]>
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, |
165efb82 | 18 | float_or_none, |
95807118 | 19 | get_exe_version, |
48844745 | 20 | is_outdated_version, |
7a340e0d NA |
21 | ISO639Utils, |
22 | orderedSet, | |
496c1923 PH |
23 | PostProcessingError, |
24 | prepend_extension, | |
f5b1bca9 | 25 | process_communicate_or_kill, |
06167fbb | 26 | replace_extension, |
7a340e0d | 27 | shell_quote, |
324ad820 | 28 | traverse_obj, |
6606817a | 29 | variadic, |
496c1923 PH |
30 | ) |
31 | ||
32 | ||
a755f825 | 33 | EXT_TO_OUT_FORMATS = { |
21bfcd3d PH |
34 | 'aac': 'adts', |
35 | 'flac': 'flac', | |
36 | 'm4a': 'ipod', | |
37 | 'mka': 'matroska', | |
38 | 'mkv': 'matroska', | |
39 | 'mpg': 'mpeg', | |
40 | 'ogv': 'ogg', | |
41 | 'ts': 'mpegts', | |
42 | 'wma': 'asf', | |
43 | 'wmv': 'asf', | |
44 | } | |
45 | ACODECS = { | |
46 | 'mp3': 'libmp3lame', | |
47 | 'aac': 'aac', | |
48 | 'flac': 'flac', | |
49 | 'm4a': 'aac', | |
d2ae7e24 | 50 | 'opus': 'libopus', |
21bfcd3d PH |
51 | 'vorbis': 'libvorbis', |
52 | 'wav': None, | |
a755f825 | 53 | } |
54 | ||
55 | ||
496c1923 PH |
56 | class FFmpegPostProcessorError(PostProcessingError): |
57 | pass | |
58 | ||
d799b47b | 59 | |
496c1923 | 60 | class FFmpegPostProcessor(PostProcessor): |
d47aeb22 | 61 | def __init__(self, downloader=None): |
496c1923 | 62 | PostProcessor.__init__(self, downloader) |
73fac4e9 | 63 | self._determine_executables() |
496c1923 | 64 | |
48844745 | 65 | def check_version(self): |
f740fae2 | 66 | if not self.available: |
beb4b92a | 67 | raise FFmpegPostProcessorError('ffmpeg not found. Please install or provide the path using --ffmpeg-location') |
48844745 | 68 | |
65bf37ef | 69 | required_version = '10-0' if self.basename == 'avconv' else '1.0' |
48844745 | 70 | if is_outdated_version( |
73fac4e9 | 71 | self._versions[self.basename], required_version): |
3aa578ca | 72 | warning = 'Your copy of %s is outdated, update %s to version %s or newer if you encounter any errors.' % ( |
73fac4e9 | 73 | self.basename, self.basename, required_version) |
f446cc66 | 74 | self.report_warning(warning) |
48844745 | 75 | |
496c1923 | 76 | @staticmethod |
73fac4e9 PH |
77 | def get_versions(downloader=None): |
78 | return FFmpegPostProcessor(downloader)._versions | |
6271f1ca | 79 | |
73fac4e9 PH |
80 | def _determine_executables(self): |
81 | programs = ['avprobe', 'avconv', 'ffmpeg', 'ffprobe'] | |
d4a24f40 | 82 | prefer_ffmpeg = True |
73fac4e9 | 83 | |
a64646e4 RA |
84 | def get_ffmpeg_version(path): |
85 | ver = get_exe_version(path, args=['-version']) | |
86 | if ver: | |
87 | regexs = [ | |
cbdc688c | 88 | r'(?:\d+:)?([0-9.]+)-[0-9]+ubuntu[0-9.]+$', # Ubuntu, see [1] |
5caa531a | 89 | r'n([0-9.]+)$', # Arch Linux |
cbdc688c | 90 | # 1. http://www.ducea.com/2006/06/17/ubuntu-package-version-naming-explanation/ |
a64646e4 RA |
91 | ] |
92 | for regex in regexs: | |
93 | mobj = re.match(regex, ver) | |
94 | if mobj: | |
95 | ver = mobj.group(1) | |
96 | return ver | |
97 | ||
73fac4e9 PH |
98 | self.basename = None |
99 | self.probe_basename = None | |
100 | ||
101 | self._paths = None | |
102 | self._versions = None | |
103 | if self._downloader: | |
f446cc66 | 104 | prefer_ffmpeg = self.get_param('prefer_ffmpeg', True) |
105 | location = self.get_param('ffmpeg_location') | |
73fac4e9 PH |
106 | if location is not None: |
107 | if not os.path.exists(location): | |
f446cc66 | 108 | self.report_warning( |
73fac4e9 | 109 | 'ffmpeg-location %s does not exist! ' |
e4172ac9 | 110 | 'Continuing without ffmpeg.' % (location)) |
73fac4e9 PH |
111 | self._versions = {} |
112 | return | |
8c0ae192 | 113 | elif os.path.isdir(location): |
114 | dirname, basename = location, None | |
115 | else: | |
73fac4e9 | 116 | basename = os.path.splitext(os.path.basename(location))[0] |
89efdc15 | 117 | basename = next((p for p in programs if basename.startswith(p)), 'ffmpeg') |
118 | dirname = os.path.dirname(os.path.abspath(location)) | |
73fac4e9 PH |
119 | if basename in ('ffmpeg', 'ffprobe'): |
120 | prefer_ffmpeg = True | |
121 | ||
122 | self._paths = dict( | |
89efdc15 | 123 | (p, os.path.join(dirname, p)) for p in programs) |
8c0ae192 | 124 | if basename: |
125 | self._paths[basename] = location | |
73fac4e9 | 126 | self._versions = dict( |
a64646e4 | 127 | (p, get_ffmpeg_version(self._paths[p])) for p in programs) |
73fac4e9 PH |
128 | if self._versions is None: |
129 | self._versions = dict( | |
a64646e4 | 130 | (p, get_ffmpeg_version(p)) for p in programs) |
73fac4e9 PH |
131 | self._paths = dict((p, p) for p in programs) |
132 | ||
d4a24f40 | 133 | if prefer_ffmpeg is False: |
d28b5171 | 134 | prefs = ('avconv', 'ffmpeg') |
d4a24f40 S |
135 | else: |
136 | prefs = ('ffmpeg', 'avconv') | |
d28b5171 PH |
137 | for p in prefs: |
138 | if self._versions[p]: | |
73fac4e9 PH |
139 | self.basename = p |
140 | break | |
76b1bd67 | 141 | |
d4a24f40 | 142 | if prefer_ffmpeg is False: |
1a253e13 | 143 | prefs = ('avprobe', 'ffprobe') |
d4a24f40 S |
144 | else: |
145 | prefs = ('ffprobe', 'avprobe') | |
1a253e13 PH |
146 | for p in prefs: |
147 | if self._versions[p]: | |
73fac4e9 PH |
148 | self.probe_basename = p |
149 | break | |
150 | ||
f740fae2 | 151 | @property |
73fac4e9 PH |
152 | def available(self): |
153 | return self.basename is not None | |
1a253e13 | 154 | |
73fac4e9 PH |
155 | @property |
156 | def executable(self): | |
157 | return self._paths[self.basename] | |
158 | ||
3da4b313 JMF |
159 | @property |
160 | def probe_available(self): | |
161 | return self.probe_basename is not None | |
162 | ||
73fac4e9 PH |
163 | @property |
164 | def probe_executable(self): | |
165 | return self._paths[self.probe_basename] | |
76b1bd67 | 166 | |
30d9e209 | 167 | def get_audio_codec(self, path): |
eb35b163 | 168 | if not self.probe_available and not self.available: |
beb4b92a | 169 | raise PostProcessingError('ffprobe and ffmpeg not found. Please install or provide the path using --ffmpeg-location') |
30d9e209 | 170 | try: |
eb35b163 RA |
171 | if self.probe_available: |
172 | cmd = [ | |
173 | encodeFilename(self.probe_executable, True), | |
174 | encodeArgument('-show_streams')] | |
175 | else: | |
176 | cmd = [ | |
177 | encodeFilename(self.executable, True), | |
178 | encodeArgument('-i')] | |
179 | cmd.append(encodeFilename(self._ffmpeg_filename_argument(path), True)) | |
f446cc66 | 180 | self.write_debug('%s command line: %s' % (self.basename, shell_quote(cmd))) |
eb35b163 RA |
181 | handle = subprocess.Popen( |
182 | cmd, stderr=subprocess.PIPE, | |
183 | stdout=subprocess.PIPE, stdin=subprocess.PIPE) | |
f5b1bca9 | 184 | stdout_data, stderr_data = process_communicate_or_kill(handle) |
eb35b163 RA |
185 | expected_ret = 0 if self.probe_available else 1 |
186 | if handle.wait() != expected_ret: | |
30d9e209 RA |
187 | return None |
188 | except (IOError, OSError): | |
189 | return None | |
eb35b163 RA |
190 | output = (stdout_data if self.probe_available else stderr_data).decode('ascii', 'ignore') |
191 | if self.probe_available: | |
192 | audio_codec = None | |
193 | for line in output.split('\n'): | |
194 | if line.startswith('codec_name='): | |
195 | audio_codec = line.split('=')[1].strip() | |
196 | elif line.strip() == 'codec_type=audio' and audio_codec is not None: | |
197 | return audio_codec | |
198 | else: | |
199 | # Stream #FILE_INDEX:STREAM_INDEX[STREAM_ID](LANGUAGE): CODEC_TYPE: CODEC_NAME | |
200 | mobj = re.search( | |
201 | r'Stream\s*#\d+:\d+(?:\[0x[0-9a-f]+\])?(?:\([a-z]{3}\))?:\s*Audio:\s*([0-9a-z]+)', | |
202 | output) | |
203 | if mobj: | |
204 | return mobj.group(1) | |
30d9e209 RA |
205 | return None |
206 | ||
06167fbb | 207 | def get_metadata_object(self, path, opts=[]): |
208 | if self.probe_basename != 'ffprobe': | |
209 | if self.probe_available: | |
210 | self.report_warning('Only ffprobe is supported for metadata extraction') | |
beb4b92a | 211 | raise PostProcessingError('ffprobe not found. Please install or provide the path using --ffmpeg-location') |
06167fbb | 212 | self.check_version() |
213 | ||
214 | cmd = [ | |
215 | encodeFilename(self.probe_executable, True), | |
216 | encodeArgument('-hide_banner'), | |
217 | encodeArgument('-show_format'), | |
218 | encodeArgument('-show_streams'), | |
219 | encodeArgument('-print_format'), | |
220 | encodeArgument('json'), | |
221 | ] | |
222 | ||
223 | cmd += opts | |
224 | cmd.append(encodeFilename(self._ffmpeg_filename_argument(path), True)) | |
06869367 | 225 | self.write_debug('ffprobe command line: %s' % shell_quote(cmd)) |
06167fbb | 226 | p = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, stdin=subprocess.PIPE) |
227 | stdout, stderr = p.communicate() | |
228 | return json.loads(stdout.decode('utf-8', 'replace')) | |
229 | ||
230 | def get_stream_number(self, path, keys, value): | |
231 | streams = self.get_metadata_object(path)['streams'] | |
232 | num = next( | |
324ad820 | 233 | (i for i, stream in enumerate(streams) if traverse_obj(stream, keys, casesense=False) == value), |
06167fbb | 234 | None) |
235 | return num, len(streams) | |
236 | ||
165efb82 | 237 | def _get_real_video_duration(self, info, fatal=True): |
238 | try: | |
239 | if '_real_duration' not in info: | |
240 | info['_real_duration'] = float_or_none( | |
241 | traverse_obj(self.get_metadata_object(info['filepath']), ('format', 'duration'))) | |
242 | if not info['_real_duration']: | |
243 | raise PostProcessingError('ffprobe returned empty duration') | |
244 | except PostProcessingError as e: | |
245 | if fatal: | |
246 | raise PostProcessingError(f'Unable to determine video duration; {e}') | |
247 | return info.setdefault('_real_duration', None) | |
248 | ||
249 | def _duration_mismatch(self, d1, d2): | |
250 | if not d1 or not d2: | |
251 | return None | |
252 | return abs(d1 - d2) > 1 | |
253 | ||
00034c14 | 254 | def run_ffmpeg_multiple_files(self, input_paths, out_path, opts, **kwargs): |
e92caff5 | 255 | return self.real_run_ffmpeg( |
256 | [(path, []) for path in input_paths], | |
00034c14 | 257 | [(out_path, opts)], **kwargs) |
e92caff5 | 258 | |
00034c14 | 259 | def real_run_ffmpeg(self, input_path_opts, output_path_opts, *, expected_retcodes=(0,)): |
48844745 | 260 | self.check_version() |
496c1923 | 261 | |
52afb2ac | 262 | oldest_mtime = min( |
7dde84f3 | 263 | os.stat(encodeFilename(path)).st_mtime for path, _ in input_path_opts if path) |
43bc8890 | 264 | |
91b6c884 | 265 | cmd = [encodeFilename(self.executable, True), encodeArgument('-y')] |
ce52c7c1 S |
266 | # avconv does not have repeat option |
267 | if self.basename == 'ffmpeg': | |
268 | cmd += [encodeArgument('-loglevel'), encodeArgument('repeat+info')] | |
5b1ecbb3 | 269 | |
e92caff5 | 270 | def make_args(file, args, name, number): |
271 | keys = ['_%s%d' % (name, number), '_%s' % name] | |
272 | if name == 'o' and number == 1: | |
273 | keys.append('') | |
274 | args += self._configuration_args(self.basename, keys) | |
275 | if name == 'i': | |
276 | args.append('-i') | |
5b1ecbb3 | 277 | return ( |
e92caff5 | 278 | [encodeArgument(arg) for arg in args] |
5b1ecbb3 | 279 | + [encodeFilename(self._ffmpeg_filename_argument(file), True)]) |
280 | ||
e92caff5 | 281 | for arg_type, path_opts in (('i', input_path_opts), ('o', output_path_opts)): |
7dde84f3 | 282 | cmd += itertools.chain.from_iterable( |
283 | make_args(path, list(opts), arg_type, i + 1) | |
284 | for i, (path, opts) in enumerate(path_opts) if path) | |
496c1923 | 285 | |
f446cc66 | 286 | self.write_debug('ffmpeg command line: %s' % shell_quote(cmd)) |
cffcbc02 | 287 | p = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, stdin=subprocess.PIPE) |
f5b1bca9 | 288 | stdout, stderr = process_communicate_or_kill(p) |
00034c14 | 289 | if p.returncode not in variadic(expected_retcodes): |
06167fbb | 290 | stderr = stderr.decode('utf-8', 'replace').strip() |
b1940459 | 291 | self.write_debug(stderr) |
06167fbb | 292 | raise FFmpegPostProcessorError(stderr.split('\n')[-1]) |
e92caff5 | 293 | for out_path, _ in output_path_opts: |
7dde84f3 | 294 | if out_path: |
295 | self.try_utime(out_path, oldest_mtime, oldest_mtime) | |
06167fbb | 296 | return stderr.decode('utf-8', 'replace') |
cc55d088 | 297 | |
00034c14 | 298 | def run_ffmpeg(self, path, out_path, opts, **kwargs): |
299 | return self.run_ffmpeg_multiple_files([path], out_path, opts, **kwargs) | |
496c1923 | 300 | |
7a340e0d NA |
301 | @staticmethod |
302 | def _ffmpeg_filename_argument(fn): | |
8a7bbd16 JMF |
303 | # Always use 'file:' because the filename may contain ':' (ffmpeg |
304 | # interprets that as a protocol) or can start with '-' (-- is broken in | |
305 | # ffmpeg, see https://ffmpeg.org/trac/ffmpeg/ticket/2127 for details) | |
b9f2fdd3 | 306 | # Also leave '-' intact in order not to break streaming to stdout. |
06167fbb | 307 | if fn.startswith(('http://', 'https://')): |
308 | return fn | |
d868f43c | 309 | return 'file:' + fn if fn != '-' else fn |
496c1923 | 310 | |
7a340e0d NA |
311 | @staticmethod |
312 | def _quote_for_ffmpeg(string): | |
313 | # See https://ffmpeg.org/ffmpeg-utils.html#toc-Quoting-and-escaping | |
314 | # A sequence of '' produces '\'''\''; | |
315 | # final replace removes the empty '' between \' \'. | |
316 | string = string.replace("'", r"'\''").replace("'''", "'") | |
317 | # Handle potential ' at string boundaries. | |
318 | string = string[1:] if string[0] == "'" else "'" + string | |
319 | return string[:-1] if string[-1] == "'" else string + "'" | |
320 | ||
321 | def force_keyframes(self, filename, timestamps): | |
322 | timestamps = orderedSet(timestamps) | |
323 | if timestamps[0] == 0: | |
324 | timestamps = timestamps[1:] | |
325 | keyframe_file = prepend_extension(filename, 'keyframes.temp') | |
326 | self.to_screen(f'Re-encoding "{filename}" with appropriate keyframes') | |
327 | self.run_ffmpeg(filename, keyframe_file, ['-force_key_frames', ','.join( | |
328 | f'{t:.6f}' for t in timestamps)]) | |
329 | return keyframe_file | |
330 | ||
331 | def concat_files(self, in_files, out_file, concat_opts=None): | |
332 | """ | |
333 | Use concat demuxer to concatenate multiple files having identical streams. | |
334 | ||
335 | Only inpoint, outpoint, and duration concat options are supported. | |
336 | See https://ffmpeg.org/ffmpeg-formats.html#concat-1 for details | |
337 | """ | |
338 | concat_file = f'{out_file}.concat' | |
339 | self.write_debug(f'Writing concat spec to {concat_file}') | |
340 | with open(concat_file, 'wt', encoding='utf-8') as f: | |
341 | f.writelines(self._concat_spec(in_files, concat_opts)) | |
342 | ||
343 | out_flags = ['-c', 'copy'] | |
344 | if out_file.rpartition('.')[-1] in ('mp4', 'mov'): | |
345 | # For some reason, '-c copy' is not enough to copy subtitles | |
346 | out_flags.extend(['-c:s', 'mov_text', '-movflags', '+faststart']) | |
347 | ||
348 | try: | |
349 | self.real_run_ffmpeg( | |
350 | [(concat_file, ['-hide_banner', '-nostdin', '-f', 'concat', '-safe', '0'])], | |
351 | [(out_file, out_flags)]) | |
352 | finally: | |
353 | os.remove(concat_file) | |
354 | ||
355 | @classmethod | |
356 | def _concat_spec(cls, in_files, concat_opts=None): | |
357 | if concat_opts is None: | |
358 | concat_opts = [{}] * len(in_files) | |
359 | yield 'ffconcat version 1.0\n' | |
360 | for file, opts in zip(in_files, concat_opts): | |
361 | yield f'file {cls._quote_for_ffmpeg(cls._ffmpeg_filename_argument(file))}\n' | |
362 | # Iterate explicitly to yield the following directives in order, ignoring the rest. | |
363 | for directive in 'inpoint', 'outpoint', 'duration': | |
364 | if directive in opts: | |
365 | yield f'{directive} {opts[directive]}\n' | |
366 | ||
496c1923 PH |
367 | |
368 | class FFmpegExtractAudioPP(FFmpegPostProcessor): | |
81a23040 | 369 | COMMON_AUDIO_EXTS = ('wav', 'flac', 'm4a', 'aiff', 'mp3', 'ogg', 'mka', 'opus', 'wma') |
370 | SUPPORTED_EXTS = ('best', 'aac', 'flac', 'mp3', 'm4a', 'opus', 'vorbis', 'wav') | |
1de75fa1 | 371 | |
496c1923 PH |
372 | def __init__(self, downloader=None, preferredcodec=None, preferredquality=None, nopostoverwrites=False): |
373 | FFmpegPostProcessor.__init__(self, downloader) | |
81a23040 | 374 | self._preferredcodec = preferredcodec or 'best' |
496c1923 PH |
375 | self._preferredquality = preferredquality |
376 | self._nopostoverwrites = nopostoverwrites | |
377 | ||
496c1923 | 378 | def run_ffmpeg(self, path, out_path, codec, more_opts): |
496c1923 PH |
379 | if codec is None: |
380 | acodec_opts = [] | |
381 | else: | |
382 | acodec_opts = ['-acodec', codec] | |
383 | opts = ['-vn'] + acodec_opts + more_opts | |
384 | try: | |
385 | FFmpegPostProcessor.run_ffmpeg(self, path, out_path, opts) | |
386 | except FFmpegPostProcessorError as err: | |
387 | raise AudioConversionError(err.msg) | |
388 | ||
8326b00a | 389 | @PostProcessor._restrict_to(images=False) |
496c1923 PH |
390 | def run(self, information): |
391 | path = information['filepath'] | |
1de75fa1 | 392 | orig_ext = information['ext'] |
393 | ||
81a23040 | 394 | if self._preferredcodec == 'best' and orig_ext in self.COMMON_AUDIO_EXTS: |
1de75fa1 | 395 | self.to_screen('Skipping audio extraction since the file is already in a common audio format') |
55b53b33 | 396 | return [], information |
496c1923 PH |
397 | |
398 | filecodec = self.get_audio_codec(path) | |
399 | if filecodec is None: | |
3aa578ca | 400 | raise PostProcessingError('WARNING: unable to obtain file audio codec with ffprobe') |
496c1923 PH |
401 | |
402 | more_opts = [] | |
403 | if self._preferredcodec == 'best' or self._preferredcodec == filecodec or (self._preferredcodec == 'm4a' and filecodec == 'aac'): | |
404 | if filecodec == 'aac' and self._preferredcodec in ['m4a', 'best']: | |
405 | # Lossless, but in another container | |
406 | acodec = 'copy' | |
407 | extension = 'm4a' | |
467d3c9a | 408 | more_opts = ['-bsf:a', 'aac_adtstoasc'] |
21bfcd3d | 409 | elif filecodec in ['aac', 'flac', 'mp3', 'vorbis', 'opus']: |
496c1923 PH |
410 | # Lossless if possible |
411 | acodec = 'copy' | |
412 | extension = filecodec | |
413 | if filecodec == 'aac': | |
414 | more_opts = ['-f', 'adts'] | |
415 | if filecodec == 'vorbis': | |
416 | extension = 'ogg' | |
417 | else: | |
418 | # MP3 otherwise. | |
419 | acodec = 'libmp3lame' | |
420 | extension = 'mp3' | |
421 | more_opts = [] | |
422 | if self._preferredquality is not None: | |
423 | if int(self._preferredquality) < 10: | |
467d3c9a | 424 | more_opts += ['-q:a', self._preferredquality] |
496c1923 | 425 | else: |
467d3c9a | 426 | more_opts += ['-b:a', self._preferredquality + 'k'] |
496c1923 | 427 | else: |
21bfcd3d PH |
428 | # We convert the audio (lossy if codec is lossy) |
429 | acodec = ACODECS[self._preferredcodec] | |
496c1923 PH |
430 | extension = self._preferredcodec |
431 | more_opts = [] | |
432 | if self._preferredquality is not None: | |
433 | # The opus codec doesn't support the -aq option | |
434 | if int(self._preferredquality) < 10 and extension != 'opus': | |
467d3c9a | 435 | more_opts += ['-q:a', self._preferredquality] |
496c1923 | 436 | else: |
467d3c9a | 437 | more_opts += ['-b:a', self._preferredquality + 'k'] |
496c1923 PH |
438 | if self._preferredcodec == 'aac': |
439 | more_opts += ['-f', 'adts'] | |
440 | if self._preferredcodec == 'm4a': | |
467d3c9a | 441 | more_opts += ['-bsf:a', 'aac_adtstoasc'] |
496c1923 PH |
442 | if self._preferredcodec == 'vorbis': |
443 | extension = 'ogg' | |
444 | if self._preferredcodec == 'wav': | |
445 | extension = 'wav' | |
446 | more_opts += ['-f', 'wav'] | |
447 | ||
3aa578ca | 448 | prefix, sep, ext = path.rpartition('.') # not os.path.splitext, since the latter does not work on unicode in all setups |
496c1923 | 449 | new_path = prefix + sep + extension |
0b94dbb1 | 450 | |
2273e2c5 PM |
451 | information['filepath'] = new_path |
452 | information['ext'] = extension | |
496c1923 PH |
453 | |
454 | # If we download foo.mp3 and convert it to... foo.mp3, then don't delete foo.mp3, silly. | |
3089bc74 S |
455 | if (new_path == path |
456 | or (self._nopostoverwrites and os.path.exists(encodeFilename(new_path)))): | |
1b77b347 | 457 | self.to_screen('Post-process file %s exists, skipping' % new_path) |
592e97e8 | 458 | return [], information |
496c1923 PH |
459 | |
460 | try: | |
1b77b347 | 461 | self.to_screen('Destination: ' + new_path) |
ce81b141 | 462 | self.run_ffmpeg(path, new_path, acodec, more_opts) |
70a1165b JMF |
463 | except AudioConversionError as e: |
464 | raise PostProcessingError( | |
465 | 'audio conversion failed: ' + e.msg) | |
466 | except Exception: | |
467 | raise PostProcessingError('error running ' + self.basename) | |
496c1923 PH |
468 | |
469 | # Try to update the date time for extracted audio file. | |
470 | if information.get('filetime') is not None: | |
dd29eb7f S |
471 | self.try_utime( |
472 | new_path, time.time(), information['filetime'], | |
473 | errnote='Cannot update utime of audio file') | |
496c1923 | 474 | |
592e97e8 | 475 | return [path], information |
496c1923 PH |
476 | |
477 | ||
857f6313 | 478 | class FFmpegVideoConvertorPP(FFmpegPostProcessor): |
81a23040 | 479 | SUPPORTED_EXTS = ('mp4', 'mkv', 'flv', 'webm', 'mov', 'avi', 'mp3', 'mka', 'm4a', 'ogg', 'opus') |
480 | FORMAT_RE = re.compile(r'{0}(?:/{0})*$'.format(r'(?:\w+>)?(?:%s)' % '|'.join(SUPPORTED_EXTS))) | |
e6f21b3d | 481 | _ACTION = 'converting' |
857f6313 | 482 | |
efe87a10 | 483 | def __init__(self, downloader=None, preferedformat=None): |
857f6313 | 484 | super(FFmpegVideoConvertorPP, self).__init__(downloader) |
06167fbb | 485 | self._preferedformats = preferedformat.lower().split('/') |
efe87a10 | 486 | |
857f6313 | 487 | def _target_ext(self, source_ext): |
06167fbb | 488 | for pair in self._preferedformats: |
489 | kv = pair.split('>') | |
857f6313 | 490 | if len(kv) == 1 or kv[0].strip() == source_ext: |
491 | return kv[-1].strip() | |
06167fbb | 492 | |
857f6313 | 493 | @staticmethod |
494 | def _options(target_ext): | |
495 | if target_ext == 'avi': | |
496 | return ['-c:v', 'libxvid', '-vtag', 'XVID'] | |
497 | return [] | |
498 | ||
8326b00a | 499 | @PostProcessor._restrict_to(images=False) |
e6f21b3d | 500 | def run(self, info): |
501 | filename, source_ext = info['filepath'], info['ext'].lower() | |
81a23040 | 502 | target_ext = self._target_ext(source_ext) |
06167fbb | 503 | _skip_msg = ( |
e6f21b3d | 504 | f'could not find a mapping for {source_ext}' if not target_ext |
505 | else f'already is in target format {source_ext}' if source_ext == target_ext | |
06167fbb | 506 | else None) |
507 | if _skip_msg: | |
e6f21b3d | 508 | self.to_screen(f'Not {self._ACTION} media file {filename!r}; {_skip_msg}') |
509 | return [], info | |
06167fbb | 510 | |
e6f21b3d | 511 | outpath = replace_extension(filename, target_ext, source_ext) |
512 | self.to_screen(f'{self._ACTION.title()} video from {source_ext} to {target_ext}; Destination: {outpath}') | |
513 | self.run_ffmpeg(filename, outpath, self._options(target_ext)) | |
857f6313 | 514 | |
e6f21b3d | 515 | info['filepath'] = outpath |
516 | info['format'] = info['ext'] = target_ext | |
517 | return [filename], info | |
efe87a10 FS |
518 | |
519 | ||
857f6313 | 520 | class FFmpegVideoRemuxerPP(FFmpegVideoConvertorPP): |
e6f21b3d | 521 | _ACTION = 'remuxing' |
496c1923 | 522 | |
857f6313 | 523 | @staticmethod |
524 | def _options(target_ext): | |
525 | options = ['-c', 'copy', '-map', '0', '-dn'] | |
526 | if target_ext in ['mp4', 'm4a', 'mov']: | |
527 | options.extend(['-movflags', '+faststart']) | |
528 | return options | |
496c1923 PH |
529 | |
530 | ||
531 | class FFmpegEmbedSubtitlePP(FFmpegPostProcessor): | |
cffab0ee | 532 | def __init__(self, downloader=None, already_have_subtitle=False): |
533 | super(FFmpegEmbedSubtitlePP, self).__init__(downloader) | |
534 | self._already_have_subtitle = already_have_subtitle | |
535 | ||
8326b00a | 536 | @PostProcessor._restrict_to(images=False) |
496c1923 | 537 | def run(self, information): |
40025ee2 | 538 | if information['ext'] not in ('mp4', 'webm', 'mkv'): |
1b77b347 | 539 | self.to_screen('Subtitles can only be embedded in mp4, webm or mkv files') |
592e97e8 | 540 | return [], information |
c84dd8a9 JMF |
541 | subtitles = information.get('requested_subtitles') |
542 | if not subtitles: | |
1b77b347 | 543 | self.to_screen('There aren\'t any subtitles to embed') |
592e97e8 | 544 | return [], information |
496c1923 | 545 | |
496c1923 | 546 | filename = information['filepath'] |
1f2a268b | 547 | if information.get('duration') and self._duration_mismatch( |
165efb82 | 548 | self._get_real_video_duration(information, False), information['duration']): |
549 | self.to_screen(f'Skipping {self.pp_key()} since the real and expected durations mismatch') | |
550 | return [], information | |
40025ee2 S |
551 | |
552 | ext = information['ext'] | |
2412044c | 553 | sub_langs, sub_names, sub_filenames = [], [], [] |
40025ee2 | 554 | webm_vtt_warn = False |
06167fbb | 555 | mp4_ass_warn = False |
40025ee2 S |
556 | |
557 | for lang, sub_info in subtitles.items(): | |
a1c39673 | 558 | if not os.path.exists(sub_info.get('filepath', '')): |
8e25d624 | 559 | self.report_warning(f'Skipping embedding {lang} subtitle because the file is missing') |
560 | continue | |
40025ee2 | 561 | sub_ext = sub_info['ext'] |
503d4a44 | 562 | if sub_ext == 'json': |
06167fbb | 563 | self.report_warning('JSON subtitles cannot be embedded') |
503d4a44 | 564 | elif ext != 'webm' or ext == 'webm' and sub_ext == 'vtt': |
40025ee2 | 565 | sub_langs.append(lang) |
2412044c | 566 | sub_names.append(sub_info.get('name')) |
dcf64d43 | 567 | sub_filenames.append(sub_info['filepath']) |
40025ee2 S |
568 | else: |
569 | if not webm_vtt_warn and ext == 'webm' and sub_ext != 'vtt': | |
570 | webm_vtt_warn = True | |
06167fbb | 571 | self.report_warning('Only WebVTT subtitles can be embedded in webm files') |
572 | if not mp4_ass_warn and ext == 'mp4' and sub_ext == 'ass': | |
573 | mp4_ass_warn = True | |
574 | self.report_warning('ASS subtitles cannot be properly embedded in mp4 files; expect issues') | |
40025ee2 S |
575 | |
576 | if not sub_langs: | |
577 | return [], information | |
578 | ||
14523ed9 | 579 | input_files = [filename] + sub_filenames |
496c1923 | 580 | |
e205db3b | 581 | opts = [ |
e0da59fe | 582 | '-c', 'copy', '-map', '0', '-dn', |
e205db3b JMF |
583 | # Don't copy the existing subtitles, we may be running the |
584 | # postprocessor a second time | |
585 | '-map', '-0:s', | |
7f903dd8 T |
586 | # Don't copy Apple TV chapters track, bin_data (see #19042, #19024, |
587 | # https://trac.ffmpeg.org/ticket/6016) | |
588 | '-map', '-0:d', | |
e205db3b | 589 | ] |
083c1bb9 N |
590 | if information['ext'] == 'mp4': |
591 | opts += ['-c:s', 'mov_text'] | |
2412044c | 592 | for i, (lang, name) in enumerate(zip(sub_langs, sub_names)): |
2875cf01 | 593 | opts.extend(['-map', '%d:0' % (i + 1)]) |
04fb6928 S |
594 | lang_code = ISO639Utils.short2long(lang) or lang |
595 | opts.extend(['-metadata:s:s:%d' % i, 'language=%s' % lang_code]) | |
2412044c | 596 | if name: |
597 | opts.extend(['-metadata:s:s:%d' % i, 'handler_name=%s' % name, | |
598 | '-metadata:s:s:%d' % i, 'title=%s' % name]) | |
496c1923 | 599 | |
2875cf01 | 600 | temp_filename = prepend_extension(filename, 'temp') |
06167fbb | 601 | self.to_screen('Embedding subtitles in "%s"' % filename) |
496c1923 | 602 | self.run_ffmpeg_multiple_files(input_files, temp_filename, opts) |
d75201a8 | 603 | os.replace(temp_filename, filename) |
496c1923 | 604 | |
cffab0ee | 605 | files_to_delete = [] if self._already_have_subtitle else sub_filenames |
606 | return files_to_delete, information | |
496c1923 PH |
607 | |
608 | ||
609 | class FFmpegMetadataPP(FFmpegPostProcessor): | |
7dde84f3 | 610 | |
7a340e0d NA |
611 | def __init__(self, downloader, add_metadata=True, add_chapters=True): |
612 | FFmpegPostProcessor.__init__(self, downloader) | |
613 | self._add_metadata = add_metadata | |
614 | self._add_chapters = add_chapters | |
615 | ||
7dde84f3 | 616 | @staticmethod |
617 | def _options(target_ext): | |
618 | yield from ('-map', '0', '-dn') | |
619 | if target_ext == 'm4a': | |
620 | yield from ('-vn', '-acodec', 'copy') | |
621 | else: | |
622 | yield from ('-c', 'copy') | |
623 | ||
8326b00a | 624 | @PostProcessor._restrict_to(images=False) |
496c1923 | 625 | def run(self, info): |
7a340e0d NA |
626 | filename, metadata_filename = info['filepath'], None |
627 | options = [] | |
628 | if self._add_chapters and info.get('chapters'): | |
629 | metadata_filename = replace_extension(filename, 'meta') | |
630 | options.extend(self._get_chapter_opts(info['chapters'], metadata_filename)) | |
631 | if self._add_metadata: | |
632 | options.extend(self._get_metadata_opts(info)) | |
633 | ||
634 | if not options: | |
635 | self.to_screen('There isn\'t any metadata to add') | |
636 | return [], info | |
637 | ||
638 | temp_filename = prepend_extension(filename, 'temp') | |
639 | self.to_screen('Adding metadata to "%s"' % filename) | |
640 | self.run_ffmpeg_multiple_files( | |
641 | (filename, metadata_filename), temp_filename, | |
642 | itertools.chain(self._options(info['ext']), *options)) | |
643 | if metadata_filename: | |
644 | os.remove(metadata_filename) | |
645 | os.replace(temp_filename, filename) | |
646 | return [], info | |
647 | ||
648 | @staticmethod | |
649 | def _get_chapter_opts(chapters, metadata_filename): | |
650 | with io.open(metadata_filename, 'wt', encoding='utf-8') as f: | |
651 | def ffmpeg_escape(text): | |
652 | return re.sub(r'([\\=;#\n])', r'\\\1', text) | |
653 | ||
654 | metadata_file_content = ';FFMETADATA1\n' | |
655 | for chapter in chapters: | |
656 | metadata_file_content += '[CHAPTER]\nTIMEBASE=1/1000\n' | |
657 | metadata_file_content += 'START=%d\n' % (chapter['start_time'] * 1000) | |
658 | metadata_file_content += 'END=%d\n' % (chapter['end_time'] * 1000) | |
659 | chapter_title = chapter.get('title') | |
660 | if chapter_title: | |
661 | metadata_file_content += 'title=%s\n' % ffmpeg_escape(chapter_title) | |
662 | f.write(metadata_file_content) | |
663 | yield ('-map_metadata', '1') | |
664 | ||
665 | def _get_metadata_opts(self, info): | |
496c1923 | 666 | metadata = {} |
4bd143a3 S |
667 | |
668 | def add(meta_list, info_list=None): | |
84601bb7 | 669 | if not meta_list: |
670 | return | |
6606817a | 671 | for info_f in variadic(info_list or meta_list): |
84601bb7 | 672 | if isinstance(info.get(info_f), (compat_str, compat_numeric_types)): |
6606817a | 673 | for meta_f in variadic(meta_list): |
4bd143a3 S |
674 | metadata[meta_f] = info[info_f] |
675 | break | |
676 | ||
2791e80b S |
677 | # See [1-4] for some info on media metadata/metadata supported |
678 | # by ffmpeg. | |
679 | # 1. https://kdenlive.org/en/project/adding-meta-data-to-mp4-video/ | |
680 | # 2. https://wiki.multimedia.cx/index.php/FFmpeg_Metadata | |
681 | # 3. https://kodi.wiki/view/Video_file_tagging | |
2791e80b | 682 | |
4bd143a3 S |
683 | add('title', ('track', 'title')) |
684 | add('date', 'upload_date') | |
cd9b384c | 685 | add(('description', 'synopsis'), 'description') |
686 | add(('purl', 'comment'), 'webpage_url') | |
4bd143a3 S |
687 | add('track', 'track_number') |
688 | add('artist', ('artist', 'creator', 'uploader', 'uploader_id')) | |
689 | add('genre') | |
690 | add('album') | |
691 | add('album_artist') | |
692 | add('disc', 'disc_number') | |
2791e80b S |
693 | add('show', 'series') |
694 | add('season_number') | |
695 | add('episode_id', ('episode', 'episode_id')) | |
696 | add('episode_sort', 'episode_number') | |
496c1923 | 697 | |
84601bb7 | 698 | prefix = 'meta_' |
699 | for key in filter(lambda k: k.startswith(prefix), info.keys()): | |
700 | add(key[len(prefix):], key) | |
701 | ||
7a340e0d NA |
702 | for name, value in metadata.items(): |
703 | yield ('-metadata', f'{name}={value}') | |
39c68260 | 704 | |
7dde84f3 | 705 | stream_idx = 0 |
706 | for fmt in info.get('requested_formats') or []: | |
707 | stream_count = 2 if 'none' not in (fmt.get('vcodec'), fmt.get('acodec')) else 1 | |
708 | if fmt.get('language'): | |
709 | lang = ISO639Utils.short2long(fmt['language']) or fmt['language'] | |
7a340e0d NA |
710 | for i in range(stream_count): |
711 | yield ('-metadata:s:%d' % (stream_idx + i), 'language=%s' % lang) | |
7dde84f3 | 712 | stream_idx += stream_count |
496c1923 | 713 | |
41712218 | 714 | if ('no-attach-info-json' not in self.get_param('compat_opts', []) |
715 | and '__infojson_filename' in info and info['ext'] in ('mkv', 'mka')): | |
7a340e0d | 716 | old_stream, new_stream = self.get_stream_number(info['filepath'], ('tags', 'mimetype'), 'application/json') |
06167fbb | 717 | if old_stream is not None: |
7a340e0d | 718 | yield ('-map', '-0:%d' % old_stream) |
06167fbb | 719 | new_stream -= 1 |
720 | ||
7a340e0d NA |
721 | yield ('-attach', info['__infojson_filename'], |
722 | '-metadata:s:%d' % new_stream, 'mimetype=application/json') | |
496c1923 PH |
723 | |
724 | ||
725 | class FFmpegMergerPP(FFmpegPostProcessor): | |
8326b00a | 726 | @PostProcessor._restrict_to(images=False) |
496c1923 PH |
727 | def run(self, info): |
728 | filename = info['filepath'] | |
5b5fbc08 | 729 | temp_filename = prepend_extension(filename, 'temp') |
d03cfdce | 730 | args = ['-c', 'copy'] |
50eff38c | 731 | audio_streams = 0 |
d03cfdce | 732 | for (i, fmt) in enumerate(info['requested_formats']): |
733 | if fmt.get('acodec') != 'none': | |
a21e0ab1 | 734 | args.extend(['-map', f'{i}:a:0']) |
9dda99f2 | 735 | aac_fixup = fmt['protocol'].startswith('m3u8') and self.get_audio_codec(fmt['filepath']) == 'aac' |
736 | if aac_fixup: | |
50eff38c | 737 | args.extend([f'-bsf:a:{audio_streams}', 'aac_adtstoasc']) |
738 | audio_streams += 1 | |
d03cfdce | 739 | if fmt.get('vcodec') != 'none': |
740 | args.extend(['-map', '%u:v:0' % (i)]) | |
1b77b347 | 741 | self.to_screen('Merging formats into "%s"' % filename) |
5b5fbc08 JMF |
742 | self.run_ffmpeg_multiple_files(info['__files_to_merge'], temp_filename, args) |
743 | os.rename(encodeFilename(temp_filename), encodeFilename(filename)) | |
d47aeb22 | 744 | return info['__files_to_merge'], info |
496c1923 | 745 | |
13763ce5 S |
746 | def can_merge(self): |
747 | # TODO: figure out merge-capable ffmpeg version | |
748 | if self.basename != 'avconv': | |
749 | return True | |
750 | ||
751 | required_version = '10-0' | |
752 | if is_outdated_version( | |
753 | self._versions[self.basename], required_version): | |
754 | warning = ('Your copy of %s is outdated and unable to properly mux separate video and audio files, ' | |
7a5c1cfe | 755 | 'yt-dlp will download single file media. ' |
13763ce5 S |
756 | 'Update %s to version %s or newer to fix this.') % ( |
757 | self.basename, self.basename, required_version) | |
f446cc66 | 758 | self.report_warning(warning) |
13763ce5 S |
759 | return False |
760 | return True | |
761 | ||
0c14e2fb | 762 | |
fd7cfb64 | 763 | class FFmpegFixupPostProcessor(FFmpegPostProcessor): |
764 | def _fixup(self, msg, filename, options): | |
6271f1ca PH |
765 | temp_filename = prepend_extension(filename, 'temp') |
766 | ||
f89b3e2d | 767 | self.to_screen(f'{msg} of "{filename}"') |
6271f1ca PH |
768 | self.run_ffmpeg(filename, temp_filename, options) |
769 | ||
d75201a8 | 770 | os.replace(temp_filename, filename) |
6271f1ca | 771 | |
fd7cfb64 | 772 | |
773 | class FFmpegFixupStretchedPP(FFmpegFixupPostProcessor): | |
774 | @PostProcessor._restrict_to(images=False, audio=False) | |
775 | def run(self, info): | |
776 | stretched_ratio = info.get('stretched_ratio') | |
777 | if stretched_ratio not in (None, 1): | |
778 | self._fixup('Fixing aspect ratio', info['filepath'], [ | |
779 | '-c', 'copy', '-map', '0', '-dn', '-aspect', '%f' % stretched_ratio]) | |
592e97e8 | 780 | return [], info |
62cd676c PH |
781 | |
782 | ||
fd7cfb64 | 783 | class FFmpegFixupM4aPP(FFmpegFixupPostProcessor): |
8326b00a | 784 | @PostProcessor._restrict_to(images=False, video=False) |
62cd676c | 785 | def run(self, info): |
fd7cfb64 | 786 | if info.get('container') == 'm4a_dash': |
787 | self._fixup('Correcting container', info['filepath'], [ | |
788 | '-c', 'copy', '-map', '0', '-dn', '-f', 'mp4']) | |
592e97e8 | 789 | return [], info |
e9fade72 JMF |
790 | |
791 | ||
fd7cfb64 | 792 | class FFmpegFixupM3u8PP(FFmpegFixupPostProcessor): |
8326b00a | 793 | @PostProcessor._restrict_to(images=False) |
f17f8651 | 794 | def run(self, info): |
fd7cfb64 | 795 | if self.get_audio_codec(info['filepath']) == 'aac': |
796 | self._fixup('Fixing malformed AAC bitstream', info['filepath'], [ | |
797 | '-c', 'copy', '-map', '0', '-dn', '-f', 'mp4', '-bsf:a', 'aac_adtstoasc']) | |
f17f8651 | 798 | return [], info |
799 | ||
800 | ||
e36d50c5 | 801 | class FFmpegFixupTimestampPP(FFmpegFixupPostProcessor): |
802 | ||
803 | def __init__(self, downloader=None, trim=0.001): | |
804 | # "trim" should be used when the video contains unintended packets | |
805 | super(FFmpegFixupTimestampPP, self).__init__(downloader) | |
806 | assert isinstance(trim, (int, float)) | |
807 | self.trim = str(trim) | |
808 | ||
809 | @PostProcessor._restrict_to(images=False) | |
810 | def run(self, info): | |
811 | required_version = '4.4' | |
812 | if is_outdated_version(self._versions[self.basename], required_version): | |
813 | self.report_warning( | |
814 | 'A re-encode is needed to fix timestamps in older versions of ffmpeg. ' | |
815 | f'Please install ffmpeg {required_version} or later to fixup without re-encoding') | |
816 | opts = ['-vf', 'setpts=PTS-STARTPTS'] | |
817 | else: | |
818 | opts = ['-c', 'copy', '-bsf', 'setts=ts=TS-STARTPTS'] | |
819 | self._fixup('Fixing frame timestamp', info['filepath'], opts + ['-map', '0', '-dn', '-ss', self.trim]) | |
820 | return [], info | |
821 | ||
822 | ||
823 | class FFmpegFixupDurationPP(FFmpegFixupPostProcessor): | |
824 | @PostProcessor._restrict_to(images=False) | |
825 | def run(self, info): | |
826 | self._fixup('Fixing video duration', info['filepath'], ['-c', 'copy', '-map', '0', '-dn']) | |
827 | return [], info | |
828 | ||
829 | ||
e9fade72 | 830 | class FFmpegSubtitlesConvertorPP(FFmpegPostProcessor): |
81a23040 | 831 | SUPPORTED_EXTS = ('srt', 'vtt', 'ass', 'lrc') |
832 | ||
e9fade72 JMF |
833 | def __init__(self, downloader=None, format=None): |
834 | super(FFmpegSubtitlesConvertorPP, self).__init__(downloader) | |
835 | self.format = format | |
836 | ||
837 | def run(self, info): | |
838 | subs = info.get('requested_subtitles') | |
e9fade72 JMF |
839 | new_ext = self.format |
840 | new_format = new_ext | |
841 | if new_format == 'vtt': | |
842 | new_format = 'webvtt' | |
843 | if subs is None: | |
1b77b347 | 844 | self.to_screen('There aren\'t any subtitles to convert') |
592e97e8 | 845 | return [], info |
1b77b347 | 846 | self.to_screen('Converting subtitles') |
e04398e3 | 847 | sub_filenames = [] |
e9fade72 | 848 | for lang, sub in subs.items(): |
a1c39673 | 849 | if not os.path.exists(sub.get('filepath', '')): |
850 | self.report_warning(f'Skipping embedding {lang} subtitle because the file is missing') | |
851 | continue | |
e9fade72 JMF |
852 | ext = sub['ext'] |
853 | if ext == new_ext: | |
1b77b347 | 854 | self.to_screen('Subtitle file for %s is already in the requested format' % new_ext) |
e9fade72 | 855 | continue |
503d4a44 | 856 | elif ext == 'json': |
1b77b347 | 857 | self.to_screen( |
858 | 'You have requested to convert json subtitles into another format, ' | |
503d4a44 | 859 | 'which is currently not possible') |
860 | continue | |
dcf64d43 | 861 | old_file = sub['filepath'] |
e04398e3 | 862 | sub_filenames.append(old_file) |
dcf64d43 | 863 | new_file = replace_extension(old_file, new_ext) |
bf6427d2 | 864 | |
40fcba5e | 865 | if ext in ('dfxp', 'ttml', 'tt'): |
f446cc66 | 866 | self.report_warning( |
1b77b347 | 867 | 'You have requested to convert dfxp (TTML) subtitles into another format, ' |
bf6427d2 YCH |
868 | 'which results in style information loss') |
869 | ||
e04398e3 | 870 | dfxp_file = old_file |
dcf64d43 | 871 | srt_file = replace_extension(old_file, 'srt') |
bf6427d2 | 872 | |
3869028f | 873 | with open(dfxp_file, 'rb') as f: |
bf6427d2 YCH |
874 | srt_data = dfxp2srt(f.read()) |
875 | ||
876 | with io.open(srt_file, 'wt', encoding='utf-8') as f: | |
877 | f.write(srt_data) | |
7e62c2eb | 878 | old_file = srt_file |
bf6427d2 | 879 | |
bf6427d2 YCH |
880 | subs[lang] = { |
881 | 'ext': 'srt', | |
dcf64d43 | 882 | 'data': srt_data, |
883 | 'filepath': srt_file, | |
bf6427d2 YCH |
884 | } |
885 | ||
886 | if new_ext == 'srt': | |
887 | continue | |
7b8b007c JMF |
888 | else: |
889 | sub_filenames.append(srt_file) | |
bf6427d2 | 890 | |
e04398e3 | 891 | self.run_ffmpeg(old_file, new_file, ['-f', new_format]) |
e9fade72 JMF |
892 | |
893 | with io.open(new_file, 'rt', encoding='utf-8') as f: | |
894 | subs[lang] = { | |
3547d265 | 895 | 'ext': new_ext, |
e9fade72 | 896 | 'data': f.read(), |
dcf64d43 | 897 | 'filepath': new_file, |
e9fade72 JMF |
898 | } |
899 | ||
dcf64d43 | 900 | info['__files_to_move'][new_file] = replace_extension( |
37242e56 | 901 | info['__files_to_move'][sub['filepath']], new_ext) |
dcf64d43 | 902 | |
e04398e3 | 903 | return sub_filenames, info |
72755351 | 904 | |
905 | ||
906 | class FFmpegSplitChaptersPP(FFmpegPostProcessor): | |
7a340e0d NA |
907 | def __init__(self, downloader, force_keyframes=False): |
908 | FFmpegPostProcessor.__init__(self, downloader) | |
909 | self._force_keyframes = force_keyframes | |
72755351 | 910 | |
911 | def _prepare_filename(self, number, chapter, info): | |
912 | info = info.copy() | |
913 | info.update({ | |
914 | 'section_number': number, | |
915 | 'section_title': chapter.get('title'), | |
916 | 'section_start': chapter.get('start_time'), | |
917 | 'section_end': chapter.get('end_time'), | |
918 | }) | |
919 | return self._downloader.prepare_filename(info, 'chapter') | |
920 | ||
921 | def _ffmpeg_args_for_chapter(self, number, chapter, info): | |
922 | destination = self._prepare_filename(number, chapter, info) | |
923 | if not self._downloader._ensure_dir_exists(encodeFilename(destination)): | |
924 | return | |
925 | ||
dcf64d43 | 926 | chapter['filepath'] = destination |
72755351 | 927 | self.to_screen('Chapter %03d; Destination: %s' % (number, destination)) |
928 | return ( | |
929 | destination, | |
930 | ['-ss', compat_str(chapter['start_time']), | |
a94bfd6c | 931 | '-t', compat_str(chapter['end_time'] - chapter['start_time'])]) |
72755351 | 932 | |
8326b00a | 933 | @PostProcessor._restrict_to(images=False) |
72755351 | 934 | def run(self, info): |
935 | chapters = info.get('chapters') or [] | |
936 | if not chapters: | |
7a340e0d | 937 | self.to_screen('Chapter information is unavailable') |
72755351 | 938 | return [], info |
939 | ||
7a340e0d NA |
940 | in_file = info['filepath'] |
941 | if self._force_keyframes and len(chapters) > 1: | |
942 | in_file = self.force_keyframes(in_file, (c['start_time'] for c in chapters)) | |
72755351 | 943 | self.to_screen('Splitting video by chapters; %d chapters found' % len(chapters)) |
944 | for idx, chapter in enumerate(chapters): | |
945 | destination, opts = self._ffmpeg_args_for_chapter(idx + 1, chapter, info) | |
7a340e0d NA |
946 | self.real_run_ffmpeg([(in_file, opts)], [(destination, ['-c', 'copy'])]) |
947 | if in_file != info['filepath']: | |
948 | os.remove(in_file) | |
72755351 | 949 | return [], info |
8fa43c73 | 950 | |
951 | ||
952 | class FFmpegThumbnailsConvertorPP(FFmpegPostProcessor): | |
81a23040 | 953 | SUPPORTED_EXTS = ('jpg', 'png') |
954 | ||
8fa43c73 | 955 | def __init__(self, downloader=None, format=None): |
956 | super(FFmpegThumbnailsConvertorPP, self).__init__(downloader) | |
957 | self.format = format | |
958 | ||
959 | @staticmethod | |
960 | def is_webp(path): | |
961 | with open(encodeFilename(path), 'rb') as f: | |
962 | b = f.read(12) | |
963 | return b[0:4] == b'RIFF' and b[8:] == b'WEBP' | |
964 | ||
965 | def fixup_webp(self, info, idx=-1): | |
966 | thumbnail_filename = info['thumbnails'][idx]['filepath'] | |
967 | _, thumbnail_ext = os.path.splitext(thumbnail_filename) | |
968 | if thumbnail_ext: | |
969 | thumbnail_ext = thumbnail_ext[1:].lower() | |
970 | if thumbnail_ext != 'webp' and self.is_webp(thumbnail_filename): | |
971 | self.to_screen('Correcting thumbnail "%s" extension to webp' % thumbnail_filename) | |
972 | webp_filename = replace_extension(thumbnail_filename, 'webp') | |
d75201a8 | 973 | os.replace(thumbnail_filename, webp_filename) |
8fa43c73 | 974 | info['thumbnails'][idx]['filepath'] = webp_filename |
975 | info['__files_to_move'][webp_filename] = replace_extension( | |
976 | info['__files_to_move'].pop(thumbnail_filename), 'webp') | |
977 | ||
81a23040 | 978 | @staticmethod |
979 | def _options(target_ext): | |
980 | if target_ext == 'jpg': | |
981 | return ['-bsf:v', 'mjpeg2jpeg'] | |
982 | return [] | |
983 | ||
984 | def convert_thumbnail(self, thumbnail_filename, target_ext): | |
81a23040 | 985 | thumbnail_conv_filename = replace_extension(thumbnail_filename, target_ext) |
337e0c62 | 986 | |
987 | self.to_screen('Converting thumbnail "%s" to %s' % (thumbnail_filename, target_ext)) | |
988 | self.real_run_ffmpeg( | |
989 | [(thumbnail_filename, ['-f', 'image2', '-pattern_type', 'none'])], | |
990 | [(thumbnail_conv_filename.replace('%', '%%'), self._options(target_ext))]) | |
a927acb1 | 991 | return thumbnail_conv_filename |
8fa43c73 | 992 | |
993 | def run(self, info): | |
8fa43c73 | 994 | files_to_delete = [] |
995 | has_thumbnail = False | |
996 | ||
997 | for idx, thumbnail_dict in enumerate(info['thumbnails']): | |
998 | if 'filepath' not in thumbnail_dict: | |
999 | continue | |
1000 | has_thumbnail = True | |
1001 | self.fixup_webp(info, idx) | |
1002 | original_thumbnail = thumbnail_dict['filepath'] | |
1003 | _, thumbnail_ext = os.path.splitext(original_thumbnail) | |
1004 | if thumbnail_ext: | |
1005 | thumbnail_ext = thumbnail_ext[1:].lower() | |
15a4fd53 | 1006 | if thumbnail_ext == 'jpeg': |
1007 | thumbnail_ext = 'jpg' | |
8fa43c73 | 1008 | if thumbnail_ext == self.format: |
1009 | self.to_screen('Thumbnail "%s" is already in the requested format' % original_thumbnail) | |
1010 | continue | |
1011 | thumbnail_dict['filepath'] = self.convert_thumbnail(original_thumbnail, self.format) | |
1012 | files_to_delete.append(original_thumbnail) | |
1013 | info['__files_to_move'][thumbnail_dict['filepath']] = replace_extension( | |
1014 | info['__files_to_move'][original_thumbnail], self.format) | |
1015 | ||
1016 | if not has_thumbnail: | |
1017 | self.to_screen('There aren\'t any thumbnails to convert') | |
1018 | return files_to_delete, info |