]>
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 | ||
b11d2101 | 13 | from ..compat import compat_str |
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, | |
d3c93ec2 | 23 | Popen, |
496c1923 PH |
24 | PostProcessingError, |
25 | prepend_extension, | |
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', | |
abad8000 | 44 | 'vtt': 'webvtt', |
21bfcd3d PH |
45 | } |
46 | ACODECS = { | |
47 | 'mp3': 'libmp3lame', | |
48 | 'aac': 'aac', | |
49 | 'flac': 'flac', | |
50 | 'm4a': 'aac', | |
d2ae7e24 | 51 | 'opus': 'libopus', |
21bfcd3d PH |
52 | 'vorbis': 'libvorbis', |
53 | 'wav': None, | |
a755f825 | 54 | } |
55 | ||
56 | ||
496c1923 PH |
57 | class FFmpegPostProcessorError(PostProcessingError): |
58 | pass | |
59 | ||
d799b47b | 60 | |
496c1923 | 61 | class FFmpegPostProcessor(PostProcessor): |
d47aeb22 | 62 | def __init__(self, downloader=None): |
496c1923 | 63 | PostProcessor.__init__(self, downloader) |
73fac4e9 | 64 | self._determine_executables() |
496c1923 | 65 | |
48844745 | 66 | def check_version(self): |
f740fae2 | 67 | if not self.available: |
beb4b92a | 68 | raise FFmpegPostProcessorError('ffmpeg not found. Please install or provide the path using --ffmpeg-location') |
48844745 | 69 | |
65bf37ef | 70 | required_version = '10-0' if self.basename == 'avconv' else '1.0' |
48844745 | 71 | if is_outdated_version( |
73fac4e9 | 72 | self._versions[self.basename], required_version): |
3aa578ca | 73 | warning = 'Your copy of %s is outdated, update %s to version %s or newer if you encounter any errors.' % ( |
73fac4e9 | 74 | self.basename, self.basename, required_version) |
f446cc66 | 75 | self.report_warning(warning) |
48844745 | 76 | |
496c1923 | 77 | @staticmethod |
73fac4e9 PH |
78 | def get_versions(downloader=None): |
79 | return FFmpegPostProcessor(downloader)._versions | |
6271f1ca | 80 | |
73fac4e9 PH |
81 | def _determine_executables(self): |
82 | programs = ['avprobe', 'avconv', 'ffmpeg', 'ffprobe'] | |
d4a24f40 | 83 | prefer_ffmpeg = True |
73fac4e9 | 84 | |
a64646e4 RA |
85 | def get_ffmpeg_version(path): |
86 | ver = get_exe_version(path, args=['-version']) | |
87 | if ver: | |
88 | regexs = [ | |
cbdc688c | 89 | r'(?:\d+:)?([0-9.]+)-[0-9]+ubuntu[0-9.]+$', # Ubuntu, see [1] |
5caa531a | 90 | r'n([0-9.]+)$', # Arch Linux |
cbdc688c | 91 | # 1. http://www.ducea.com/2006/06/17/ubuntu-package-version-naming-explanation/ |
a64646e4 RA |
92 | ] |
93 | for regex in regexs: | |
94 | mobj = re.match(regex, ver) | |
95 | if mobj: | |
96 | ver = mobj.group(1) | |
97 | return ver | |
98 | ||
73fac4e9 PH |
99 | self.basename = None |
100 | self.probe_basename = None | |
101 | ||
102 | self._paths = None | |
103 | self._versions = None | |
104 | if self._downloader: | |
f446cc66 | 105 | prefer_ffmpeg = self.get_param('prefer_ffmpeg', True) |
106 | location = self.get_param('ffmpeg_location') | |
73fac4e9 PH |
107 | if location is not None: |
108 | if not os.path.exists(location): | |
f446cc66 | 109 | self.report_warning( |
73fac4e9 | 110 | 'ffmpeg-location %s does not exist! ' |
e4172ac9 | 111 | 'Continuing without ffmpeg.' % (location)) |
73fac4e9 PH |
112 | self._versions = {} |
113 | return | |
8c0ae192 | 114 | elif os.path.isdir(location): |
115 | dirname, basename = location, None | |
116 | else: | |
73fac4e9 | 117 | basename = os.path.splitext(os.path.basename(location))[0] |
89efdc15 | 118 | basename = next((p for p in programs if basename.startswith(p)), 'ffmpeg') |
119 | dirname = os.path.dirname(os.path.abspath(location)) | |
73fac4e9 PH |
120 | if basename in ('ffmpeg', 'ffprobe'): |
121 | prefer_ffmpeg = True | |
122 | ||
123 | self._paths = dict( | |
89efdc15 | 124 | (p, os.path.join(dirname, p)) for p in programs) |
8c0ae192 | 125 | if basename: |
126 | self._paths[basename] = location | |
73fac4e9 | 127 | self._versions = dict( |
a64646e4 | 128 | (p, get_ffmpeg_version(self._paths[p])) for p in programs) |
73fac4e9 PH |
129 | if self._versions is None: |
130 | self._versions = dict( | |
a64646e4 | 131 | (p, get_ffmpeg_version(p)) for p in programs) |
73fac4e9 PH |
132 | self._paths = dict((p, p) for p in programs) |
133 | ||
d4a24f40 | 134 | if prefer_ffmpeg is False: |
d28b5171 | 135 | prefs = ('avconv', 'ffmpeg') |
d4a24f40 S |
136 | else: |
137 | prefs = ('ffmpeg', 'avconv') | |
d28b5171 PH |
138 | for p in prefs: |
139 | if self._versions[p]: | |
73fac4e9 PH |
140 | self.basename = p |
141 | break | |
76b1bd67 | 142 | |
d4a24f40 | 143 | if prefer_ffmpeg is False: |
1a253e13 | 144 | prefs = ('avprobe', 'ffprobe') |
d4a24f40 S |
145 | else: |
146 | prefs = ('ffprobe', 'avprobe') | |
1a253e13 PH |
147 | for p in prefs: |
148 | if self._versions[p]: | |
73fac4e9 PH |
149 | self.probe_basename = p |
150 | break | |
151 | ||
f740fae2 | 152 | @property |
73fac4e9 PH |
153 | def available(self): |
154 | return self.basename is not None | |
1a253e13 | 155 | |
73fac4e9 PH |
156 | @property |
157 | def executable(self): | |
158 | return self._paths[self.basename] | |
159 | ||
3da4b313 JMF |
160 | @property |
161 | def probe_available(self): | |
162 | return self.probe_basename is not None | |
163 | ||
73fac4e9 PH |
164 | @property |
165 | def probe_executable(self): | |
166 | return self._paths[self.probe_basename] | |
76b1bd67 | 167 | |
30d9e209 | 168 | def get_audio_codec(self, path): |
eb35b163 | 169 | if not self.probe_available and not self.available: |
beb4b92a | 170 | raise PostProcessingError('ffprobe and ffmpeg not found. Please install or provide the path using --ffmpeg-location') |
30d9e209 | 171 | try: |
eb35b163 RA |
172 | if self.probe_available: |
173 | cmd = [ | |
174 | encodeFilename(self.probe_executable, True), | |
175 | encodeArgument('-show_streams')] | |
176 | else: | |
177 | cmd = [ | |
178 | encodeFilename(self.executable, True), | |
179 | encodeArgument('-i')] | |
180 | cmd.append(encodeFilename(self._ffmpeg_filename_argument(path), True)) | |
f446cc66 | 181 | self.write_debug('%s command line: %s' % (self.basename, shell_quote(cmd))) |
d3c93ec2 | 182 | handle = Popen(cmd, stdin=subprocess.PIPE, stdout=subprocess.PIPE, stderr=subprocess.PIPE) |
183 | stdout_data, stderr_data = handle.communicate_or_kill() | |
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)) |
d3c93ec2 | 225 | p = Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, stdin=subprocess.PIPE) |
06167fbb | 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 | ||
165efb82 | 236 | def _get_real_video_duration(self, info, fatal=True): |
237 | try: | |
238 | if '_real_duration' not in info: | |
239 | info['_real_duration'] = float_or_none( | |
240 | traverse_obj(self.get_metadata_object(info['filepath']), ('format', 'duration'))) | |
241 | if not info['_real_duration']: | |
242 | raise PostProcessingError('ffprobe returned empty duration') | |
243 | except PostProcessingError as e: | |
244 | if fatal: | |
245 | raise PostProcessingError(f'Unable to determine video duration; {e}') | |
246 | return info.setdefault('_real_duration', None) | |
247 | ||
248 | def _duration_mismatch(self, d1, d2): | |
249 | if not d1 or not d2: | |
250 | return None | |
251 | return abs(d1 - d2) > 1 | |
252 | ||
00034c14 | 253 | def run_ffmpeg_multiple_files(self, input_paths, out_path, opts, **kwargs): |
e92caff5 | 254 | return self.real_run_ffmpeg( |
255 | [(path, []) for path in input_paths], | |
00034c14 | 256 | [(out_path, opts)], **kwargs) |
e92caff5 | 257 | |
00034c14 | 258 | def real_run_ffmpeg(self, input_path_opts, output_path_opts, *, expected_retcodes=(0,)): |
48844745 | 259 | self.check_version() |
496c1923 | 260 | |
52afb2ac | 261 | oldest_mtime = min( |
7dde84f3 | 262 | os.stat(encodeFilename(path)).st_mtime for path, _ in input_path_opts if path) |
43bc8890 | 263 | |
91b6c884 | 264 | cmd = [encodeFilename(self.executable, True), encodeArgument('-y')] |
ce52c7c1 S |
265 | # avconv does not have repeat option |
266 | if self.basename == 'ffmpeg': | |
267 | cmd += [encodeArgument('-loglevel'), encodeArgument('repeat+info')] | |
5b1ecbb3 | 268 | |
e92caff5 | 269 | def make_args(file, args, name, number): |
270 | keys = ['_%s%d' % (name, number), '_%s' % name] | |
271 | if name == 'o' and number == 1: | |
272 | keys.append('') | |
273 | args += self._configuration_args(self.basename, keys) | |
274 | if name == 'i': | |
275 | args.append('-i') | |
5b1ecbb3 | 276 | return ( |
e92caff5 | 277 | [encodeArgument(arg) for arg in args] |
5b1ecbb3 | 278 | + [encodeFilename(self._ffmpeg_filename_argument(file), True)]) |
279 | ||
e92caff5 | 280 | for arg_type, path_opts in (('i', input_path_opts), ('o', output_path_opts)): |
7dde84f3 | 281 | cmd += itertools.chain.from_iterable( |
282 | make_args(path, list(opts), arg_type, i + 1) | |
283 | for i, (path, opts) in enumerate(path_opts) if path) | |
496c1923 | 284 | |
f446cc66 | 285 | self.write_debug('ffmpeg command line: %s' % shell_quote(cmd)) |
d3c93ec2 | 286 | p = Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, stdin=subprocess.PIPE) |
287 | stdout, stderr = p.communicate_or_kill() | |
00034c14 | 288 | if p.returncode not in variadic(expected_retcodes): |
06167fbb | 289 | stderr = stderr.decode('utf-8', 'replace').strip() |
b1940459 | 290 | self.write_debug(stderr) |
06167fbb | 291 | raise FFmpegPostProcessorError(stderr.split('\n')[-1]) |
e92caff5 | 292 | for out_path, _ in output_path_opts: |
7dde84f3 | 293 | if out_path: |
294 | self.try_utime(out_path, oldest_mtime, oldest_mtime) | |
06167fbb | 295 | return stderr.decode('utf-8', 'replace') |
cc55d088 | 296 | |
00034c14 | 297 | def run_ffmpeg(self, path, out_path, opts, **kwargs): |
298 | return self.run_ffmpeg_multiple_files([path], out_path, opts, **kwargs) | |
496c1923 | 299 | |
7a340e0d NA |
300 | @staticmethod |
301 | def _ffmpeg_filename_argument(fn): | |
8a7bbd16 JMF |
302 | # Always use 'file:' because the filename may contain ':' (ffmpeg |
303 | # interprets that as a protocol) or can start with '-' (-- is broken in | |
304 | # ffmpeg, see https://ffmpeg.org/trac/ffmpeg/ticket/2127 for details) | |
b9f2fdd3 | 305 | # Also leave '-' intact in order not to break streaming to stdout. |
06167fbb | 306 | if fn.startswith(('http://', 'https://')): |
307 | return fn | |
d868f43c | 308 | return 'file:' + fn if fn != '-' else fn |
496c1923 | 309 | |
7a340e0d NA |
310 | @staticmethod |
311 | def _quote_for_ffmpeg(string): | |
312 | # See https://ffmpeg.org/ffmpeg-utils.html#toc-Quoting-and-escaping | |
313 | # A sequence of '' produces '\'''\''; | |
314 | # final replace removes the empty '' between \' \'. | |
315 | string = string.replace("'", r"'\''").replace("'''", "'") | |
316 | # Handle potential ' at string boundaries. | |
317 | string = string[1:] if string[0] == "'" else "'" + string | |
318 | return string[:-1] if string[-1] == "'" else string + "'" | |
319 | ||
320 | def force_keyframes(self, filename, timestamps): | |
321 | timestamps = orderedSet(timestamps) | |
322 | if timestamps[0] == 0: | |
323 | timestamps = timestamps[1:] | |
324 | keyframe_file = prepend_extension(filename, 'keyframes.temp') | |
325 | self.to_screen(f'Re-encoding "{filename}" with appropriate keyframes') | |
326 | self.run_ffmpeg(filename, keyframe_file, ['-force_key_frames', ','.join( | |
327 | f'{t:.6f}' for t in timestamps)]) | |
328 | return keyframe_file | |
329 | ||
330 | def concat_files(self, in_files, out_file, concat_opts=None): | |
331 | """ | |
332 | Use concat demuxer to concatenate multiple files having identical streams. | |
333 | ||
334 | Only inpoint, outpoint, and duration concat options are supported. | |
335 | See https://ffmpeg.org/ffmpeg-formats.html#concat-1 for details | |
336 | """ | |
337 | concat_file = f'{out_file}.concat' | |
338 | self.write_debug(f'Writing concat spec to {concat_file}') | |
339 | with open(concat_file, 'wt', encoding='utf-8') as f: | |
340 | f.writelines(self._concat_spec(in_files, concat_opts)) | |
341 | ||
342 | out_flags = ['-c', 'copy'] | |
343 | if out_file.rpartition('.')[-1] in ('mp4', 'mov'): | |
344 | # For some reason, '-c copy' is not enough to copy subtitles | |
345 | out_flags.extend(['-c:s', 'mov_text', '-movflags', '+faststart']) | |
346 | ||
347 | try: | |
348 | self.real_run_ffmpeg( | |
349 | [(concat_file, ['-hide_banner', '-nostdin', '-f', 'concat', '-safe', '0'])], | |
350 | [(out_file, out_flags)]) | |
351 | finally: | |
352 | os.remove(concat_file) | |
353 | ||
354 | @classmethod | |
355 | def _concat_spec(cls, in_files, concat_opts=None): | |
356 | if concat_opts is None: | |
357 | concat_opts = [{}] * len(in_files) | |
358 | yield 'ffconcat version 1.0\n' | |
359 | for file, opts in zip(in_files, concat_opts): | |
360 | yield f'file {cls._quote_for_ffmpeg(cls._ffmpeg_filename_argument(file))}\n' | |
361 | # Iterate explicitly to yield the following directives in order, ignoring the rest. | |
362 | for directive in 'inpoint', 'outpoint', 'duration': | |
363 | if directive in opts: | |
364 | yield f'{directive} {opts[directive]}\n' | |
365 | ||
496c1923 PH |
366 | |
367 | class FFmpegExtractAudioPP(FFmpegPostProcessor): | |
81a23040 | 368 | COMMON_AUDIO_EXTS = ('wav', 'flac', 'm4a', 'aiff', 'mp3', 'ogg', 'mka', 'opus', 'wma') |
369 | SUPPORTED_EXTS = ('best', 'aac', 'flac', 'mp3', 'm4a', 'opus', 'vorbis', 'wav') | |
1de75fa1 | 370 | |
496c1923 PH |
371 | def __init__(self, downloader=None, preferredcodec=None, preferredquality=None, nopostoverwrites=False): |
372 | FFmpegPostProcessor.__init__(self, downloader) | |
81a23040 | 373 | self._preferredcodec = preferredcodec or 'best' |
31c49255 | 374 | self._preferredquality = float_or_none(preferredquality) |
496c1923 PH |
375 | self._nopostoverwrites = nopostoverwrites |
376 | ||
31c49255 | 377 | def _quality_args(self, codec): |
378 | if self._preferredquality is None: | |
379 | return [] | |
380 | elif self._preferredquality > 10: | |
381 | return ['-b:a', f'{self._preferredquality}k'] | |
382 | ||
383 | limits = { | |
384 | 'libmp3lame': (10, 0), | |
385 | 'aac': (0.1, 11), | |
386 | 'vorbis': (0, 10), | |
387 | 'opus': None, # doesn't support -q:a | |
388 | 'wav': None, | |
389 | 'flac': None, | |
390 | }[codec] | |
391 | if not limits: | |
392 | return [] | |
393 | ||
394 | q = limits[1] + (limits[0] - limits[1]) * (self._preferredquality / 10) | |
395 | return ['-q:a', f'{q}'] | |
396 | ||
496c1923 | 397 | def run_ffmpeg(self, path, out_path, codec, more_opts): |
496c1923 PH |
398 | if codec is None: |
399 | acodec_opts = [] | |
400 | else: | |
401 | acodec_opts = ['-acodec', codec] | |
402 | opts = ['-vn'] + acodec_opts + more_opts | |
403 | try: | |
404 | FFmpegPostProcessor.run_ffmpeg(self, path, out_path, opts) | |
405 | except FFmpegPostProcessorError as err: | |
406 | raise AudioConversionError(err.msg) | |
407 | ||
8326b00a | 408 | @PostProcessor._restrict_to(images=False) |
496c1923 PH |
409 | def run(self, information): |
410 | path = information['filepath'] | |
1de75fa1 | 411 | orig_ext = information['ext'] |
412 | ||
81a23040 | 413 | if self._preferredcodec == 'best' and orig_ext in self.COMMON_AUDIO_EXTS: |
1de75fa1 | 414 | self.to_screen('Skipping audio extraction since the file is already in a common audio format') |
55b53b33 | 415 | return [], information |
496c1923 PH |
416 | |
417 | filecodec = self.get_audio_codec(path) | |
418 | if filecodec is None: | |
3aa578ca | 419 | raise PostProcessingError('WARNING: unable to obtain file audio codec with ffprobe') |
496c1923 PH |
420 | |
421 | more_opts = [] | |
422 | if self._preferredcodec == 'best' or self._preferredcodec == filecodec or (self._preferredcodec == 'm4a' and filecodec == 'aac'): | |
423 | if filecodec == 'aac' and self._preferredcodec in ['m4a', 'best']: | |
424 | # Lossless, but in another container | |
425 | acodec = 'copy' | |
426 | extension = 'm4a' | |
467d3c9a | 427 | more_opts = ['-bsf:a', 'aac_adtstoasc'] |
21bfcd3d | 428 | elif filecodec in ['aac', 'flac', 'mp3', 'vorbis', 'opus']: |
496c1923 PH |
429 | # Lossless if possible |
430 | acodec = 'copy' | |
431 | extension = filecodec | |
432 | if filecodec == 'aac': | |
433 | more_opts = ['-f', 'adts'] | |
434 | if filecodec == 'vorbis': | |
435 | extension = 'ogg' | |
436 | else: | |
437 | # MP3 otherwise. | |
438 | acodec = 'libmp3lame' | |
439 | extension = 'mp3' | |
31c49255 | 440 | more_opts = self._quality_args(acodec) |
496c1923 | 441 | else: |
21bfcd3d PH |
442 | # We convert the audio (lossy if codec is lossy) |
443 | acodec = ACODECS[self._preferredcodec] | |
496c1923 | 444 | extension = self._preferredcodec |
31c49255 | 445 | more_opts = self._quality_args(acodec) |
496c1923 PH |
446 | if self._preferredcodec == 'aac': |
447 | more_opts += ['-f', 'adts'] | |
448 | if self._preferredcodec == 'm4a': | |
467d3c9a | 449 | more_opts += ['-bsf:a', 'aac_adtstoasc'] |
496c1923 PH |
450 | if self._preferredcodec == 'vorbis': |
451 | extension = 'ogg' | |
452 | if self._preferredcodec == 'wav': | |
453 | extension = 'wav' | |
454 | more_opts += ['-f', 'wav'] | |
455 | ||
3aa578ca | 456 | prefix, sep, ext = path.rpartition('.') # not os.path.splitext, since the latter does not work on unicode in all setups |
496c1923 | 457 | new_path = prefix + sep + extension |
0b94dbb1 | 458 | |
2273e2c5 PM |
459 | information['filepath'] = new_path |
460 | information['ext'] = extension | |
496c1923 PH |
461 | |
462 | # If we download foo.mp3 and convert it to... foo.mp3, then don't delete foo.mp3, silly. | |
3089bc74 S |
463 | if (new_path == path |
464 | or (self._nopostoverwrites and os.path.exists(encodeFilename(new_path)))): | |
1b77b347 | 465 | self.to_screen('Post-process file %s exists, skipping' % new_path) |
592e97e8 | 466 | return [], information |
496c1923 PH |
467 | |
468 | try: | |
1b77b347 | 469 | self.to_screen('Destination: ' + new_path) |
ce81b141 | 470 | self.run_ffmpeg(path, new_path, acodec, more_opts) |
70a1165b JMF |
471 | except AudioConversionError as e: |
472 | raise PostProcessingError( | |
473 | 'audio conversion failed: ' + e.msg) | |
474 | except Exception: | |
475 | raise PostProcessingError('error running ' + self.basename) | |
496c1923 PH |
476 | |
477 | # Try to update the date time for extracted audio file. | |
478 | if information.get('filetime') is not None: | |
dd29eb7f S |
479 | self.try_utime( |
480 | new_path, time.time(), information['filetime'], | |
481 | errnote='Cannot update utime of audio file') | |
496c1923 | 482 | |
592e97e8 | 483 | return [path], information |
496c1923 PH |
484 | |
485 | ||
857f6313 | 486 | class FFmpegVideoConvertorPP(FFmpegPostProcessor): |
81a23040 | 487 | SUPPORTED_EXTS = ('mp4', 'mkv', 'flv', 'webm', 'mov', 'avi', 'mp3', 'mka', 'm4a', 'ogg', 'opus') |
488 | FORMAT_RE = re.compile(r'{0}(?:/{0})*$'.format(r'(?:\w+>)?(?:%s)' % '|'.join(SUPPORTED_EXTS))) | |
e6f21b3d | 489 | _ACTION = 'converting' |
857f6313 | 490 | |
efe87a10 | 491 | def __init__(self, downloader=None, preferedformat=None): |
857f6313 | 492 | super(FFmpegVideoConvertorPP, self).__init__(downloader) |
06167fbb | 493 | self._preferedformats = preferedformat.lower().split('/') |
efe87a10 | 494 | |
857f6313 | 495 | def _target_ext(self, source_ext): |
06167fbb | 496 | for pair in self._preferedformats: |
497 | kv = pair.split('>') | |
857f6313 | 498 | if len(kv) == 1 or kv[0].strip() == source_ext: |
499 | return kv[-1].strip() | |
06167fbb | 500 | |
857f6313 | 501 | @staticmethod |
502 | def _options(target_ext): | |
503 | if target_ext == 'avi': | |
504 | return ['-c:v', 'libxvid', '-vtag', 'XVID'] | |
505 | return [] | |
506 | ||
8326b00a | 507 | @PostProcessor._restrict_to(images=False) |
e6f21b3d | 508 | def run(self, info): |
509 | filename, source_ext = info['filepath'], info['ext'].lower() | |
81a23040 | 510 | target_ext = self._target_ext(source_ext) |
06167fbb | 511 | _skip_msg = ( |
e6f21b3d | 512 | f'could not find a mapping for {source_ext}' if not target_ext |
513 | else f'already is in target format {source_ext}' if source_ext == target_ext | |
06167fbb | 514 | else None) |
515 | if _skip_msg: | |
e6f21b3d | 516 | self.to_screen(f'Not {self._ACTION} media file {filename!r}; {_skip_msg}') |
517 | return [], info | |
06167fbb | 518 | |
e6f21b3d | 519 | outpath = replace_extension(filename, target_ext, source_ext) |
520 | self.to_screen(f'{self._ACTION.title()} video from {source_ext} to {target_ext}; Destination: {outpath}') | |
521 | self.run_ffmpeg(filename, outpath, self._options(target_ext)) | |
857f6313 | 522 | |
e6f21b3d | 523 | info['filepath'] = outpath |
524 | info['format'] = info['ext'] = target_ext | |
525 | return [filename], info | |
efe87a10 FS |
526 | |
527 | ||
857f6313 | 528 | class FFmpegVideoRemuxerPP(FFmpegVideoConvertorPP): |
e6f21b3d | 529 | _ACTION = 'remuxing' |
496c1923 | 530 | |
857f6313 | 531 | @staticmethod |
532 | def _options(target_ext): | |
533 | options = ['-c', 'copy', '-map', '0', '-dn'] | |
534 | if target_ext in ['mp4', 'm4a', 'mov']: | |
535 | options.extend(['-movflags', '+faststart']) | |
536 | return options | |
496c1923 PH |
537 | |
538 | ||
539 | class FFmpegEmbedSubtitlePP(FFmpegPostProcessor): | |
cffab0ee | 540 | def __init__(self, downloader=None, already_have_subtitle=False): |
541 | super(FFmpegEmbedSubtitlePP, self).__init__(downloader) | |
542 | self._already_have_subtitle = already_have_subtitle | |
543 | ||
8326b00a | 544 | @PostProcessor._restrict_to(images=False) |
496c1923 | 545 | def run(self, information): |
40025ee2 | 546 | if information['ext'] not in ('mp4', 'webm', 'mkv'): |
1b77b347 | 547 | self.to_screen('Subtitles can only be embedded in mp4, webm or mkv files') |
592e97e8 | 548 | return [], information |
c84dd8a9 JMF |
549 | subtitles = information.get('requested_subtitles') |
550 | if not subtitles: | |
1b77b347 | 551 | self.to_screen('There aren\'t any subtitles to embed') |
592e97e8 | 552 | return [], information |
496c1923 | 553 | |
496c1923 | 554 | filename = information['filepath'] |
1f2a268b | 555 | if information.get('duration') and self._duration_mismatch( |
165efb82 | 556 | self._get_real_video_duration(information, False), information['duration']): |
557 | self.to_screen(f'Skipping {self.pp_key()} since the real and expected durations mismatch') | |
558 | return [], information | |
40025ee2 S |
559 | |
560 | ext = information['ext'] | |
2412044c | 561 | sub_langs, sub_names, sub_filenames = [], [], [] |
40025ee2 | 562 | webm_vtt_warn = False |
06167fbb | 563 | mp4_ass_warn = False |
40025ee2 S |
564 | |
565 | for lang, sub_info in subtitles.items(): | |
a1c39673 | 566 | if not os.path.exists(sub_info.get('filepath', '')): |
8e25d624 | 567 | self.report_warning(f'Skipping embedding {lang} subtitle because the file is missing') |
568 | continue | |
40025ee2 | 569 | sub_ext = sub_info['ext'] |
503d4a44 | 570 | if sub_ext == 'json': |
06167fbb | 571 | self.report_warning('JSON subtitles cannot be embedded') |
503d4a44 | 572 | elif ext != 'webm' or ext == 'webm' and sub_ext == 'vtt': |
40025ee2 | 573 | sub_langs.append(lang) |
2412044c | 574 | sub_names.append(sub_info.get('name')) |
dcf64d43 | 575 | sub_filenames.append(sub_info['filepath']) |
40025ee2 S |
576 | else: |
577 | if not webm_vtt_warn and ext == 'webm' and sub_ext != 'vtt': | |
578 | webm_vtt_warn = True | |
06167fbb | 579 | self.report_warning('Only WebVTT subtitles can be embedded in webm files') |
580 | if not mp4_ass_warn and ext == 'mp4' and sub_ext == 'ass': | |
581 | mp4_ass_warn = True | |
582 | self.report_warning('ASS subtitles cannot be properly embedded in mp4 files; expect issues') | |
40025ee2 S |
583 | |
584 | if not sub_langs: | |
585 | return [], information | |
586 | ||
14523ed9 | 587 | input_files = [filename] + sub_filenames |
496c1923 | 588 | |
e205db3b | 589 | opts = [ |
e0da59fe | 590 | '-c', 'copy', '-map', '0', '-dn', |
e205db3b JMF |
591 | # Don't copy the existing subtitles, we may be running the |
592 | # postprocessor a second time | |
593 | '-map', '-0:s', | |
7f903dd8 T |
594 | # Don't copy Apple TV chapters track, bin_data (see #19042, #19024, |
595 | # https://trac.ffmpeg.org/ticket/6016) | |
596 | '-map', '-0:d', | |
e205db3b | 597 | ] |
083c1bb9 N |
598 | if information['ext'] == 'mp4': |
599 | opts += ['-c:s', 'mov_text'] | |
2412044c | 600 | for i, (lang, name) in enumerate(zip(sub_langs, sub_names)): |
2875cf01 | 601 | opts.extend(['-map', '%d:0' % (i + 1)]) |
04fb6928 S |
602 | lang_code = ISO639Utils.short2long(lang) or lang |
603 | opts.extend(['-metadata:s:s:%d' % i, 'language=%s' % lang_code]) | |
2412044c | 604 | if name: |
605 | opts.extend(['-metadata:s:s:%d' % i, 'handler_name=%s' % name, | |
606 | '-metadata:s:s:%d' % i, 'title=%s' % name]) | |
496c1923 | 607 | |
2875cf01 | 608 | temp_filename = prepend_extension(filename, 'temp') |
06167fbb | 609 | self.to_screen('Embedding subtitles in "%s"' % filename) |
496c1923 | 610 | self.run_ffmpeg_multiple_files(input_files, temp_filename, opts) |
d75201a8 | 611 | os.replace(temp_filename, filename) |
496c1923 | 612 | |
cffab0ee | 613 | files_to_delete = [] if self._already_have_subtitle else sub_filenames |
614 | return files_to_delete, information | |
496c1923 PH |
615 | |
616 | ||
617 | class FFmpegMetadataPP(FFmpegPostProcessor): | |
7dde84f3 | 618 | |
7a340e0d NA |
619 | def __init__(self, downloader, add_metadata=True, add_chapters=True): |
620 | FFmpegPostProcessor.__init__(self, downloader) | |
621 | self._add_metadata = add_metadata | |
622 | self._add_chapters = add_chapters | |
623 | ||
7dde84f3 | 624 | @staticmethod |
625 | def _options(target_ext): | |
626 | yield from ('-map', '0', '-dn') | |
627 | if target_ext == 'm4a': | |
628 | yield from ('-vn', '-acodec', 'copy') | |
629 | else: | |
630 | yield from ('-c', 'copy') | |
631 | ||
8326b00a | 632 | @PostProcessor._restrict_to(images=False) |
496c1923 | 633 | def run(self, info): |
7a340e0d NA |
634 | filename, metadata_filename = info['filepath'], None |
635 | options = [] | |
636 | if self._add_chapters and info.get('chapters'): | |
637 | metadata_filename = replace_extension(filename, 'meta') | |
638 | options.extend(self._get_chapter_opts(info['chapters'], metadata_filename)) | |
639 | if self._add_metadata: | |
640 | options.extend(self._get_metadata_opts(info)) | |
641 | ||
642 | if not options: | |
643 | self.to_screen('There isn\'t any metadata to add') | |
644 | return [], info | |
645 | ||
646 | temp_filename = prepend_extension(filename, 'temp') | |
647 | self.to_screen('Adding metadata to "%s"' % filename) | |
648 | self.run_ffmpeg_multiple_files( | |
649 | (filename, metadata_filename), temp_filename, | |
650 | itertools.chain(self._options(info['ext']), *options)) | |
651 | if metadata_filename: | |
652 | os.remove(metadata_filename) | |
653 | os.replace(temp_filename, filename) | |
654 | return [], info | |
655 | ||
656 | @staticmethod | |
657 | def _get_chapter_opts(chapters, metadata_filename): | |
658 | with io.open(metadata_filename, 'wt', encoding='utf-8') as f: | |
659 | def ffmpeg_escape(text): | |
660 | return re.sub(r'([\\=;#\n])', r'\\\1', text) | |
661 | ||
662 | metadata_file_content = ';FFMETADATA1\n' | |
663 | for chapter in chapters: | |
664 | metadata_file_content += '[CHAPTER]\nTIMEBASE=1/1000\n' | |
665 | metadata_file_content += 'START=%d\n' % (chapter['start_time'] * 1000) | |
666 | metadata_file_content += 'END=%d\n' % (chapter['end_time'] * 1000) | |
667 | chapter_title = chapter.get('title') | |
668 | if chapter_title: | |
669 | metadata_file_content += 'title=%s\n' % ffmpeg_escape(chapter_title) | |
670 | f.write(metadata_file_content) | |
671 | yield ('-map_metadata', '1') | |
672 | ||
673 | def _get_metadata_opts(self, info): | |
496c1923 | 674 | metadata = {} |
b11d2101 | 675 | meta_prefix = 'meta_' |
4bd143a3 S |
676 | |
677 | def add(meta_list, info_list=None): | |
b11d2101 | 678 | value = next(( |
679 | str(info[key]) for key in [meta_prefix] + list(variadic(info_list or meta_list)) | |
680 | if info.get(key) is not None), None) | |
681 | if value not in ('', None): | |
682 | metadata.update({meta_f: value for meta_f in variadic(meta_list)}) | |
4bd143a3 | 683 | |
2791e80b S |
684 | # See [1-4] for some info on media metadata/metadata supported |
685 | # by ffmpeg. | |
686 | # 1. https://kdenlive.org/en/project/adding-meta-data-to-mp4-video/ | |
687 | # 2. https://wiki.multimedia.cx/index.php/FFmpeg_Metadata | |
688 | # 3. https://kodi.wiki/view/Video_file_tagging | |
2791e80b | 689 | |
4bd143a3 S |
690 | add('title', ('track', 'title')) |
691 | add('date', 'upload_date') | |
cd9b384c | 692 | add(('description', 'synopsis'), 'description') |
693 | add(('purl', 'comment'), 'webpage_url') | |
4bd143a3 S |
694 | add('track', 'track_number') |
695 | add('artist', ('artist', 'creator', 'uploader', 'uploader_id')) | |
696 | add('genre') | |
697 | add('album') | |
698 | add('album_artist') | |
699 | add('disc', 'disc_number') | |
2791e80b S |
700 | add('show', 'series') |
701 | add('season_number') | |
702 | add('episode_id', ('episode', 'episode_id')) | |
703 | add('episode_sort', 'episode_number') | |
496c1923 | 704 | |
b11d2101 | 705 | for key, value in info.items(): |
706 | if value is not None and key != meta_prefix and key.startswith(meta_prefix): | |
707 | metadata[key[len(meta_prefix):]] = value | |
84601bb7 | 708 | |
7a340e0d NA |
709 | for name, value in metadata.items(): |
710 | yield ('-metadata', f'{name}={value}') | |
39c68260 | 711 | |
7dde84f3 | 712 | stream_idx = 0 |
713 | for fmt in info.get('requested_formats') or []: | |
714 | stream_count = 2 if 'none' not in (fmt.get('vcodec'), fmt.get('acodec')) else 1 | |
715 | if fmt.get('language'): | |
716 | lang = ISO639Utils.short2long(fmt['language']) or fmt['language'] | |
7a340e0d NA |
717 | for i in range(stream_count): |
718 | yield ('-metadata:s:%d' % (stream_idx + i), 'language=%s' % lang) | |
7dde84f3 | 719 | stream_idx += stream_count |
496c1923 | 720 | |
41712218 | 721 | if ('no-attach-info-json' not in self.get_param('compat_opts', []) |
722 | and '__infojson_filename' in info and info['ext'] in ('mkv', 'mka')): | |
7a340e0d | 723 | old_stream, new_stream = self.get_stream_number(info['filepath'], ('tags', 'mimetype'), 'application/json') |
06167fbb | 724 | if old_stream is not None: |
7a340e0d | 725 | yield ('-map', '-0:%d' % old_stream) |
06167fbb | 726 | new_stream -= 1 |
727 | ||
7a340e0d NA |
728 | yield ('-attach', info['__infojson_filename'], |
729 | '-metadata:s:%d' % new_stream, 'mimetype=application/json') | |
496c1923 PH |
730 | |
731 | ||
732 | class FFmpegMergerPP(FFmpegPostProcessor): | |
8326b00a | 733 | @PostProcessor._restrict_to(images=False) |
496c1923 PH |
734 | def run(self, info): |
735 | filename = info['filepath'] | |
5b5fbc08 | 736 | temp_filename = prepend_extension(filename, 'temp') |
d03cfdce | 737 | args = ['-c', 'copy'] |
50eff38c | 738 | audio_streams = 0 |
d03cfdce | 739 | for (i, fmt) in enumerate(info['requested_formats']): |
740 | if fmt.get('acodec') != 'none': | |
a21e0ab1 | 741 | args.extend(['-map', f'{i}:a:0']) |
9dda99f2 | 742 | aac_fixup = fmt['protocol'].startswith('m3u8') and self.get_audio_codec(fmt['filepath']) == 'aac' |
743 | if aac_fixup: | |
50eff38c | 744 | args.extend([f'-bsf:a:{audio_streams}', 'aac_adtstoasc']) |
745 | audio_streams += 1 | |
d03cfdce | 746 | if fmt.get('vcodec') != 'none': |
747 | args.extend(['-map', '%u:v:0' % (i)]) | |
1b77b347 | 748 | self.to_screen('Merging formats into "%s"' % filename) |
5b5fbc08 JMF |
749 | self.run_ffmpeg_multiple_files(info['__files_to_merge'], temp_filename, args) |
750 | os.rename(encodeFilename(temp_filename), encodeFilename(filename)) | |
d47aeb22 | 751 | return info['__files_to_merge'], info |
496c1923 | 752 | |
13763ce5 S |
753 | def can_merge(self): |
754 | # TODO: figure out merge-capable ffmpeg version | |
755 | if self.basename != 'avconv': | |
756 | return True | |
757 | ||
758 | required_version = '10-0' | |
759 | if is_outdated_version( | |
760 | self._versions[self.basename], required_version): | |
761 | warning = ('Your copy of %s is outdated and unable to properly mux separate video and audio files, ' | |
7a5c1cfe | 762 | 'yt-dlp will download single file media. ' |
13763ce5 S |
763 | 'Update %s to version %s or newer to fix this.') % ( |
764 | self.basename, self.basename, required_version) | |
f446cc66 | 765 | self.report_warning(warning) |
13763ce5 S |
766 | return False |
767 | return True | |
768 | ||
0c14e2fb | 769 | |
fd7cfb64 | 770 | class FFmpegFixupPostProcessor(FFmpegPostProcessor): |
771 | def _fixup(self, msg, filename, options): | |
6271f1ca PH |
772 | temp_filename = prepend_extension(filename, 'temp') |
773 | ||
f89b3e2d | 774 | self.to_screen(f'{msg} of "{filename}"') |
6271f1ca PH |
775 | self.run_ffmpeg(filename, temp_filename, options) |
776 | ||
d75201a8 | 777 | os.replace(temp_filename, filename) |
6271f1ca | 778 | |
fd7cfb64 | 779 | |
780 | class FFmpegFixupStretchedPP(FFmpegFixupPostProcessor): | |
781 | @PostProcessor._restrict_to(images=False, audio=False) | |
782 | def run(self, info): | |
783 | stretched_ratio = info.get('stretched_ratio') | |
784 | if stretched_ratio not in (None, 1): | |
785 | self._fixup('Fixing aspect ratio', info['filepath'], [ | |
786 | '-c', 'copy', '-map', '0', '-dn', '-aspect', '%f' % stretched_ratio]) | |
592e97e8 | 787 | return [], info |
62cd676c PH |
788 | |
789 | ||
fd7cfb64 | 790 | class FFmpegFixupM4aPP(FFmpegFixupPostProcessor): |
8326b00a | 791 | @PostProcessor._restrict_to(images=False, video=False) |
62cd676c | 792 | def run(self, info): |
fd7cfb64 | 793 | if info.get('container') == 'm4a_dash': |
794 | self._fixup('Correcting container', info['filepath'], [ | |
795 | '-c', 'copy', '-map', '0', '-dn', '-f', 'mp4']) | |
592e97e8 | 796 | return [], info |
e9fade72 JMF |
797 | |
798 | ||
fd7cfb64 | 799 | class FFmpegFixupM3u8PP(FFmpegFixupPostProcessor): |
8326b00a | 800 | @PostProcessor._restrict_to(images=False) |
f17f8651 | 801 | def run(self, info): |
fd7cfb64 | 802 | if self.get_audio_codec(info['filepath']) == 'aac': |
803 | self._fixup('Fixing malformed AAC bitstream', info['filepath'], [ | |
804 | '-c', 'copy', '-map', '0', '-dn', '-f', 'mp4', '-bsf:a', 'aac_adtstoasc']) | |
f17f8651 | 805 | return [], info |
806 | ||
807 | ||
e36d50c5 | 808 | class FFmpegFixupTimestampPP(FFmpegFixupPostProcessor): |
809 | ||
810 | def __init__(self, downloader=None, trim=0.001): | |
811 | # "trim" should be used when the video contains unintended packets | |
812 | super(FFmpegFixupTimestampPP, self).__init__(downloader) | |
813 | assert isinstance(trim, (int, float)) | |
814 | self.trim = str(trim) | |
815 | ||
816 | @PostProcessor._restrict_to(images=False) | |
817 | def run(self, info): | |
818 | required_version = '4.4' | |
819 | if is_outdated_version(self._versions[self.basename], required_version): | |
820 | self.report_warning( | |
821 | 'A re-encode is needed to fix timestamps in older versions of ffmpeg. ' | |
822 | f'Please install ffmpeg {required_version} or later to fixup without re-encoding') | |
823 | opts = ['-vf', 'setpts=PTS-STARTPTS'] | |
824 | else: | |
825 | opts = ['-c', 'copy', '-bsf', 'setts=ts=TS-STARTPTS'] | |
826 | self._fixup('Fixing frame timestamp', info['filepath'], opts + ['-map', '0', '-dn', '-ss', self.trim]) | |
827 | return [], info | |
828 | ||
829 | ||
830 | class FFmpegFixupDurationPP(FFmpegFixupPostProcessor): | |
831 | @PostProcessor._restrict_to(images=False) | |
832 | def run(self, info): | |
833 | self._fixup('Fixing video duration', info['filepath'], ['-c', 'copy', '-map', '0', '-dn']) | |
834 | return [], info | |
835 | ||
836 | ||
e9fade72 | 837 | class FFmpegSubtitlesConvertorPP(FFmpegPostProcessor): |
81a23040 | 838 | SUPPORTED_EXTS = ('srt', 'vtt', 'ass', 'lrc') |
839 | ||
e9fade72 JMF |
840 | def __init__(self, downloader=None, format=None): |
841 | super(FFmpegSubtitlesConvertorPP, self).__init__(downloader) | |
842 | self.format = format | |
843 | ||
844 | def run(self, info): | |
845 | subs = info.get('requested_subtitles') | |
e9fade72 JMF |
846 | new_ext = self.format |
847 | new_format = new_ext | |
848 | if new_format == 'vtt': | |
849 | new_format = 'webvtt' | |
850 | if subs is None: | |
1b77b347 | 851 | self.to_screen('There aren\'t any subtitles to convert') |
592e97e8 | 852 | return [], info |
1b77b347 | 853 | self.to_screen('Converting subtitles') |
e04398e3 | 854 | sub_filenames = [] |
e9fade72 | 855 | for lang, sub in subs.items(): |
a1c39673 | 856 | if not os.path.exists(sub.get('filepath', '')): |
857 | self.report_warning(f'Skipping embedding {lang} subtitle because the file is missing') | |
858 | continue | |
e9fade72 JMF |
859 | ext = sub['ext'] |
860 | if ext == new_ext: | |
1b77b347 | 861 | self.to_screen('Subtitle file for %s is already in the requested format' % new_ext) |
e9fade72 | 862 | continue |
503d4a44 | 863 | elif ext == 'json': |
1b77b347 | 864 | self.to_screen( |
865 | 'You have requested to convert json subtitles into another format, ' | |
503d4a44 | 866 | 'which is currently not possible') |
867 | continue | |
dcf64d43 | 868 | old_file = sub['filepath'] |
e04398e3 | 869 | sub_filenames.append(old_file) |
dcf64d43 | 870 | new_file = replace_extension(old_file, new_ext) |
bf6427d2 | 871 | |
40fcba5e | 872 | if ext in ('dfxp', 'ttml', 'tt'): |
f446cc66 | 873 | self.report_warning( |
1b77b347 | 874 | 'You have requested to convert dfxp (TTML) subtitles into another format, ' |
bf6427d2 YCH |
875 | 'which results in style information loss') |
876 | ||
e04398e3 | 877 | dfxp_file = old_file |
dcf64d43 | 878 | srt_file = replace_extension(old_file, 'srt') |
bf6427d2 | 879 | |
3869028f | 880 | with open(dfxp_file, 'rb') as f: |
bf6427d2 YCH |
881 | srt_data = dfxp2srt(f.read()) |
882 | ||
883 | with io.open(srt_file, 'wt', encoding='utf-8') as f: | |
884 | f.write(srt_data) | |
7e62c2eb | 885 | old_file = srt_file |
bf6427d2 | 886 | |
bf6427d2 YCH |
887 | subs[lang] = { |
888 | 'ext': 'srt', | |
dcf64d43 | 889 | 'data': srt_data, |
890 | 'filepath': srt_file, | |
bf6427d2 YCH |
891 | } |
892 | ||
893 | if new_ext == 'srt': | |
894 | continue | |
7b8b007c JMF |
895 | else: |
896 | sub_filenames.append(srt_file) | |
bf6427d2 | 897 | |
e04398e3 | 898 | self.run_ffmpeg(old_file, new_file, ['-f', new_format]) |
e9fade72 JMF |
899 | |
900 | with io.open(new_file, 'rt', encoding='utf-8') as f: | |
901 | subs[lang] = { | |
3547d265 | 902 | 'ext': new_ext, |
e9fade72 | 903 | 'data': f.read(), |
dcf64d43 | 904 | 'filepath': new_file, |
e9fade72 JMF |
905 | } |
906 | ||
dcf64d43 | 907 | info['__files_to_move'][new_file] = replace_extension( |
37242e56 | 908 | info['__files_to_move'][sub['filepath']], new_ext) |
dcf64d43 | 909 | |
e04398e3 | 910 | return sub_filenames, info |
72755351 | 911 | |
912 | ||
913 | class FFmpegSplitChaptersPP(FFmpegPostProcessor): | |
7a340e0d NA |
914 | def __init__(self, downloader, force_keyframes=False): |
915 | FFmpegPostProcessor.__init__(self, downloader) | |
916 | self._force_keyframes = force_keyframes | |
72755351 | 917 | |
918 | def _prepare_filename(self, number, chapter, info): | |
919 | info = info.copy() | |
920 | info.update({ | |
921 | 'section_number': number, | |
922 | 'section_title': chapter.get('title'), | |
923 | 'section_start': chapter.get('start_time'), | |
924 | 'section_end': chapter.get('end_time'), | |
925 | }) | |
926 | return self._downloader.prepare_filename(info, 'chapter') | |
927 | ||
928 | def _ffmpeg_args_for_chapter(self, number, chapter, info): | |
929 | destination = self._prepare_filename(number, chapter, info) | |
930 | if not self._downloader._ensure_dir_exists(encodeFilename(destination)): | |
931 | return | |
932 | ||
dcf64d43 | 933 | chapter['filepath'] = destination |
72755351 | 934 | self.to_screen('Chapter %03d; Destination: %s' % (number, destination)) |
935 | return ( | |
936 | destination, | |
937 | ['-ss', compat_str(chapter['start_time']), | |
a94bfd6c | 938 | '-t', compat_str(chapter['end_time'] - chapter['start_time'])]) |
72755351 | 939 | |
8326b00a | 940 | @PostProcessor._restrict_to(images=False) |
72755351 | 941 | def run(self, info): |
942 | chapters = info.get('chapters') or [] | |
943 | if not chapters: | |
7a340e0d | 944 | self.to_screen('Chapter information is unavailable') |
72755351 | 945 | return [], info |
946 | ||
7a340e0d NA |
947 | in_file = info['filepath'] |
948 | if self._force_keyframes and len(chapters) > 1: | |
949 | in_file = self.force_keyframes(in_file, (c['start_time'] for c in chapters)) | |
72755351 | 950 | self.to_screen('Splitting video by chapters; %d chapters found' % len(chapters)) |
951 | for idx, chapter in enumerate(chapters): | |
952 | destination, opts = self._ffmpeg_args_for_chapter(idx + 1, chapter, info) | |
7a340e0d NA |
953 | self.real_run_ffmpeg([(in_file, opts)], [(destination, ['-c', 'copy'])]) |
954 | if in_file != info['filepath']: | |
955 | os.remove(in_file) | |
72755351 | 956 | return [], info |
8fa43c73 | 957 | |
958 | ||
959 | class FFmpegThumbnailsConvertorPP(FFmpegPostProcessor): | |
81a23040 | 960 | SUPPORTED_EXTS = ('jpg', 'png') |
961 | ||
8fa43c73 | 962 | def __init__(self, downloader=None, format=None): |
963 | super(FFmpegThumbnailsConvertorPP, self).__init__(downloader) | |
964 | self.format = format | |
965 | ||
966 | @staticmethod | |
967 | def is_webp(path): | |
968 | with open(encodeFilename(path), 'rb') as f: | |
969 | b = f.read(12) | |
970 | return b[0:4] == b'RIFF' and b[8:] == b'WEBP' | |
971 | ||
972 | def fixup_webp(self, info, idx=-1): | |
973 | thumbnail_filename = info['thumbnails'][idx]['filepath'] | |
974 | _, thumbnail_ext = os.path.splitext(thumbnail_filename) | |
975 | if thumbnail_ext: | |
976 | thumbnail_ext = thumbnail_ext[1:].lower() | |
977 | if thumbnail_ext != 'webp' and self.is_webp(thumbnail_filename): | |
978 | self.to_screen('Correcting thumbnail "%s" extension to webp' % thumbnail_filename) | |
979 | webp_filename = replace_extension(thumbnail_filename, 'webp') | |
d75201a8 | 980 | os.replace(thumbnail_filename, webp_filename) |
8fa43c73 | 981 | info['thumbnails'][idx]['filepath'] = webp_filename |
982 | info['__files_to_move'][webp_filename] = replace_extension( | |
983 | info['__files_to_move'].pop(thumbnail_filename), 'webp') | |
984 | ||
81a23040 | 985 | @staticmethod |
986 | def _options(target_ext): | |
987 | if target_ext == 'jpg': | |
988 | return ['-bsf:v', 'mjpeg2jpeg'] | |
989 | return [] | |
990 | ||
991 | def convert_thumbnail(self, thumbnail_filename, target_ext): | |
81a23040 | 992 | thumbnail_conv_filename = replace_extension(thumbnail_filename, target_ext) |
337e0c62 | 993 | |
994 | self.to_screen('Converting thumbnail "%s" to %s' % (thumbnail_filename, target_ext)) | |
995 | self.real_run_ffmpeg( | |
996 | [(thumbnail_filename, ['-f', 'image2', '-pattern_type', 'none'])], | |
997 | [(thumbnail_conv_filename.replace('%', '%%'), self._options(target_ext))]) | |
a927acb1 | 998 | return thumbnail_conv_filename |
8fa43c73 | 999 | |
1000 | def run(self, info): | |
8fa43c73 | 1001 | files_to_delete = [] |
1002 | has_thumbnail = False | |
1003 | ||
1004 | for idx, thumbnail_dict in enumerate(info['thumbnails']): | |
1005 | if 'filepath' not in thumbnail_dict: | |
1006 | continue | |
1007 | has_thumbnail = True | |
1008 | self.fixup_webp(info, idx) | |
1009 | original_thumbnail = thumbnail_dict['filepath'] | |
1010 | _, thumbnail_ext = os.path.splitext(original_thumbnail) | |
1011 | if thumbnail_ext: | |
1012 | thumbnail_ext = thumbnail_ext[1:].lower() | |
15a4fd53 | 1013 | if thumbnail_ext == 'jpeg': |
1014 | thumbnail_ext = 'jpg' | |
8fa43c73 | 1015 | if thumbnail_ext == self.format: |
1016 | self.to_screen('Thumbnail "%s" is already in the requested format' % original_thumbnail) | |
1017 | continue | |
1018 | thumbnail_dict['filepath'] = self.convert_thumbnail(original_thumbnail, self.format) | |
1019 | files_to_delete.append(original_thumbnail) | |
1020 | info['__files_to_move'][thumbnail_dict['filepath']] = replace_extension( | |
1021 | info['__files_to_move'][original_thumbnail], self.format) | |
1022 | ||
1023 | if not has_thumbnail: | |
1024 | self.to_screen('There aren\'t any thumbnails to convert') | |
1025 | return files_to_delete, info |