aboutsummaryrefslogtreecommitdiff
path: root/youtube_dl/PostProcessor.py
blob: 6da24f986f855cded32a76a2fc60d3d420c8fdb2 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
#!/usr/bin/env python
# -*- coding: utf-8 -*-

from __future__ import absolute_import

import os
import subprocess
import sys
import time

from .utils import *


class PostProcessor(object):
    """Post Processor class.

    PostProcessor objects can be added to downloaders with their
    add_post_processor() method. When the downloader has finished a
    successful download, it will take its internal chain of PostProcessors
    and start calling the run() method on each one of them, first with
    an initial argument and then with the returned value of the previous
    PostProcessor.

    The chain will be stopped if one of them ever returns None or the end
    of the chain is reached.

    PostProcessor objects follow a "mutual registration" process similar
    to InfoExtractor objects.
    """

    _downloader = None

    def __init__(self, downloader=None):
        self._downloader = downloader

    def set_downloader(self, downloader):
        """Sets the downloader for this PP."""
        self._downloader = downloader

    def run(self, information):
        """Run the PostProcessor.

        The "information" argument is a dictionary like the ones
        composed by InfoExtractors. The only difference is that this
        one has an extra field called "filepath" that points to the
        downloaded file.

        When this method returns None, the postprocessing chain is
        stopped. However, this method may return an information
        dictionary that will be passed to the next postprocessing
        object in the chain. It can be the one it received after
        changing some fields.

        In addition, this method may raise a PostProcessingError
        exception that will be taken into account by the downloader
        it was called from.
        """
        return information # by default, do nothing

class AudioConversionError(BaseException):
    def __init__(self, message):
        self.message = message

class FFmpegExtractAudioPP(PostProcessor):
    def __init__(self, downloader=None, preferredcodec=None, preferredquality=None, keepvideo=False, nopostoverwrites=False):
        PostProcessor.__init__(self, downloader)
        if preferredcodec is None:
            preferredcodec = 'best'
        self._preferredcodec = preferredcodec
        self._preferredquality = preferredquality
        self._keepvideo = keepvideo
        self._nopostoverwrites = nopostoverwrites
        self._exes = self.detect_executables()

    @staticmethod
    def detect_executables():
        def executable(exe):
            try:
                subprocess.Popen([exe, '-version'], stdout=subprocess.PIPE, stderr=subprocess.PIPE).communicate()
            except OSError:
                return False
            return exe
        programs = ['avprobe', 'avconv', 'ffmpeg', 'ffprobe']
        return dict((program, executable(program)) for program in programs)

    def get_audio_codec(self, path):
        if not self._exes['ffprobe'] and not self._exes['avprobe']: return None
        try:
            cmd = [self._exes['avprobe'] or self._exes['ffprobe'], '-show_streams', '--', encodeFilename(path)]
            handle = subprocess.Popen(cmd, stderr=compat_subprocess_get_DEVNULL(), stdout=subprocess.PIPE)
            output = handle.communicate()[0]
            if handle.wait() != 0:
                return None
        except (IOError, OSError):
            return None
        audio_codec = None
        for line in output.decode('ascii', 'ignore').split('\n'):
            if line.startswith('codec_name='):
                audio_codec = line.split('=')[1].strip()
            elif line.strip() == 'codec_type=audio' and audio_codec is not None:
                return audio_codec
        return None

    def run_ffmpeg(self, path, out_path, codec, more_opts):
        if not self._exes['ffmpeg'] and not self._exes['avconv']:
            raise AudioConversionError('ffmpeg or avconv not found. Please install one.')
        if codec is None:
            acodec_opts = []
        else:
            acodec_opts = ['-acodec', codec]
        if self._nopostoverwrites and self._exes['ffmpeg']:
            overwrite_opts = '-n'
        else:
            overwrite_opts = '-y'
        cmd = ([self._exes['avconv'] or self._exes['ffmpeg'], overwrite_opts, '-i', encodeFilename(path), '-vn']
               + acodec_opts + more_opts +
               ['--', encodeFilename(out_path)])
        p = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
        stdout,stderr = p.communicate()
        if p.returncode != 0:
            msg = stderr.strip().split('\n')[-1]
            raise AudioConversionError(msg)

    def run(self, information):
        path = information['filepath']

        filecodec = self.get_audio_codec(path)
        if filecodec is None:
            self._downloader.to_stderr(u'WARNING: unable to obtain file audio codec with ffprobe')
            return None

        more_opts = []
        if self._preferredcodec == 'best' or self._preferredcodec == filecodec or (self._preferredcodec == 'm4a' and filecodec == 'aac'):
            if self._preferredcodec == 'm4a' and filecodec == 'aac':
                # Lossless, but in another container
                acodec = 'copy'
                extension = self._preferredcodec
                more_opts = [self._exes['avconv'] and '-bsf:a' or '-absf', 'aac_adtstoasc']
            elif filecodec in ['aac', 'mp3', 'vorbis']:
                # Lossless if possible
                acodec = 'copy'
                extension = filecodec
                if filecodec == 'aac':
                    more_opts = ['-f', 'adts']
                if filecodec == 'vorbis':
                    extension = 'ogg'
            else:
                # MP3 otherwise.
                acodec = 'libmp3lame'
                extension = 'mp3'
                more_opts = []
                if self._preferredquality is not None:
                    if int(self._preferredquality) < 10:
                        more_opts += [self._exes['avconv'] and '-q:a' or '-aq', self._preferredquality]
                    else:
                        more_opts += [self._exes['avconv'] and '-b:a' or '-ab', self._preferredquality + 'k']
        else:
            # We convert the audio (lossy)
            acodec = {'mp3': 'libmp3lame', 'aac': 'aac', 'm4a': 'aac', 'vorbis': 'libvorbis', 'wav': None}[self._preferredcodec]
            extension = self._preferredcodec
            more_opts = []
            if self._preferredquality is not None:
                if int(self._preferredquality) < 10:
                    more_opts += [self._exes['avconv'] and '-q:a' or '-aq', self._preferredquality]
                else:
                    more_opts += [self._exes['avconv'] and '-b:a' or '-ab', self._preferredquality + 'k']
            if self._preferredcodec == 'aac':
                more_opts += ['-f', 'adts']
            if self._preferredcodec == 'm4a':
                more_opts += [self._exes['avconv'] and '-bsf:a' or '-absf', 'aac_adtstoasc']
            if self._preferredcodec == 'vorbis':
                extension = 'ogg'
            if self._preferredcodec == 'wav':
                extension = 'wav'
                more_opts += ['-f', 'wav']

        prefix, sep, ext = path.rpartition(u'.') # not os.path.splitext, since the latter does not work on unicode in all setups
        new_path = prefix + sep + extension
        self._downloader.to_screen(u'[' + (self._exes['avconv'] and 'avconv' or 'ffmpeg') + '] Destination: ' + new_path)
        try:
            self.run_ffmpeg(path, new_path, acodec, more_opts)
        except:
            etype,e,tb = sys.exc_info()
            if isinstance(e, AudioConversionError):
                self._downloader.to_stderr(u'ERROR: audio conversion failed: ' + e.message)
            else:
                self._downloader.to_stderr(u'ERROR: error running ' + (self._exes['avconv'] and 'avconv' or 'ffmpeg'))
            return None

        # Try to update the date time for extracted audio file.
        if information.get('filetime') is not None:
            try:
                os.utime(encodeFilename(new_path), (time.time(), information['filetime']))
            except:
                self._downloader.to_stderr(u'WARNING: Cannot update utime of audio file')

        if not self._keepvideo:
            try:
                os.remove(encodeFilename(path))
            except (IOError, OSError):
                self._downloader.to_stderr(u'WARNING: Unable to remove downloaded video file')
                return None

        information['filepath'] = new_path
        return information