yt-dlp/yt_dlp/downloader/common.py

433 lines
16 KiB
Python
Raw Normal View History

from __future__ import division, unicode_literals
2014-11-16 14:06:59 +00:00
import copy
import os
import re
import time
import random
from ..utils import (
decodeArgument,
encodeFilename,
error_to_compat_str,
format_bytes,
shell_quote,
timeconvert,
)
from ..minicurses import (
MultilineLogger,
MultilinePrinter,
QuietMultilinePrinter,
BreaklineStatusPrinter
)
class FileDownloader(object):
"""File Downloader class.
File downloader objects are the ones responsible of downloading the
actual video file and writing it to disk.
File downloaders accept a lot of parameters. In order not to saturate
the object constructor with arguments, it receives a dictionary of
options instead.
Available options:
verbose: Print additional info to stdout.
quiet: Do not print messages to stdout.
ratelimit: Download speed limit, in bytes/sec.
throttledratelimit: Assume the download is being throttled below this speed (bytes/sec)
retries: Number of times to retry for HTTP error 5xx
buffersize: Size of download buffer in bytes.
noresizebuffer: Do not automatically resize the download buffer.
continuedl: Try to continue downloads if possible.
noprogress: Do not print the progress bar.
nopart: Do not use temporary .part files.
updatetime: Use the Last-modified header to set output file timestamps.
test: Download only first bytes to test the downloader.
min_filesize: Skip files smaller than this size
max_filesize: Skip files larger than this size
xattr_set_filesize: Set ytdl.filesize user xattribute with expected size.
external_downloader_args: A dictionary of downloader keys (in lower case)
and a list of additional command-line arguments for the
executable. Use 'default' as the name for arguments to be
passed to all downloaders. For compatibility with youtube-dl,
a single list of args can also be used
hls_use_mpegts: Use the mpegts container for HLS videos.
2018-02-17 12:10:12 +00:00
http_chunk_size: Size of a chunk for chunk-based HTTP downloading. May be
2018-02-03 19:53:50 +00:00
useful for bypassing bandwidth throttling imposed by
a webserver (experimental)
progress_template: See YoutubeDL.py
Subclasses of this one must re-define the real_download method.
"""
2014-09-24 12:38:40 +00:00
_TEST_FILE_SIZE = 10241
params = None
def __init__(self, ydl, params):
"""Create a FileDownloader object with the given options."""
self.ydl = ydl
self._progress_hooks = []
self.params = params
self._prepare_multiline_status()
self.add_progress_hook(self.report_progress)
@staticmethod
def format_seconds(seconds):
(mins, secs) = divmod(seconds, 60)
(hours, mins) = divmod(mins, 60)
if hours > 99:
return '--:--:--'
if hours == 0:
return '%02d:%02d' % (mins, secs)
else:
return '%02d:%02d:%02d' % (hours, mins, secs)
@staticmethod
def calc_percent(byte_counter, data_len):
if data_len is None:
return None
return float(byte_counter) / float(data_len) * 100.0
@staticmethod
def format_percent(percent):
if percent is None:
return '---.-%'
return '%6s' % ('%3.1f%%' % percent)
@staticmethod
def calc_eta(start, now, total, current):
if total is None:
return None
if now is None:
now = time.time()
dif = now - start
2014-11-23 19:41:03 +00:00
if current == 0 or dif < 0.001: # One millisecond
return None
rate = float(current) / dif
return int((float(total) - float(current)) / rate)
@staticmethod
def format_eta(eta):
if eta is None:
return '--:--'
return FileDownloader.format_seconds(eta)
@staticmethod
def calc_speed(start, now, bytes):
dif = now - start
2014-11-23 19:41:03 +00:00
if bytes == 0 or dif < 0.001: # One millisecond
return None
return float(bytes) / dif
@staticmethod
def format_speed(speed):
if speed is None:
return '%10s' % '---b/s'
return '%10s' % ('%s/s' % format_bytes(speed))
@staticmethod
def format_retries(retries):
return 'inf' if retries == float('inf') else '%.0f' % retries
@staticmethod
def best_block_size(elapsed_time, bytes):
new_min = max(bytes / 2.0, 1.0)
2014-11-23 19:41:03 +00:00
new_max = min(max(bytes * 2.0, 1.0), 4194304) # Do not surpass 4 MB
if elapsed_time < 0.001:
return int(new_max)
rate = bytes / elapsed_time
if rate > new_max:
return int(new_max)
if rate < new_min:
return int(new_min)
return int(rate)
@staticmethod
def parse_bytes(bytestr):
"""Parse a string indicating a byte quantity into an integer."""
matchobj = re.match(r'(?i)^(\d+(?:\.\d+)?)([kMGTPEZY]?)$', bytestr)
if matchobj is None:
return None
number = float(matchobj.group(1))
multiplier = 1024.0 ** 'bkmgtpezy'.index(matchobj.group(2).lower())
return int(round(number * multiplier))
def to_screen(self, *args, **kargs):
self.ydl.to_stdout(*args, quiet=self.params.get('quiet'), **kargs)
def to_stderr(self, message):
self.ydl.to_stderr(message)
def to_console_title(self, message):
self.ydl.to_console_title(message)
def trouble(self, *args, **kargs):
self.ydl.trouble(*args, **kargs)
def report_warning(self, *args, **kargs):
self.ydl.report_warning(*args, **kargs)
def report_error(self, *args, **kargs):
self.ydl.report_error(*args, **kargs)
2021-05-19 12:04:17 +00:00
def write_debug(self, *args, **kargs):
self.ydl.write_debug(*args, **kargs)
def slow_down(self, start_time, now, byte_counter):
"""Sleep if the download speed is over the rate limit."""
rate_limit = self.params.get('ratelimit')
if rate_limit is None or byte_counter == 0:
return
if now is None:
now = time.time()
elapsed = now - start_time
if elapsed <= 0.0:
return
speed = float(byte_counter) / elapsed
if speed > rate_limit:
sleep_time = float(byte_counter) / rate_limit - elapsed
if sleep_time > 0:
time.sleep(sleep_time)
def temp_name(self, filename):
"""Returns a temporary filename for the given filename."""
2014-11-16 14:06:59 +00:00
if self.params.get('nopart', False) or filename == '-' or \
(os.path.exists(encodeFilename(filename)) and not os.path.isfile(encodeFilename(filename))):
return filename
2014-11-16 14:06:59 +00:00
return filename + '.part'
def undo_temp_name(self, filename):
2014-11-16 14:06:59 +00:00
if filename.endswith('.part'):
return filename[:-len('.part')]
return filename
def ytdl_filename(self, filename):
return filename + '.ytdl'
def try_rename(self, old_filename, new_filename):
if old_filename == new_filename:
return
try:
os.replace(old_filename, new_filename)
except (IOError, OSError) as err:
self.report_error(f'unable to rename file: {err}')
def try_utime(self, filename, last_modified_hdr):
"""Try to set the last-modified time of the given file."""
if last_modified_hdr is None:
return
if not os.path.isfile(encodeFilename(filename)):
return
timestr = last_modified_hdr
if timestr is None:
return
filetime = timeconvert(timestr)
if filetime is None:
return filetime
# Ignore obviously invalid dates
if filetime == 0:
return
try:
os.utime(filename, (time.time(), filetime))
except Exception:
pass
return filetime
def report_destination(self, filename):
"""Report destination filename."""
2014-11-16 14:06:59 +00:00
self.to_screen('[download] Destination: ' + filename)
def _prepare_multiline_status(self, lines=1):
if self.params.get('noprogress'):
self._multiline = QuietMultilinePrinter()
elif self.ydl.params.get('logger'):
self._multiline = MultilineLogger(self.ydl.params['logger'], lines)
elif self.params.get('progress_with_newline'):
self._multiline = BreaklineStatusPrinter(self.ydl._screen_file, lines)
else:
self._multiline = MultilinePrinter(self.ydl._screen_file, lines, not self.params.get('quiet'))
def _finish_multiline_status(self):
self._multiline.end()
def _report_progress_status(self, s):
progress_dict = s.copy()
progress_dict.pop('info_dict')
progress_dict = {'info': s['info_dict'], 'progress': progress_dict}
progress_template = self.params.get('progress_template', {})
self._multiline.print_at_line(self.ydl.evaluate_outtmpl(
progress_template.get('download') or '[download] %(progress._default_template)s',
progress_dict), s.get('progress_idx') or 0)
self.to_console_title(self.ydl.evaluate_outtmpl(
progress_template.get('download-title') or 'yt-dlp %(progress._default_template)s',
progress_dict))
def report_progress(self, s):
if s['status'] == 'finished':
if self.params.get('noprogress'):
self.to_screen('[download] Download completed')
msg_template = '100%%'
if s.get('total_bytes') is not None:
s['_total_bytes_str'] = format_bytes(s['total_bytes'])
msg_template += ' of %(_total_bytes_str)s'
if s.get('elapsed') is not None:
s['_elapsed_str'] = self.format_seconds(s['elapsed'])
msg_template += ' in %(_elapsed_str)s'
s['_percent_str'] = self.format_percent(100)
s['_default_template'] = msg_template % s
self._report_progress_status(s)
return
if s['status'] != 'downloading':
return
if s.get('eta') is not None:
s['_eta_str'] = self.format_eta(s['eta'])
else:
s['_eta_str'] = 'Unknown ETA'
if s.get('total_bytes') and s.get('downloaded_bytes') is not None:
s['_percent_str'] = self.format_percent(100 * s['downloaded_bytes'] / s['total_bytes'])
elif s.get('total_bytes_estimate') and s.get('downloaded_bytes') is not None:
s['_percent_str'] = self.format_percent(100 * s['downloaded_bytes'] / s['total_bytes_estimate'])
else:
if s.get('downloaded_bytes') == 0:
s['_percent_str'] = self.format_percent(0)
else:
s['_percent_str'] = 'Unknown %'
if s.get('speed') is not None:
s['_speed_str'] = self.format_speed(s['speed'])
else:
s['_speed_str'] = 'Unknown speed'
if s.get('total_bytes') is not None:
s['_total_bytes_str'] = format_bytes(s['total_bytes'])
msg_template = '%(_percent_str)s of %(_total_bytes_str)s at %(_speed_str)s ETA %(_eta_str)s'
elif s.get('total_bytes_estimate') is not None:
s['_total_bytes_estimate_str'] = format_bytes(s['total_bytes_estimate'])
msg_template = '%(_percent_str)s of ~%(_total_bytes_estimate_str)s at %(_speed_str)s ETA %(_eta_str)s'
else:
if s.get('downloaded_bytes') is not None:
s['_downloaded_bytes_str'] = format_bytes(s['downloaded_bytes'])
if s.get('elapsed'):
s['_elapsed_str'] = self.format_seconds(s['elapsed'])
msg_template = '%(_downloaded_bytes_str)s at %(_speed_str)s (%(_elapsed_str)s)'
else:
msg_template = '%(_downloaded_bytes_str)s at %(_speed_str)s'
else:
msg_template = '%(_percent_str)s % at %(_speed_str)s ETA %(_eta_str)s'
s['_default_template'] = msg_template % s
self._report_progress_status(s)
def report_resuming_byte(self, resume_len):
"""Report attempt to resume at given byte."""
2014-11-16 14:06:59 +00:00
self.to_screen('[download] Resuming download at byte %s' % resume_len)
def report_retry(self, err, count, retries):
"""Report retry in case of HTTP error 5xx"""
self.to_screen(
2021-03-01 00:09:50 +00:00
'[download] Got server HTTP error: %s. Retrying (attempt %d of %s) ...'
% (error_to_compat_str(err), count, self.format_retries(retries)))
2021-08-07 11:20:46 +00:00
def report_file_already_downloaded(self, *args, **kwargs):
"""Report file has already been fully downloaded."""
2021-08-07 11:20:46 +00:00
return self.ydl.report_file_already_downloaded(*args, **kwargs)
def report_unable_to_resume(self):
"""Report it was impossible to resume download."""
2014-11-16 14:06:59 +00:00
self.to_screen('[download] Unable to resume')
@staticmethod
def supports_manifest(manifest):
""" Whether the downloader can download the fragments from the manifest.
Redefine in subclasses if needed. """
pass
def download(self, filename, info_dict, subtitle=False):
"""Download to a filename using the info from info_dict
Return True on success and False otherwise
"""
nooverwrites_and_exists = (
not self.params.get('overwrites', True)
and os.path.exists(encodeFilename(filename))
)
if not hasattr(filename, 'write'):
continuedl_and_exists = (
self.params.get('continuedl', True)
and os.path.isfile(encodeFilename(filename))
and not self.params.get('nopart', False)
)
# Check file already present
if filename != '-' and (nooverwrites_and_exists or continuedl_and_exists):
self.report_file_already_downloaded(filename)
self._hook_progress({
'filename': filename,
'status': 'finished',
'total_bytes': os.path.getsize(encodeFilename(filename)),
}, info_dict)
2020-11-15 00:28:41 +00:00
return True, False
2013-12-23 15:03:06 +00:00
if subtitle is False:
min_sleep_interval = self.params.get('sleep_interval')
if min_sleep_interval:
max_sleep_interval = self.params.get('max_sleep_interval', min_sleep_interval)
sleep_interval = random.uniform(min_sleep_interval, max_sleep_interval)
self.to_screen(
2021-03-01 00:09:50 +00:00
'[download] Sleeping %s seconds ...' % (
int(sleep_interval) if sleep_interval.is_integer()
else '%.2f' % sleep_interval))
time.sleep(sleep_interval)
else:
[common] Make sure self.params.get('sleep_interval_subtitles') is int This can happen if another software is using yt-dlc'API (ie: tubeup). The stack trace would be: $ tubeup 'https://youtube.com/watch?v=JyE9OF03cao' [debug] Encodings: locale UTF-8, fs utf-8, out UTF-8, pref UTF-8 [debug] youtube-dlc version 2020.10.25 [debug] Python version 3.7.3 (CPython) - Linux-5.8.0-0.bpo.2-amd64-x86_64-with-debian-10.6 [debug] exe versions: ffmpeg 3.3.9, ffprobe 3.3.9 [debug] Proxy map: {} There are no annotations to write. ERROR: '>' not supported between instances of 'NoneType' and 'int' Traceback (most recent call last): File "/mnt/data2/Backup/Wiki/.local/lib/python3.7/site-packages/youtube_dlc/YoutubeDL.py", line 846, in extract_info return self.process_ie_result(ie_result, download, extra_info) File "/mnt/data2/Backup/Wiki/.local/lib/python3.7/site-packages/youtube_dlc/YoutubeDL.py", line 901, in process_ie_result return self.process_video_result(ie_result, download=download) File "/mnt/data2/Backup/Wiki/.local/lib/python3.7/site-packages/youtube_dlc/YoutubeDL.py", line 1696, in process_video_result self.process_info(new_info) File "/mnt/data2/Backup/Wiki/.local/lib/python3.7/site-packages/youtube_dlc/YoutubeDL.py", line 1894, in process_info dl(sub_filename, sub_info, subtitle=True) File "/mnt/data2/Backup/Wiki/.local/lib/python3.7/site-packages/youtube_dlc/YoutubeDL.py", line 1866, in dl return fd.download(name, info, subtitle) File "/mnt/data2/Backup/Wiki/.local/lib/python3.7/site-packages/youtube_dlc/downloader/common.py", line 367, in download if self.params.get('sleep_interval_subtitles') > 0: TypeError: '>' not supported between instances of 'NoneType' and 'int'
2020-11-08 07:36:26 +00:00
sleep_interval_sub = 0
if type(self.params.get('sleep_interval_subtitles')) is int:
2020-11-01 02:36:16 +00:00
sleep_interval_sub = self.params.get('sleep_interval_subtitles')
[common] Make sure self.params.get('sleep_interval_subtitles') is int This can happen if another software is using yt-dlc'API (ie: tubeup). The stack trace would be: $ tubeup 'https://youtube.com/watch?v=JyE9OF03cao' [debug] Encodings: locale UTF-8, fs utf-8, out UTF-8, pref UTF-8 [debug] youtube-dlc version 2020.10.25 [debug] Python version 3.7.3 (CPython) - Linux-5.8.0-0.bpo.2-amd64-x86_64-with-debian-10.6 [debug] exe versions: ffmpeg 3.3.9, ffprobe 3.3.9 [debug] Proxy map: {} There are no annotations to write. ERROR: '>' not supported between instances of 'NoneType' and 'int' Traceback (most recent call last): File "/mnt/data2/Backup/Wiki/.local/lib/python3.7/site-packages/youtube_dlc/YoutubeDL.py", line 846, in extract_info return self.process_ie_result(ie_result, download, extra_info) File "/mnt/data2/Backup/Wiki/.local/lib/python3.7/site-packages/youtube_dlc/YoutubeDL.py", line 901, in process_ie_result return self.process_video_result(ie_result, download=download) File "/mnt/data2/Backup/Wiki/.local/lib/python3.7/site-packages/youtube_dlc/YoutubeDL.py", line 1696, in process_video_result self.process_info(new_info) File "/mnt/data2/Backup/Wiki/.local/lib/python3.7/site-packages/youtube_dlc/YoutubeDL.py", line 1894, in process_info dl(sub_filename, sub_info, subtitle=True) File "/mnt/data2/Backup/Wiki/.local/lib/python3.7/site-packages/youtube_dlc/YoutubeDL.py", line 1866, in dl return fd.download(name, info, subtitle) File "/mnt/data2/Backup/Wiki/.local/lib/python3.7/site-packages/youtube_dlc/downloader/common.py", line 367, in download if self.params.get('sleep_interval_subtitles') > 0: TypeError: '>' not supported between instances of 'NoneType' and 'int'
2020-11-08 07:36:26 +00:00
if sleep_interval_sub > 0:
2020-11-01 02:36:16 +00:00
self.to_screen(
2021-03-01 00:09:50 +00:00
'[download] Sleeping %s seconds ...' % (
2020-11-01 02:36:16 +00:00
sleep_interval_sub))
time.sleep(sleep_interval_sub)
ret = self.real_download(filename, info_dict)
self._finish_multiline_status()
return ret, True
def real_download(self, filename, info_dict):
"""Real download process. Redefine in subclasses."""
2014-11-16 14:06:59 +00:00
raise NotImplementedError('This method must be implemented by subclasses')
def _hook_progress(self, status, info_dict):
if not self._progress_hooks:
return
info_dict = dict(info_dict)
for key in ('__original_infodict', '__postprocessors'):
info_dict.pop(key, None)
# youtube-dl passes the same status object to all the hooks.
# Some third party scripts seems to be relying on this.
# So keep this behavior if possible
status['info_dict'] = copy.deepcopy(info_dict)
for ph in self._progress_hooks:
ph(status)
def add_progress_hook(self, ph):
# See YoutubeDl.py (search for progress_hooks) for a description of
# this interface
self._progress_hooks.append(ph)
2015-04-25 22:30:45 +00:00
def _debug_cmd(self, args, exe=None):
if not self.params.get('verbose', False):
return
2015-04-25 22:30:45 +00:00
str_args = [decodeArgument(a) for a in args]
if exe is None:
2015-04-25 22:30:45 +00:00
exe = os.path.basename(str_args[0])
2021-05-14 07:45:29 +00:00
self.write_debug('%s command line: %s' % (exe, shell_quote(str_args)))