yt-dlp/youtube_dlc/postprocessor/sponskrub.py

91 lines
3.7 KiB
Python
Raw Normal View History

2020-11-15 00:28:41 +00:00
from __future__ import unicode_literals
import os
import subprocess
from .common import PostProcessor
from ..compat import compat_shlex_split
from ..utils import (
check_executable,
encodeArgument,
2021-01-17 23:52:47 +00:00
encodeFilename,
2020-11-15 00:28:41 +00:00
shell_quote,
str_or_none,
2020-11-15 00:28:41 +00:00
PostProcessingError,
2021-01-17 23:52:47 +00:00
prepend_extension,
2020-11-15 00:28:41 +00:00
)
class SponSkrubPP(PostProcessor):
_temp_ext = 'spons'
_exe_name = 'sponskrub'
def __init__(self, downloader, path='', args=None, ignoreerror=False, cut=False, force=False):
PostProcessor.__init__(self, downloader)
self.force = force
self.cutout = cut
self.args = str_or_none(args) or '' # For backward compatibility
2020-11-15 00:28:41 +00:00
self.path = self.get_exe(path)
if not ignoreerror and self.path is None:
if path:
raise PostProcessingError('sponskrub not found in "%s"' % path)
else:
raise PostProcessingError('sponskrub not found. Please install or provide the path using --sponskrub-path.')
def get_exe(self, path=''):
if not path or not check_executable(path, ['-h']):
path = os.path.join(path, self._exe_name)
if not check_executable(path, ['-h']):
return None
return path
def run(self, information):
if self.path is None:
return [], information
if information['extractor_key'].lower() != 'youtube':
self.to_screen('Skipping sponskrub since it is not a YouTube video')
2020-11-15 00:28:41 +00:00
return [], information
if self.cutout and not self.force and not information.get('__real_download', False):
self.report_warning(
'Skipping sponskrub since the video was already downloaded. '
2020-11-15 00:28:41 +00:00
'Use --sponskrub-force to run sponskrub anyway')
return [], information
self.to_screen('Trying to %s sponsor sections' % ('remove' if self.cutout else 'mark'))
2020-11-15 00:28:41 +00:00
if self.cutout:
self.report_warning('Cutting out sponsor segments will cause the subtitles to go out of sync.')
2020-11-15 00:28:41 +00:00
if not information.get('__real_download', False):
self.report_warning('If sponskrub is run multiple times, unintended parts of the video could be cut out.')
2020-11-15 00:28:41 +00:00
filename = information['filepath']
2021-01-17 23:52:47 +00:00
temp_filename = prepend_extension(filename, self._temp_ext)
if os.path.exists(encodeFilename(temp_filename)):
os.remove(encodeFilename(temp_filename))
2020-11-15 00:28:41 +00:00
cmd = [self.path]
if not self.cutout:
cmd += ['-chapter']
cmd += compat_shlex_split(self.args) # For backward compatibility
cmd += self._configuration_args(exe=self._exe_name)
2020-11-15 00:28:41 +00:00
cmd += ['--', information['id'], filename, temp_filename]
cmd = [encodeArgument(i) for i in cmd]
self.write_debug('sponskrub command line: %s' % shell_quote(cmd))
2020-11-15 00:28:41 +00:00
p = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, stdin=subprocess.PIPE)
stdout, stderr = p.communicate()
if p.returncode == 0:
2021-01-17 23:52:47 +00:00
os.remove(encodeFilename(filename))
os.rename(encodeFilename(temp_filename), encodeFilename(filename))
self.to_screen('Sponsor sections have been %s' % ('removed' if self.cutout else 'marked'))
elif p.returncode == 3:
self.to_screen('No segments in the SponsorBlock database')
else:
msg = stderr.decode('utf-8', 'replace').strip() or stdout.decode('utf-8', 'replace').strip()
self.write_debug(msg, prefix=False)
line = 0 if msg[:12].lower() == 'unrecognised' else -1
msg = msg.split('\n')[line]
raise PostProcessingError(msg if msg else 'sponskrub failed with error code %s' % p.returncode)
2020-11-15 00:28:41 +00:00
return [], information