373 | def run(self, information): |
374 | path = information['filepath'] | |
1de75fa1 | 375 | orig_ext = information['ext'] |
376 | ||
81a23040 | 377 | if self._preferredcodec == 'best' and orig_ext in self.COMMON_AUDIO_EXTS: |
1de75fa1 | 378 | self.to_screen('Skipping audio extraction since the file is already in a common audio format') |
55b53b33 | 379 | return [], information |
496c1923 PH |
380 | |
381 | filecodec = self.get_audio_codec(path) | |
382 | if filecodec is None: | |
3aa578ca | 383 | raise PostProcessingError('WARNING: unable to obtain file audio codec with ffprobe') |
496c1923 PH |
384 | |
385 | more_opts = [] | |
386 | if self._preferredcodec == 'best' or self._preferredcodec == filecodec or (self._preferredcodec == 'm4a' and filecodec == 'aac'): | |
387 | if filecodec == 'aac' and self._preferredcodec in ['m4a', 'best']: | |
388 | # Lossless, but in another container | |
389 | acodec = 'copy' | |
390 | extension = 'm4a' | |
467d3c9a | 391 | more_opts = ['-bsf:a', 'aac_adtstoasc'] |
21bfcd3d | 392 | elif filecodec in ['aac', 'flac', 'mp3', 'vorbis', 'opus']: |
496c1923 PH |
393 | # Lossless if possible |
394 | acodec = 'copy' | |
395 | extension = filecodec | |
396 | if filecodec == 'aac': | |
397 | more_opts = ['-f', 'adts'] | |
398 | if filecodec == 'vorbis': | |
399 | extension = 'ogg' | |
400 | else: | |
401 | # MP3 otherwise. | |
402 | acodec = 'libmp3lame' | |
403 | extension = 'mp3' | |
404 | more_opts = [] | |
405 | if self._preferredquality is not None: | |
406 | if int(self._preferredquality) < 10: | |
467d3c9a | 407 | more_opts += ['-q:a', self._preferredquality] |
496c1923 | 408 | else: |
467d3c9a | 409 | more_opts += ['-b:a', self._preferredquality + 'k'] |
496c1923 | 410 | else: |
21bfcd3d PH |
411 | # We convert the audio (lossy if codec is lossy) |
412 | acodec = ACODECS[self._preferredcodec] | |
496c1923 PH |
413 | extension = self._preferredcodec |
414 | more_opts = [] | |
415 | if self._preferredquality is not None: | |
416 | # The opus codec doesn't support the -aq option | |
417 | if int(self._preferredquality) < 10 and extension != 'opus': | |
467d3c9a | 418 | more_opts += ['-q:a', self._preferredquality] |
496c1923 | 419 | else: |
467d3c9a | 420 | more_opts += ['-b:a', self._preferredquality + 'k'] |
496c1923 PH |
421 | if self._preferredcodec == 'aac': |
422 | more_opts += ['-f', 'adts'] | |
423 | if self._preferredcodec == 'm4a': | |
467d3c9a | 424 | more_opts += ['-bsf:a', 'aac_adtstoasc'] |
496c1923 PH |
425 | if self._preferredcodec == 'vorbis': |
426 | extension = 'ogg' | |
427 | if self._preferredcodec == 'wav': | |
428 | extension = 'wav' | |
429 | more_opts += ['-f', 'wav'] | |
430 | ||
3aa578ca | 431 | prefix, sep, ext = path.rpartition('.') # not os.path.splitext, since the latter does not work on unicode in all setups |
496c1923 | 432 | new_path = prefix + sep + extension |
0b94dbb1 | 433 | |
2273e2c5 PM |
434 | information['filepath'] = new_path |
435 | information['ext'] = extension | |
496c1923 PH |
436 | |
437 | # If we download foo.mp3 and convert it to... foo.mp3, then don't delete foo.mp3, silly. | |
3089bc74 S |
438 | if (new_path == path |
439 | or (self._nopostoverwrites and os.path.exists(encodeFilename(new_path)))): | |
1b77b347 | 440 | self.to_screen('Post-process file %s exists, skipping' % new_path) |
592e97e8 | 441 | return [], information |
496c1923 PH |
442 | |
443 | try: | |
1b77b347 | 444 | self.to_screen('Destination: ' + new_path) |
ce81b141 | 445 | self.run_ffmpeg(path, new_path, acodec, more_opts) |
70a1165b JMF |
446 | except AudioConversionError as e: |
447 | raise PostProcessingError( | |
448 | 'audio conversion failed: ' + e.msg) | |
449 | except Exception: | |
450 | raise PostProcessingError('error running ' + self.basename) | |
496c1923 PH |
451 | |
452 | # Try to update the date time for extracted audio file. | |
453 | if information.get('filetime') is not None: | |
dd29eb7f S |
454 | self.try_utime( |
455 | new_path, time.time(), information['filetime'], | |
456 | errnote='Cannot update utime of audio file') | |
496c1923 | 457 | |
592e97e8 | 458 | return [path], information |
496c1923 PH |
459 | |
460 | ||
857f6313 | 461 | class FFmpegVideoConvertorPP(FFmpegPostProcessor): |
81a23040 | 462 | SUPPORTED_EXTS = ('mp4', 'mkv', 'flv', 'webm', 'mov', 'avi', 'mp3', 'mka', 'm4a', 'ogg', 'opus') |
463 | FORMAT_RE = re.compile(r'{0}(?:/{0})*$'.format(r'(?:\w+>)?(?:%s)' % '|'.join(SUPPORTED_EXTS))) | |
857f6313 | 464 | _action = 'converting' |
465 | ||
efe87a10 | 466 | def __init__(self, downloader=None, preferedformat=None): |
857f6313 | 467 | super(FFmpegVideoConvertorPP, self).__init__(downloader) |
06167fbb | 468 | self._preferedformats = preferedformat.lower().split('/') |
efe87a10 | 469 | |
857f6313 | 470 | def _target_ext(self, source_ext): |
06167fbb | 471 | for pair in self._preferedformats: |
472 | kv = pair.split('>') | |
857f6313 | 473 | if len(kv) == 1 or kv[0].strip() == source_ext: |
474 | return kv[-1].strip() | |
06167fbb | 475 | |
857f6313 | 476 | @staticmethod |
477 | def _options(target_ext): | |
478 | if target_ext == 'avi': | |
479 | return ['-c:v', 'libxvid', '-vtag', 'XVID'] | |
480 | return [] | |
481 | ||
8326b00a | 482 | @PostProcessor._restrict_to(images=False) |
857f6313 | 483 | def run(self, information): |
81a23040 | 484 | path, source_ext = information['filepath'], information['ext'].lower() |
485 | target_ext = self._target_ext(source_ext) | |
06167fbb | 486 | _skip_msg = ( |
857f6313 | 487 | 'could not find a mapping for %s' if not target_ext |
488 | else 'already is in target format %s' if source_ext == target_ext | |
06167fbb | 489 | else None) |
490 | if _skip_msg: | |
81a23040 | 491 | self.to_screen('Not %s media file "%s"; %s' % (self._action, path, _skip_msg % source_ext)) |
efe87a10 | 492 | return [], information |
06167fbb | 493 | |
06167fbb | 494 | prefix, sep, oldext = path.rpartition('.') |
857f6313 | 495 | outpath = prefix + sep + target_ext |
496 | self.to_screen('%s video from %s to %s; Destination: %s' % (self._action.title(), source_ext, target_ext, outpath)) | |
497 | self.run_ffmpeg(path, outpath, self._options(target_ext)) | |
498 | ||
efe87a10 | 499 | information['filepath'] = outpath |
857f6313 | 500 | information['format'] = information['ext'] = target_ext |
efe87a10 FS |
501 | return [path], information |
502 | ||
503 | ||
857f6313 | 504 | class FFmpegVideoRemuxerPP(FFmpegVideoConvertorPP): |
505 | _action = 'remuxing' | |
496c1923 | 506 | |
857f6313 | 507 | @staticmethod |
508 | def _options(target_ext): | |
509 | options = ['-c', 'copy', '-map', '0', '-dn'] | |
510 | if target_ext in ['mp4', 'm4a', 'mov']: | |
511 | options.extend(['-movflags', '+faststart']) | |
512 | return options | |
496c1923 PH |
513 | |
514 | ||
515 | class FFmpegEmbedSubtitlePP(FFmpegPostProcessor): | |
cffab0ee | 516 | def __init__(self, downloader=None, already_have_subtitle=False): |
517 | super(FFmpegEmbedSubtitlePP, self).__init__(downloader) | |
518 | self._already_have_subtitle = already_have_subtitle | |
519 | ||
8326b00a | 520 | @PostProcessor._restrict_to(images=False) |
496c1923 | 521 | def run(self, information): |
40025ee2 | 522 | if information['ext'] not in ('mp4', 'webm', 'mkv'): |
1b77b347 | 523 | self.to_screen('Subtitles can only be embedded in mp4, webm or mkv files') |
592e97e8 | 524 | return [], information |
c84dd8a9 JMF |
525 | subtitles = information.get('requested_subtitles') |
526 | if not subtitles: | |
1b77b347 | 527 | self.to_screen('There aren\'t any subtitles to embed') |
592e97e8 | 528 | return [], information |
496c1923 | 529 | |
496c1923 | 530 | filename = information['filepath'] |
40025ee2 S |
531 | |
532 | ext = information['ext'] | |
2412044c | 533 | sub_langs, sub_names, sub_filenames = [], [], [] |
40025ee2 | 534 | webm_vtt_warn = False |
06167fbb | 535 | mp4_ass_warn = False |
40025ee2 S |
536 | |
537 | for lang, sub_info in subtitles.items(): | |
8e25d624 | 538 | if not os.path.exists(information.get('filepath', '')): |
539 | self.report_warning(f'Skipping embedding {lang} subtitle because the file is missing') | |
540 | continue | |
40025ee2 | 541 | sub_ext = sub_info['ext'] |
503d4a44 | 542 | if sub_ext == 'json': |
06167fbb | 543 | self.report_warning('JSON subtitles cannot be embedded') |
503d4a44 | 544 | elif ext != 'webm' or ext == 'webm' and sub_ext == 'vtt': |
40025ee2 | 545 | sub_langs.append(lang) |
2412044c | 546 | sub_names.append(sub_info.get('name')) |
dcf64d43 | 547 | sub_filenames.append(sub_info['filepath']) |
40025ee2 S |
548 | else: |
549 | if not webm_vtt_warn and ext == 'webm' and sub_ext != 'vtt': | |
550 | webm_vtt_warn = True | |
06167fbb | 551 | self.report_warning('Only WebVTT subtitles can be embedded in webm files') |
552 | if not mp4_ass_warn and ext == 'mp4' and sub_ext == 'ass': | |
553 | mp4_ass_warn = True | |
554 | self.report_warning('ASS subtitles cannot be properly embedded in mp4 files; expect issues') | |
40025ee2 S |
555 | |
556 | if not sub_langs: | |
557 | return [], information | |
558 | ||
14523ed9 | 559 | input_files = [filename] + sub_filenames |
496c1923 | 560 | |
e205db3b | 561 | opts = [ |
e0da59fe | 562 | '-c', 'copy', '-map', '0', '-dn', |
e205db3b JMF |
563 | # Don't copy the existing subtitles, we may be running the |
564 | # postprocessor a second time | |
565 | '-map', '-0:s', | |
7f903dd8 T |
566 | # Don't copy Apple TV chapters track, bin_data (see #19042, #19024, |
567 | # https://trac.ffmpeg.org/ticket/6016) | |
568 | '-map', '-0:d', | |
e205db3b | 569 | ] |
083c1bb9 N |
570 | if information['ext'] == 'mp4': |
571 | opts += ['-c:s', 'mov_text'] | |
2412044c | 572 | for i, (lang, name) in enumerate(zip(sub_langs, sub_names)): |
2875cf01 | 573 | opts.extend(['-map', '%d:0' % (i + 1)]) |
04fb6928 S |
574 | lang_code = ISO639Utils.short2long(lang) or lang |
575 | opts.extend(['-metadata:s:s:%d' % i, 'language=%s' % lang_code]) | |
2412044c | 576 | if name: |
577 | opts.extend(['-metadata:s:s:%d' % i, 'handler_name=%s' % name, | |
578 | '-metadata:s:s:%d' % i, 'title=%s' % name]) | |
496c1923 | 579 | |
2875cf01 | 580 | temp_filename = prepend_extension(filename, 'temp') |
06167fbb | 581 | self.to_screen('Embedding subtitles in "%s"' % filename) |
496c1923 | 582 | self.run_ffmpeg_multiple_files(input_files, temp_filename, opts) |
d75201a8 | 583 | os.replace(temp_filename, filename) |
496c1923 | 584 | |
cffab0ee | 585 | files_to_delete = [] if self._already_have_subtitle else sub_filenames |
586 | return files_to_delete, information | |
496c1923 PH |
587 | |
588 | ||
589 | class FFmpegMetadataPP(FFmpegPostProcessor): | |
7dde84f3 | 590 | |
7a340e0d NA |
591 | def __init__(self, downloader, add_metadata=True, add_chapters=True): |
592 | FFmpegPostProcessor.__init__(self, downloader) | |
593 | self._add_metadata = add_metadata | |
594 | self._add_chapters = add_chapters | |
595 | ||
7dde84f3 | 596 | @staticmethod |
597 | def _options(target_ext): | |
598 | yield from ('-map', '0', '-dn') | |
599 | if target_ext == 'm4a': | |
600 | yield from ('-vn', '-acodec', 'copy') | |
601 | else: | |
602 | yield from ('-c', 'copy') | |
603 | ||
8326b00a | 604 | @PostProcessor._restrict_to(images=False) |
496c1923 | 605 | def run(self, info): |
7a340e0d NA |
606 | filename, metadata_filename = info['filepath'], None |
607 | options = [] | |
608 | if self._add_chapters and info.get('chapters'): | |
609 | metadata_filename = replace_extension(filename, 'meta') | |
610 | options.extend(self._get_chapter_opts(info['chapters'], metadata_filename)) | |
611 | if self._add_metadata: | |
612 | options.extend(self._get_metadata_opts(info)) | |
613 | ||
614 | if not options: | |
615 | self.to_screen('There isn\'t any metadata to add') | |
616 | return [], info | |
617 | ||
618 | temp_filename = prepend_extension(filename, 'temp') | |
619 | self.to_screen('Adding metadata to "%s"' % filename) | |
620 | self.run_ffmpeg_multiple_files( | |
621 | (filename, metadata_filename), temp_filename, | |
622 | itertools.chain(self._options(info['ext']), *options)) | |
623 | if metadata_filename: | |
624 | os.remove(metadata_filename) | |
625 | os.replace(temp_filename, filename) | |
626 | return [], info | |
627 | ||
628 | @staticmethod | |
629 | def _get_chapter_opts(chapters, metadata_filename): | |
630 | with io.open(metadata_filename, 'wt', encoding='utf-8') as f: | |
631 | def ffmpeg_escape(text): | |
632 | return re.sub(r'([\\=;#\n])', r'\\\1', text) | |
633 | ||
634 | metadata_file_content = ';FFMETADATA1\n' | |
635 | for chapter in chapters: | |
636 | metadata_file_content += '[CHAPTER]\nTIMEBASE=1/1000\n' | |
637 | metadata_file_content += 'START=%d\n' % (chapter['start_time'] * 1000) | |
638 | metadata_file_content += 'END=%d\n' % (chapter['end_time'] * 1000) | |
639 | chapter_title = chapter.get('title') | |
640 | if chapter_title: | |
641 | metadata_file_content += 'title=%s\n' % ffmpeg_escape(chapter_title) | |
642 | f.write(metadata_file_content) | |
643 | yield ('-map_metadata', '1') | |
644 | ||
645 | def _get_metadata_opts(self, info): | |
496c1923 | 646 | metadata = {} |
4bd143a3 S |
647 | |
648 | def add(meta_list, info_list=None): | |
84601bb7 | 649 | if not meta_list: |
650 | return | |
6606817a | 651 | for info_f in variadic(info_list or meta_list): |
84601bb7 | 652 | if isinstance(info.get(info_f), (compat_str, compat_numeric_types)): |
6606817a | 653 | for meta_f in variadic(meta_list): |
4bd143a3 S |
654 | metadata[meta_f] = info[info_f] |
655 | break | |
656 | ||
2791e80b S |
657 | # See [1-4] for some info on media metadata/metadata supported |
658 | # by ffmpeg. | |
659 | # 1. https://kdenlive.org/en/project/adding-meta-data-to-mp4-video/ | |
660 | # 2. https://wiki.multimedia.cx/index.php/FFmpeg_Metadata | |
661 | # 3. https://kodi.wiki/view/Video_file_tagging | |
2791e80b | 662 | |
4bd143a3 S |
663 | add('title', ('track', 'title')) |
664 | add('date', 'upload_date') | |
cd9b384c | 665 | add(('description', 'synopsis'), 'description') |
666 | add(('purl', 'comment'), 'webpage_url') | |
4bd143a3 S |
667 | add('track', 'track_number') |
668 | add('artist', ('artist', 'creator', 'uploader', 'uploader_id')) | |
669 | add('genre') | |
670 | add('album') | |
671 | add('album_artist') | |
672 | add('disc', 'disc_number') | |
2791e80b S |
673 | add('show', 'series') |
674 | add('season_number') | |
675 | add('episode_id', ('episode', 'episode_id')) | |
676 | add('episode_sort', 'episode_number') | |
496c1923 | 677 | |
84601bb7 | 678 | prefix = 'meta_' |
679 | for key in filter(lambda k: k.startswith(prefix), info.keys()): | |
680 | add(key[len(prefix):], key) | |
681 | ||
7a340e0d NA |
682 | for name, value in metadata.items(): |
683 | yield ('-metadata', f'{name}={value}') | |
39c68260 | 684 | |
7dde84f3 | 685 | stream_idx = 0 |
686 | for fmt in info.get('requested_formats') or []: | |
687 | stream_count = 2 if 'none' not in (fmt.get('vcodec'), fmt.get('acodec')) else 1 | |
688 | if fmt.get('language'): | |
689 | lang = ISO639Utils.short2long(fmt['language']) or fmt['language'] | |
7a340e0d NA |
690 | for i in range(stream_count): |
691 | yield ('-metadata:s:%d' % (stream_idx + i), 'language=%s' % lang) | |
7dde84f3 | 692 | stream_idx += stream_count |
496c1923 | 693 | |
41712218 | 694 | if ('no-attach-info-json' not in self.get_param('compat_opts', []) |
695 | and '__infojson_filename' in info and info['ext'] in ('mkv', 'mka')): | |
7a340e0d | 696 | old_stream, new_stream = self.get_stream_number(info['filepath'], ('tags', 'mimetype'), 'application/json') |
06167fbb | 697 | if old_stream is not None: |
7a340e0d | 698 | yield ('-map', '-0:%d' % old_stream) |
06167fbb | 699 | new_stream -= 1 |
700 | ||
7a340e0d NA |
701 | yield ('-attach', info['__infojson_filename'], |
702 | '-metadata:s:%d' % new_stream, 'mimetype=application/json') | |
496c1923 PH |
703 | |
704 | ||
705 | class FFmpegMergerPP(FFmpegPostProcessor): | |
8326b00a | 706 | @PostProcessor._restrict_to(images=False) |
496c1923 PH |
707 | def run(self, info): |
708 | filename = info['filepath'] | |
5b5fbc08 | 709 | temp_filename = prepend_extension(filename, 'temp') |
d03cfdce | 710 | args = ['-c', 'copy'] |
711 | for (i, fmt) in enumerate(info['requested_formats']): | |
712 | if fmt.get('acodec') != 'none': | |
713 | args.extend(['-map', '%u:a:0' % (i)]) | |
714 | if fmt.get('vcodec') != 'none': | |
715 | args.extend(['-map', '%u:v:0' % (i)]) | |
1b77b347 | 716 | self.to_screen('Merging formats into "%s"' % filename) |
5b5fbc08 JMF |
717 | self.run_ffmpeg_multiple_files(info['__files_to_merge'], temp_filename, args) |
718 | os.rename(encodeFilename(temp_filename), encodeFilename(filename)) | |
d47aeb22 | 719 | return info['__files_to_merge'], info |
496c1923 | 720 | |
13763ce5 S |
721 | def can_merge(self): |
722 | # TODO: figure out merge-capable ffmpeg version | |
723 | if self.basename != 'avconv': | |
724 | return True | |
725 | ||
726 | required_version = '10-0' | |
727 | if is_outdated_version( | |
728 | self._versions[self.basename], required_version): | |
729 | warning = ('Your copy of %s is outdated and unable to properly mux separate video and audio files, ' | |
7a5c1cfe | 730 | 'yt-dlp will download single file media. ' |
13763ce5 S |
731 | 'Update %s to version %s or newer to fix this.') % ( |
732 | self.basename, self.basename, required_version) | |
f446cc66 | 733 | self.report_warning(warning) |
13763ce5 S |
734 | return False |
735 | return True | |
736 | ||
0c14e2fb | 737 | |
fd7cfb64 | 738 | class FFmpegFixupPostProcessor(FFmpegPostProcessor): |
739 | def _fixup(self, msg, filename, options): | |
6271f1ca PH |
740 | temp_filename = prepend_extension(filename, 'temp') |
741 | ||
f89b3e2d | 742 | self.to_screen(f'{msg} of "{filename}"') |
6271f1ca PH |
743 | self.run_ffmpeg(filename, temp_filename, options) |
744 | ||
d75201a8 | 745 | os.replace(temp_filename, filename) |
6271f1ca | 746 | |
fd7cfb64 | 747 | |
748 | class FFmpegFixupStretchedPP(FFmpegFixupPostProcessor): | |
749 | @PostProcessor._restrict_to(images=False, audio=False) | |
750 | def run(self, info): | |
751 | stretched_ratio = info.get('stretched_ratio') | |
752 | if stretched_ratio not in (None, 1): | |
753 | self._fixup('Fixing aspect ratio', info['filepath'], [ | |
754 | '-c', 'copy', '-map', '0', '-dn', '-aspect', '%f' % stretched_ratio]) | |
592e97e8 | 755 | return [], info |
62cd676c PH |
756 | |
757 | ||
fd7cfb64 | 758 | class FFmpegFixupM4aPP(FFmpegFixupPostProcessor): |
8326b00a | 759 | @PostProcessor._restrict_to(images=False, video=False) |
62cd676c | 760 | def run(self, info): |
fd7cfb64 | 761 | if info.get('container') == 'm4a_dash': |
762 | self._fixup('Correcting container', info['filepath'], [ | |
763 | '-c', 'copy', '-map', '0', '-dn', '-f', 'mp4']) | |
592e97e8 | 764 | return [], info |
e9fade72 JMF |
765 | |
766 | ||
fd7cfb64 | 767 | class FFmpegFixupM3u8PP(FFmpegFixupPostProcessor): |
8326b00a | 768 | @PostProcessor._restrict_to(images=False) |
f17f8651 | 769 | def run(self, info): |
fd7cfb64 | 770 | if self.get_audio_codec(info['filepath']) == 'aac': |
771 | self._fixup('Fixing malformed AAC bitstream', info['filepath'], [ | |
772 | '-c', 'copy', '-map', '0', '-dn', '-f', 'mp4', '-bsf:a', 'aac_adtstoasc']) | |
f17f8651 | 773 | return [], info |
774 | ||
775 | ||
e36d50c5 | 776 | class FFmpegFixupTimestampPP(FFmpegFixupPostProcessor): |
777 | ||
778 | def __init__(self, downloader=None, trim=0.001): | |
779 | # "trim" should be used when the video contains unintended packets | |
780 | super(FFmpegFixupTimestampPP, self).__init__(downloader) | |
781 | assert isinstance(trim, (int, float)) | |
782 | self.trim = str(trim) | |
783 | ||
784 | @PostProcessor._restrict_to(images=False) | |
785 | def run(self, info): | |
786 | required_version = '4.4' | |
787 | if is_outdated_version(self._versions[self.basename], required_version): | |
788 | self.report_warning( | |
789 | 'A re-encode is needed to fix timestamps in older versions of ffmpeg. ' | |
790 | f'Please install ffmpeg {required_version} or later to fixup without re-encoding') | |
791 | opts = ['-vf', 'setpts=PTS-STARTPTS'] | |
792 | else: | |
793 | opts = ['-c', 'copy', '-bsf', 'setts=ts=TS-STARTPTS'] | |
794 | self._fixup('Fixing frame timestamp', info['filepath'], opts + ['-map', '0', '-dn', '-ss', self.trim]) | |
795 | return [], info | |
796 | ||
797 | ||
798 | class FFmpegFixupDurationPP(FFmpegFixupPostProcessor): | |
799 | @PostProcessor._restrict_to(images=False) | |
800 | def run(self, info): | |
801 | self._fixup('Fixing video duration', info['filepath'], ['-c', 'copy', '-map', '0', '-dn']) | |
802 | return [], info | |
803 | ||
804 | ||
e9fade72 | 805 | class FFmpegSubtitlesConvertorPP(FFmpegPostProcessor): |
81a23040 | 806 | SUPPORTED_EXTS = ('srt', 'vtt', 'ass', 'lrc') |
807 | ||
e9fade72 JMF |
808 | def __init__(self, downloader=None, format=None): |
809 | super(FFmpegSubtitlesConvertorPP, self).__init__(downloader) | |
810 | self.format = format | |
811 | ||
812 | def run(self, info): | |
813 | subs = info.get('requested_subtitles') | |
e9fade72 JMF |
814 | new_ext = self.format |
815 | new_format = new_ext | |
816 | if new_format == 'vtt': | |
817 | new_format = 'webvtt' | |
818 | if subs is None: | |
1b77b347 | 819 | self.to_screen('There aren\'t any subtitles to convert') |
592e97e8 | 820 | return [], info |
1b77b347 | 821 | self.to_screen('Converting subtitles') |
e04398e3 | 822 | sub_filenames = [] |
e9fade72 JMF |
823 | for lang, sub in subs.items(): |
824 | ext = sub['ext'] | |
825 | if ext == new_ext: | |
1b77b347 | 826 | self.to_screen('Subtitle file for %s is already in the requested format' % new_ext) |
e9fade72 | 827 | continue |
503d4a44 | 828 | elif ext == 'json': |
1b77b347 | 829 | self.to_screen( |
830 | 'You have requested to convert json subtitles into another format, ' | |
503d4a44 | 831 | 'which is currently not possible') |
832 | continue | |
dcf64d43 | 833 | old_file = sub['filepath'] |
e04398e3 | 834 | sub_filenames.append(old_file) |
dcf64d43 | 835 | new_file = replace_extension(old_file, new_ext) |
bf6427d2 | 836 | |
40fcba5e | 837 | if ext in ('dfxp', 'ttml', 'tt'): |
f446cc66 | 838 | self.report_warning( |
1b77b347 | 839 | 'You have requested to convert dfxp (TTML) subtitles into another format, ' |
bf6427d2 YCH |
840 | 'which results in style information loss') |
841 | ||
e04398e3 | 842 | dfxp_file = old_file |
dcf64d43 | 843 | srt_file = replace_extension(old_file, 'srt') |
bf6427d2 | 844 | |
3869028f | 845 | with open(dfxp_file, 'rb') as f: |
bf6427d2 YCH |
846 | srt_data = dfxp2srt(f.read()) |
847 | ||
848 | with io.open(srt_file, 'wt', encoding='utf-8') as f: | |
849 | f.write(srt_data) | |
7e62c2eb | 850 | old_file = srt_file |
bf6427d2 | 851 | |
bf6427d2 YCH |
852 | subs[lang] = { |
853 | 'ext': 'srt', | |
dcf64d43 | 854 | 'data': srt_data, |
855 | 'filepath': srt_file, | |
bf6427d2 YCH |
856 | } |
857 | ||
858 | if new_ext == 'srt': | |
859 | continue | |
7b8b007c JMF |
860 | else: |
861 | sub_filenames.append(srt_file) | |
bf6427d2 | 862 | |
e04398e3 | 863 | self.run_ffmpeg(old_file, new_file, ['-f', new_format]) |
e9fade72 JMF |
864 | |
865 | with io.open(new_file, 'rt', encoding='utf-8') as f: | |
866 | subs[lang] = { | |
3547d265 | 867 | 'ext': new_ext, |
e9fade72 | 868 | 'data': f.read(), |
dcf64d43 | 869 | 'filepath': new_file, |
e9fade72 JMF |
870 | } |
871 | ||
dcf64d43 | 872 | info['__files_to_move'][new_file] = replace_extension( |
37242e56 | 873 | info['__files_to_move'][sub['filepath']], new_ext) |
dcf64d43 | 874 | |
e04398e3 | 875 | return sub_filenames, info |
72755351 | 876 | |
877 | ||
878 | class FFmpegSplitChaptersPP(FFmpegPostProcessor): | |
7a340e0d NA |
879 | def __init__(self, downloader, force_keyframes=False): |
880 | FFmpegPostProcessor.__init__(self, downloader) | |
881 | self._force_keyframes = force_keyframes | |
72755351 | 882 | |
883 | def _prepare_filename(self, number, chapter, info): | |
884 | info = info.copy() | |
885 | info.update({ | |
886 | 'section_number': number, | |
887 | 'section_title': chapter.get('title'), | |
888 | 'section_start': chapter.get('start_time'), | |
889 | 'section_end': chapter.get('end_time'), | |
890 | }) | |
891 | return self._downloader.prepare_filename(info, 'chapter') | |
892 | ||
893 | def _ffmpeg_args_for_chapter(self, number, chapter, info): | |
894 | destination = self._prepare_filename(number, chapter, info) | |
895 | if not self._downloader._ensure_dir_exists(encodeFilename(destination)): | |
896 | return | |
897 | ||
dcf64d43 | 898 | chapter['filepath'] = destination |
72755351 | 899 | self.to_screen('Chapter %03d; Destination: %s' % (number, destination)) |
900 | return ( | |
901 | destination, | |
902 | ['-ss', compat_str(chapter['start_time']), | |
a94bfd6c | 903 | '-t', compat_str(chapter['end_time'] - chapter['start_time'])]) |
72755351 | 904 | |
8326b00a | 905 | @PostProcessor._restrict_to(images=False) |
72755351 | 906 | def run(self, info): |
907 | chapters = info.get('chapters') or [] | |
908 | if not chapters: | |
7a340e0d | 909 | self.to_screen('Chapter information is unavailable') |
72755351 | 910 | return [], info |
911 | ||
7a340e0d NA |
912 | in_file = info['filepath'] |
913 | if self._force_keyframes and len(chapters) > 1: | |
914 | in_file = self.force_keyframes(in_file, (c['start_time'] for c in chapters)) | |
72755351 | 915 | self.to_screen('Splitting video by chapters; %d chapters found' % len(chapters)) |
916 | for idx, chapter in enumerate(chapters): | |
917 | destination, opts = self._ffmpeg_args_for_chapter(idx + 1, chapter, info) | |
7a340e0d NA |
918 | self.real_run_ffmpeg([(in_file, opts)], [(destination, ['-c', 'copy'])]) |
919 | if in_file != info['filepath']: | |
920 | os.remove(in_file) | |
72755351 | 921 | return [], info |
8fa43c73 | 922 | |
923 | ||
924 | class FFmpegThumbnailsConvertorPP(FFmpegPostProcessor): | |
81a23040 | 925 | SUPPORTED_EXTS = ('jpg', 'png') |
926 | ||
8fa43c73 | 927 | def __init__(self, downloader=None, format=None): |
928 | super(FFmpegThumbnailsConvertorPP, self).__init__(downloader) | |
929 | self.format = format | |
930 | ||
931 | @staticmethod | |
932 | def is_webp(path): | |
933 | with open(encodeFilename(path), 'rb') as f: | |
934 | b = f.read(12) | |
935 | return b[0:4] == b'RIFF' and b[8:] == b'WEBP' | |
936 | ||
937 | def fixup_webp(self, info, idx=-1): | |
938 | thumbnail_filename = info['thumbnails'][idx]['filepath'] | |
939 | _, thumbnail_ext = os.path.splitext(thumbnail_filename) | |
940 | if thumbnail_ext: | |
941 | thumbnail_ext = thumbnail_ext[1:].lower() | |
942 | if thumbnail_ext != 'webp' and self.is_webp(thumbnail_filename): | |
943 | self.to_screen('Correcting thumbnail "%s" extension to webp' % thumbnail_filename) | |
944 | webp_filename = replace_extension(thumbnail_filename, 'webp') | |
d75201a8 | 945 | os.replace(thumbnail_filename, webp_filename) |
8fa43c73 | 946 | info['thumbnails'][idx]['filepath'] = webp_filename |
947 | info['__files_to_move'][webp_filename] = replace_extension( | |
948 | info['__files_to_move'].pop(thumbnail_filename), 'webp') | |
949 | ||
81a23040 | 950 | @staticmethod |
951 | def _options(target_ext): | |
952 | if target_ext == 'jpg': | |
953 | return ['-bsf:v', 'mjpeg2jpeg'] | |
954 | return [] | |
955 | ||
956 | def convert_thumbnail(self, thumbnail_filename, target_ext): | |
81a23040 | 957 | thumbnail_conv_filename = replace_extension(thumbnail_filename, target_ext) |
337e0c62 | 958 | |
959 | self.to_screen('Converting thumbnail "%s" to %s' % (thumbnail_filename, target_ext)) | |
960 | self.real_run_ffmpeg( | |
961 | [(thumbnail_filename, ['-f', 'image2', '-pattern_type', 'none'])], | |
962 | [(thumbnail_conv_filename.replace('%', '%%'), self._options(target_ext))]) | |
a927acb1 | 963 | return thumbnail_conv_filename |
8fa43c73 | 964 | |
965 | def run(self, info): | |
8fa43c73 | 966 | files_to_delete = [] |
967 | has_thumbnail = False | |
968 | ||
969 | for idx, thumbnail_dict in enumerate(info['thumbnails']): | |
970 | if 'filepath' not in thumbnail_dict: | |
971 | continue | |
972 | has_thumbnail = True | |
973 | self.fixup_webp(info, idx) | |
974 | original_thumbnail = thumbnail_dict['filepath'] | |
975 | _, thumbnail_ext = os.path.splitext(original_thumbnail) | |
976 | if thumbnail_ext: | |
977 | thumbnail_ext = thumbnail_ext[1:].lower() | |
15a4fd53 | 978 | if thumbnail_ext == 'jpeg': |
979 | thumbnail_ext = 'jpg' | |
8fa43c73 | 980 | if thumbnail_ext == self.format: |
981 | self.to_screen('Thumbnail "%s" is already in the requested format' % original_thumbnail) | |
982 | continue | |
983 | thumbnail_dict['filepath'] = self.convert_thumbnail(original_thumbnail, self.format) | |
984 | files_to_delete.append(original_thumbnail) | |
985 | info['__files_to_move'][thumbnail_dict['filepath']] = replace_extension( | |
986 | info['__files_to_move'][original_thumbnail], self.format) | |
987 | ||
988 | if not has_thumbnail: | |
989 | self.to_screen('There aren\'t any thumbnails to convert') | |
990 | return files_to_delete, info |