diff options
author | Jesús <heckyel@hyperbola.info> | 2022-12-02 05:21:10 +0800 |
---|---|---|
committer | Jesús <heckyel@hyperbola.info> | 2022-12-02 05:21:10 +0800 |
commit | eaeeef9c1d1bedb76fea953c332ef84d53bffe2c (patch) | |
tree | c3cb5582247e47fc67c24cd7ff8ea857fb76821e /hypervideo_dl/downloader | |
parent | 1e5a50b71d8f0eae6007bedc329eecb24bb5aba3 (diff) | |
download | hypervideo-eaeeef9c1d1bedb76fea953c332ef84d53bffe2c.tar.lz hypervideo-eaeeef9c1d1bedb76fea953c332ef84d53bffe2c.tar.xz hypervideo-eaeeef9c1d1bedb76fea953c332ef84d53bffe2c.zip |
update from upstream
Diffstat (limited to 'hypervideo_dl/downloader')
-rw-r--r-- | hypervideo_dl/downloader/__init__.py | 31 | ||||
-rw-r--r-- | hypervideo_dl/downloader/common.py | 333 | ||||
-rw-r--r-- | hypervideo_dl/downloader/dash.py | 23 | ||||
-rw-r--r-- | hypervideo_dl/downloader/external.py | 219 | ||||
-rw-r--r-- | hypervideo_dl/downloader/f4m.py | 54 | ||||
-rw-r--r-- | hypervideo_dl/downloader/fc2.py | 11 | ||||
-rw-r--r-- | hypervideo_dl/downloader/fragment.py | 191 | ||||
-rw-r--r-- | hypervideo_dl/downloader/hls.py | 66 | ||||
-rw-r--r-- | hypervideo_dl/downloader/http.py | 98 | ||||
-rw-r--r-- | hypervideo_dl/downloader/ism.py | 64 | ||||
-rw-r--r-- | hypervideo_dl/downloader/mhtml.py | 30 | ||||
-rw-r--r-- | hypervideo_dl/downloader/niconico.py | 13 | ||||
-rw-r--r-- | hypervideo_dl/downloader/rtmp.py | 12 | ||||
-rw-r--r-- | hypervideo_dl/downloader/rtsp.py | 9 | ||||
-rw-r--r-- | hypervideo_dl/downloader/websocket.py | 19 | ||||
-rw-r--r-- | hypervideo_dl/downloader/youtube_live_chat.py | 53 |
16 files changed, 566 insertions, 660 deletions
diff --git a/hypervideo_dl/downloader/__init__.py b/hypervideo_dl/downloader/__init__.py index 96d484d..c34dbce 100644 --- a/hypervideo_dl/downloader/__init__.py +++ b/hypervideo_dl/downloader/__init__.py @@ -1,10 +1,4 @@ -from __future__ import unicode_literals - -from ..compat import compat_str -from ..utils import ( - determine_protocol, - NO_DEFAULT -) +from ..utils import NO_DEFAULT, determine_protocol def get_suitable_downloader(info_dict, params={}, default=NO_DEFAULT, protocol=None, to_stdout=False): @@ -29,21 +23,18 @@ def get_suitable_downloader(info_dict, params={}, default=NO_DEFAULT, protocol=N # Some of these require get_suitable_downloader from .common import FileDownloader from .dash import DashSegmentsFD +from .external import FFmpegFD, get_external_downloader from .f4m import F4mFD from .fc2 import FC2LiveFD from .hls import HlsFD from .http import HttpFD -from .rtmp import RtmpFD -from .rtsp import RtspFD from .ism import IsmFD from .mhtml import MhtmlFD from .niconico import NiconicoDmcFD +from .rtmp import RtmpFD +from .rtsp import RtspFD from .websocket import WebSocketFragmentFD from .youtube_live_chat import YoutubeLiveChatFD -from .external import ( - get_external_downloader, - FFmpegFD, -) PROTOCOL_MAP = { 'rtmp': RtmpFD, @@ -68,10 +59,11 @@ PROTOCOL_MAP = { def shorten_protocol_name(proto, simplify=False): short_protocol_names = { - 'm3u8_native': 'm3u8_n', - 'rtmp_ffmpeg': 'rtmp_f', + 'm3u8_native': 'm3u8', + 'm3u8': 'm3u8F', + 'rtmp_ffmpeg': 'rtmpF', 'http_dash_segments': 'dash', - 'http_dash_segments_generator': 'dash_g', + 'http_dash_segments_generator': 'dashG', 'niconico_dmc': 'dmc', 'websocket_frag': 'WSfrag', } @@ -79,6 +71,7 @@ def shorten_protocol_name(proto, simplify=False): short_protocol_names.update({ 'https': 'http', 'ftps': 'ftp', + 'm3u8': 'm3u8', # Reverse above m3u8 mapping 'm3u8_native': 'm3u8', 'http_dash_segments_generator': 'dash', 'rtmp_ffmpeg': 'rtmp', @@ -93,13 +86,13 @@ def _get_suitable_downloader(info_dict, protocol, params, default): if default is NO_DEFAULT: default = HttpFD - # if (info_dict.get('start_time') or info_dict.get('end_time')) and not info_dict.get('requested_formats') and FFmpegFD.can_download(info_dict): - # return FFmpegFD + if (info_dict.get('section_start') or info_dict.get('section_end')) and FFmpegFD.can_download(info_dict): + return FFmpegFD info_dict['protocol'] = protocol downloaders = params.get('external_downloader') external_downloader = ( - downloaders if isinstance(downloaders, compat_str) or downloaders is None + downloaders if isinstance(downloaders, str) or downloaders is None else downloaders.get(shorten_protocol_name(protocol, True), downloaders.get('default'))) if external_downloader is None: diff --git a/hypervideo_dl/downloader/common.py b/hypervideo_dl/downloader/common.py index 7cef3e8..72d4822 100644 --- a/hypervideo_dl/downloader/common.py +++ b/hypervideo_dl/downloader/common.py @@ -1,30 +1,39 @@ -from __future__ import division, unicode_literals - +import contextlib +import errno +import functools import os +import random import re import time -import random -import errno +from ..minicurses import ( + BreaklineStatusPrinter, + MultilineLogger, + MultilinePrinter, + QuietMultilinePrinter, +) from ..utils import ( + IDENTITY, + NO_DEFAULT, + LockingUnsupportedError, + Namespace, + RetryManager, + classproperty, decodeArgument, encodeFilename, - error_to_compat_str, format_bytes, + join_nonempty, + parse_bytes, + remove_start, sanitize_open, shell_quote, timeconvert, timetuple_from_msec, -) -from ..minicurses import ( - MultilineLogger, - MultilinePrinter, - QuietMultilinePrinter, - BreaklineStatusPrinter + try_call, ) -class FileDownloader(object): +class FileDownloader: """File Downloader class. File downloader objects are the ones responsible of downloading the @@ -39,6 +48,7 @@ class FileDownloader(object): verbose: Print additional info to stdout. quiet: Do not print messages to stdout. ratelimit: Download speed limit, in bytes/sec. + continuedl: Attempt to continue downloads if possible throttledratelimit: Assume the download is being throttled below this speed (bytes/sec) retries: Number of times to retry for HTTP error 5xx file_access_retries: Number of times to retry on file access error @@ -62,6 +72,7 @@ class FileDownloader(object): useful for bypassing bandwidth throttling imposed by a webserver (experimental) progress_template: See YoutubeDL.py + retry_sleep_functions: See YoutubeDL.py Subclasses of this one must re-define the real_download method. """ @@ -71,21 +82,51 @@ class FileDownloader(object): def __init__(self, ydl, params): """Create a FileDownloader object with the given options.""" - self.ydl = ydl + self._set_ydl(ydl) self._progress_hooks = [] self.params = params self._prepare_multiline_status() self.add_progress_hook(self.report_progress) + def _set_ydl(self, ydl): + self.ydl = ydl + + for func in ( + 'deprecation_warning', + 'deprecated_feature', + 'report_error', + 'report_file_already_downloaded', + 'report_warning', + 'to_console_title', + 'to_stderr', + 'trouble', + 'write_debug', + ): + if not hasattr(self, func): + setattr(self, func, getattr(ydl, func)) + + def to_screen(self, *args, **kargs): + self.ydl.to_screen(*args, quiet=self.params.get('quiet'), **kargs) + + __to_screen = to_screen + + @classproperty + def FD_NAME(cls): + return re.sub(r'(?<=[a-z])(?=[A-Z])', '_', cls.__name__[:-2]).lower() + @staticmethod def format_seconds(seconds): + if seconds is None: + return ' Unknown' time = timetuple_from_msec(seconds * 1000) if time.hours > 99: return '--:--:--' - if not time.hours: - return '%02d:%02d' % time[1:-1] return '%02d:%02d:%02d' % time[:-1] + @classmethod + def format_eta(cls, seconds): + return f'{remove_start(cls.format_seconds(seconds), "00:"):>8s}' + @staticmethod def calc_percent(byte_counter, data_len): if data_len is None: @@ -94,11 +135,7 @@ class FileDownloader(object): @staticmethod def format_percent(percent): - if percent is None: - return '---.-%' - elif percent == 100: - return '100%' - return '%6s' % ('%3.1f%%' % percent) + return ' N/A%' if percent is None else f'{percent:>5.1f}%' @staticmethod def calc_eta(start, now, total, current): @@ -113,12 +150,6 @@ class FileDownloader(object): 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 if bytes == 0 or dif < 0.001: # One millisecond @@ -127,13 +158,11 @@ class FileDownloader(object): @staticmethod def format_speed(speed): - if speed is None: - return '%10s' % '---b/s' - return '%10s' % ('%s/s' % format_bytes(speed)) + return ' Unknown B/s' if speed is None else f'{format_bytes(speed):>10s}/s' @staticmethod def format_retries(retries): - return 'inf' if retries == float('inf') else '%.0f' % retries + return 'inf' if retries == float('inf') else int(retries) @staticmethod def best_block_size(elapsed_time, bytes): @@ -151,33 +180,7 @@ class FileDownloader(object): @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_screen(*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) - - def write_debug(self, *args, **kargs): - self.ydl.write_debug(*args, **kargs) + parse_bytes(bytestr) def slow_down(self, start_time, now, byte_counter): """Sleep if the download speed is over the rate limit.""" @@ -211,30 +214,31 @@ class FileDownloader(object): return filename + '.ytdl' def wrap_file_access(action, *, fatal=False): - def outer(func): - def inner(self, *args, **kwargs): - file_access_retries = self.params.get('file_access_retries', 0) - retry = 0 - while True: - try: - return func(self, *args, **kwargs) - except (IOError, OSError) as err: - retry = retry + 1 - if retry > file_access_retries or err.errno not in (errno.EACCES, errno.EINVAL): - if not fatal: - self.report_error(f'unable to {action} file: {err}') - return - raise - self.to_screen( - f'[download] Unable to {action} file due to file access error. ' - f'Retrying (attempt {retry} of {self.format_retries(file_access_retries)}) ...') - time.sleep(0.01) - return inner - return outer + def error_callback(err, count, retries, *, fd): + return RetryManager.report_retry( + err, count, retries, info=fd.__to_screen, + warn=lambda e: (time.sleep(0.01), fd.to_screen(f'[download] Unable to {action} file: {e}')), + error=None if fatal else lambda e: fd.report_error(f'Unable to {action} file: {e}'), + sleep_func=fd.params.get('retry_sleep_functions', {}).get('file_access')) + + def wrapper(self, func, *args, **kwargs): + for retry in RetryManager(self.params.get('file_access_retries'), error_callback, fd=self): + try: + return func(self, *args, **kwargs) + except OSError as err: + if err.errno in (errno.EACCES, errno.EINVAL): + retry.error = err + continue + retry.error_callback(err, 1, 0) + + return functools.partial(functools.partialmethod, wrapper) @wrap_file_access('open', fatal=True) def sanitize_open(self, filename, open_mode): - return sanitize_open(filename, open_mode) + f, filename = sanitize_open(filename, open_mode) + if not getattr(f, 'locked', None): + self.write_debug(f'{LockingUnsupportedError.msg}. Proceeding without locking', only_once=True) + return f, filename @wrap_file_access('remove') def try_remove(self, filename): @@ -261,10 +265,8 @@ class FileDownloader(object): # Ignore obviously invalid dates if filetime == 0: return - try: + with contextlib.suppress(Exception): os.utime(filename, (time.time(), filetime)) - except Exception: - pass return filetime def report_destination(self, filename): @@ -277,26 +279,26 @@ class FileDownloader(object): 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._out_files['screen'], lines) + self._multiline = BreaklineStatusPrinter(self.ydl._out_files.out, lines) else: - self._multiline = MultilinePrinter(self.ydl._out_files['screen'], lines, not self.params.get('quiet')) + self._multiline = MultilinePrinter(self.ydl._out_files.out, lines, not self.params.get('quiet')) self._multiline.allow_colors = self._multiline._HAVE_FULLCAP and not self.params.get('no_color') def _finish_multiline_status(self): self._multiline.end() - _progress_styles = { - 'downloaded_bytes': 'light blue', - 'percent': 'light blue', - 'eta': 'yellow', - 'speed': 'green', - 'elapsed': 'bold white', - 'total_bytes': '', - 'total_bytes_estimate': '', - } + ProgressStyles = Namespace( + downloaded_bytes='light blue', + percent='light blue', + eta='yellow', + speed='green', + elapsed='bold white', + total_bytes='', + total_bytes_estimate='', + ) def _report_progress_status(self, s, default_template): - for name, style in self._progress_styles.items(): + for name, style in self.ProgressStyles.items_: name = f'_{name}_str' if name not in s: continue @@ -320,78 +322,73 @@ class FileDownloader(object): self._multiline.stream, self._multiline.allow_colors, *args, **kwargs) def report_progress(self, s): + def with_fields(*tups, default=''): + for *fields, tmpl in tups: + if all(s.get(f) is not None for f in fields): + return tmpl + return default + + _format_bytes = lambda k: f'{format_bytes(s.get(k)):>10s}' + 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) - self._report_progress_status(s, msg_template) - return + speed = try_call(lambda: s['total_bytes'] / s['elapsed']) + s.update({ + 'speed': speed, + '_speed_str': self.format_speed(speed).strip(), + '_total_bytes_str': _format_bytes('total_bytes'), + '_elapsed_str': self.format_seconds(s.get('elapsed')), + '_percent_str': self.format_percent(100), + }) + self._report_progress_status(s, join_nonempty( + '100%%', + with_fields(('total_bytes', 'of %(_total_bytes_str)s')), + with_fields(('elapsed', 'in %(_elapsed_str)s')), + with_fields(('speed', 'at %(_speed_str)s')), + delim=' ')) 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' - - 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' - if s.get('fragment_index') and s.get('fragment_count'): - msg_template += ' (frag %(fragment_index)s/%(fragment_count)s)' - elif s.get('fragment_index'): - msg_template += ' (frag %(fragment_index)s)' + s.update({ + '_eta_str': self.format_eta(s.get('eta')).strip(), + '_speed_str': self.format_speed(s.get('speed')), + '_percent_str': self.format_percent(try_call( + lambda: 100 * s['downloaded_bytes'] / s['total_bytes'], + lambda: 100 * s['downloaded_bytes'] / s['total_bytes_estimate'], + lambda: s['downloaded_bytes'] == 0 and 0)), + '_total_bytes_str': _format_bytes('total_bytes'), + '_total_bytes_estimate_str': _format_bytes('total_bytes_estimate'), + '_downloaded_bytes_str': _format_bytes('downloaded_bytes'), + '_elapsed_str': self.format_seconds(s.get('elapsed')), + }) + + msg_template = with_fields( + ('total_bytes', '%(_percent_str)s of %(_total_bytes_str)s at %(_speed_str)s ETA %(_eta_str)s'), + ('total_bytes_estimate', '%(_percent_str)s of ~%(_total_bytes_estimate_str)s at %(_speed_str)s ETA %(_eta_str)s'), + ('downloaded_bytes', 'elapsed', '%(_downloaded_bytes_str)s at %(_speed_str)s (%(_elapsed_str)s)'), + ('downloaded_bytes', '%(_downloaded_bytes_str)s at %(_speed_str)s'), + default='%(_percent_str)s at %(_speed_str)s ETA %(_eta_str)s') + + msg_template += with_fields( + ('fragment_index', 'fragment_count', ' (frag %(fragment_index)s/%(fragment_count)s)'), + ('fragment_index', ' (frag %(fragment_index)s)')) self._report_progress_status(s, msg_template) def report_resuming_byte(self, resume_len): """Report attempt to resume at given byte.""" 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( - '[download] Got server HTTP error: %s. Retrying (attempt %d of %s) ...' - % (error_to_compat_str(err), count, self.format_retries(retries))) - - def report_file_already_downloaded(self, *args, **kwargs): - """Report file has already been fully downloaded.""" - return self.ydl.report_file_already_downloaded(*args, **kwargs) + def report_retry(self, err, count, retries, frag_index=NO_DEFAULT, fatal=True): + """Report retry""" + is_frag = False if frag_index is NO_DEFAULT else 'fragment' + RetryManager.report_retry( + err, count, retries, info=self.__to_screen, + warn=lambda msg: self.__to_screen(f'[download] Got error: {msg}'), + error=IDENTITY if not fatal else lambda e: self.report_error(f'\r[download] Got error: {e}'), + sleep_func=self.params.get('retry_sleep_functions', {}).get(is_frag or 'http'), + suffix=f'fragment{"s" if frag_index is None else f" {frag_index}"}' if is_frag else None) def report_unable_to_resume(self): """Report it was impossible to resume download.""" @@ -431,25 +428,16 @@ class FileDownloader(object): self._finish_multiline_status() return True, False - 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( - '[download] Sleeping %s seconds ...' % ( - int(sleep_interval) if sleep_interval.is_integer() - else '%.2f' % sleep_interval)) - time.sleep(sleep_interval) + if subtitle: + sleep_interval = self.params.get('sleep_interval_subtitles') or 0 else: - sleep_interval_sub = 0 - if type(self.params.get('sleep_interval_subtitles')) is int: - sleep_interval_sub = self.params.get('sleep_interval_subtitles') - if sleep_interval_sub > 0: - self.to_screen( - '[download] Sleeping %s seconds ...' % ( - sleep_interval_sub)) - time.sleep(sleep_interval_sub) + min_sleep_interval = self.params.get('sleep_interval') or 0 + sleep_interval = random.uniform( + min_sleep_interval, self.params.get('max_sleep_interval') or min_sleep_interval) + if sleep_interval > 0: + self.to_screen(f'[download] Sleeping {sleep_interval:.2f} seconds ...') + time.sleep(sleep_interval) + ret = self.real_download(filename, info_dict) self._finish_multiline_status() return ret, True @@ -459,8 +447,7 @@ class FileDownloader(object): raise NotImplementedError('This method must be implemented by subclasses') def _hook_progress(self, status, info_dict): - if not self._progress_hooks: - return + # Ideally we want to make a copy of the dict, but that is too slow status['info_dict'] = info_dict # youtube-dl passes the same status object to all the hooks. # Some third party scripts seems to be relying on this. @@ -482,4 +469,4 @@ class FileDownloader(object): if exe is None: exe = os.path.basename(str_args[0]) - self.write_debug('%s command line: %s' % (exe, shell_quote(str_args))) + self.write_debug(f'{exe} command line: {shell_quote(str_args)}') diff --git a/hypervideo_dl/downloader/dash.py b/hypervideo_dl/downloader/dash.py index a845ee7..4328d73 100644 --- a/hypervideo_dl/downloader/dash.py +++ b/hypervideo_dl/downloader/dash.py @@ -1,10 +1,9 @@ -from __future__ import unicode_literals import time +import urllib.parse -from ..downloader import get_suitable_downloader +from . import get_suitable_downloader from .fragment import FragmentFD - -from ..utils import urljoin +from ..utils import update_url_query, urljoin class DashSegmentsFD(FragmentFD): @@ -42,24 +41,29 @@ class DashSegmentsFD(FragmentFD): self._prepare_and_start_frag_download(ctx, fmt) ctx['start'] = real_start - fragments_to_download = self._get_fragments(fmt, ctx) + extra_query = None + extra_param_to_segment_url = info_dict.get('extra_param_to_segment_url') + if extra_param_to_segment_url: + extra_query = urllib.parse.parse_qs(extra_param_to_segment_url) + + fragments_to_download = self._get_fragments(fmt, ctx, extra_query) if real_downloader: self.to_screen( - '[%s] Fragment downloads will be delegated to %s' % (self.FD_NAME, real_downloader.get_basename())) + f'[{self.FD_NAME}] Fragment downloads will be delegated to {real_downloader.get_basename()}') info_dict['fragments'] = list(fragments_to_download) fd = real_downloader(self.ydl, self.params) return fd.real_download(filename, info_dict) args.append([ctx, fragments_to_download, fmt]) - return self.download_and_append_fragments_multiple(*args) + return self.download_and_append_fragments_multiple(*args, is_fatal=lambda idx: idx == 0) def _resolve_fragments(self, fragments, ctx): fragments = fragments(ctx) if callable(fragments) else fragments return [next(iter(fragments))] if self.params.get('test') else fragments - def _get_fragments(self, fmt, ctx): + def _get_fragments(self, fmt, ctx, extra_query): fragment_base_url = fmt.get('fragment_base_url') fragments = self._resolve_fragments(fmt['fragments'], ctx) @@ -72,9 +76,12 @@ class DashSegmentsFD(FragmentFD): if not fragment_url: assert fragment_base_url fragment_url = urljoin(fragment_base_url, fragment['path']) + if extra_query: + fragment_url = update_url_query(fragment_url, extra_query) yield { 'frag_index': frag_index, + 'fragment_count': fragment.get('fragment_count'), 'index': i, 'url': fragment_url, } diff --git a/hypervideo_dl/downloader/external.py b/hypervideo_dl/downloader/external.py index b99dc37..75257a7 100644 --- a/hypervideo_dl/downloader/external.py +++ b/hypervideo_dl/downloader/external.py @@ -1,5 +1,4 @@ -from __future__ import unicode_literals - +import enum import os.path import re import subprocess @@ -7,30 +6,35 @@ import sys import time from .fragment import FragmentFD -from ..compat import ( - compat_setenv, - compat_str, -) -from ..postprocessor.ffmpeg import FFmpegPostProcessor, EXT_TO_OUT_FORMATS +from ..compat import functools +from ..postprocessor.ffmpeg import EXT_TO_OUT_FORMATS, FFmpegPostProcessor from ..utils import ( + Popen, + RetryManager, + _configuration_args, + check_executable, classproperty, + cli_bool_option, cli_option, cli_valueless_option, - cli_bool_option, - _configuration_args, determine_ext, - encodeFilename, encodeArgument, + encodeFilename, handle_youtubedl_headers, - check_executable, - Popen, remove_end, + traverse_obj, ) +class Features(enum.Enum): + TO_STDOUT = enum.auto() + MULTIPLE_FORMATS = enum.auto() + + class ExternalFD(FragmentFD): SUPPORTED_PROTOCOLS = ('http', 'https', 'ftp', 'ftps') - can_download_to_stdout = False + SUPPORTED_FEATURES = () + _CAPTURE_STDERR = True def real_download(self, filename, info_dict): self.report_destination(filename) @@ -56,7 +60,7 @@ class ExternalFD(FragmentFD): } if filename != '-': fsize = os.path.getsize(encodeFilename(tmpfilename)) - self.to_screen('\r[%s] Downloaded %s bytes' % (self.get_basename(), fsize)) + self.to_screen(f'\r[{self.get_basename()}] Downloaded {fsize} bytes') self.try_rename(tmpfilename, filename) status.update({ 'downloaded_bytes': fsize, @@ -78,7 +82,7 @@ class ExternalFD(FragmentFD): def EXE_NAME(cls): return cls.get_basename() - @property + @functools.cached_property def exe(self): return self.EXE_NAME @@ -94,9 +98,11 @@ class ExternalFD(FragmentFD): @classmethod def supports(cls, info_dict): - return ( - (cls.can_download_to_stdout or not info_dict.get('to_stdout')) - and info_dict['protocol'] in cls.SUPPORTED_PROTOCOLS) + return all(( + not info_dict.get('to_stdout') or Features.TO_STDOUT in cls.SUPPORTED_FEATURES, + '+' not in info_dict['protocol'] or Features.MULTIPLE_FORMATS in cls.SUPPORTED_FEATURES, + all(proto in cls.SUPPORTED_PROTOCOLS for proto in info_dict['protocol'].split('+')), + )) @classmethod def can_download(cls, info_dict, path=None): @@ -123,33 +129,28 @@ class ExternalFD(FragmentFD): self._debug_cmd(cmd) if 'fragments' not in info_dict: - p = Popen(cmd, stderr=subprocess.PIPE) - _, stderr = p.communicate_or_kill() - if p.returncode != 0: - self.to_stderr(stderr.decode('utf-8', 'replace')) - return p.returncode + _, stderr, returncode = Popen.run( + cmd, text=True, stderr=subprocess.PIPE if self._CAPTURE_STDERR else None) + if returncode and stderr: + self.to_stderr(stderr) + return returncode - fragment_retries = self.params.get('fragment_retries', 0) skip_unavailable_fragments = self.params.get('skip_unavailable_fragments', True) - count = 0 - while count <= fragment_retries: - p = Popen(cmd, stderr=subprocess.PIPE) - _, stderr = p.communicate_or_kill() - if p.returncode == 0: + retry_manager = RetryManager(self.params.get('fragment_retries'), self.report_retry, + frag_index=None, fatal=not skip_unavailable_fragments) + for retry in retry_manager: + _, stderr, returncode = Popen.run(cmd, text=True, stderr=subprocess.PIPE) + if not returncode: break # TODO: Decide whether to retry based on error code # https://aria2.github.io/manual/en/html/aria2c.html#exit-status - self.to_stderr(stderr.decode('utf-8', 'replace')) - count += 1 - if count <= fragment_retries: - self.to_screen( - '[%s] Got error. Retrying fragments (attempt %d of %s)...' - % (self.get_basename(), count, self.format_retries(fragment_retries))) - if count > fragment_retries: - if not skip_unavailable_fragments: - self.report_error('Giving up after %s fragment retries' % fragment_retries) - return -1 + if stderr: + self.to_stderr(stderr) + retry.error = Exception() + continue + if not skip_unavailable_fragments and retry_manager.error: + return -1 decrypt_fragment = self.decrypter(info_dict) dest, _ = self.sanitize_open(tmpfilename, 'wb') @@ -157,7 +158,7 @@ class ExternalFD(FragmentFD): fragment_filename = '%s-Frag%d' % (tmpfilename, frag_index) try: src, _ = self.sanitize_open(fragment_filename, 'rb') - except IOError as err: + except OSError as err: if skip_unavailable_fragments and frag_index > 1: self.report_skip_fragment(frag_index, err) continue @@ -174,12 +175,13 @@ class ExternalFD(FragmentFD): class CurlFD(ExternalFD): AVAILABLE_OPT = '-V' + _CAPTURE_STDERR = False # curl writes the progress to stderr def _make_cmd(self, tmpfilename, info_dict): cmd = [self.exe, '--location', '-o', tmpfilename, '--compressed'] if info_dict.get('http_headers') is not None: for key, val in info_dict['http_headers'].items(): - cmd += ['--header', '%s: %s' % (key, val)] + cmd += ['--header', f'{key}: {val}'] cmd += self._bool_option('--continue-at', 'continuedl', '-', '0') cmd += self._valueless_option('--silent', 'noprogress') @@ -198,16 +200,6 @@ class CurlFD(ExternalFD): cmd += ['--', info_dict['url']] return cmd - def _call_downloader(self, tmpfilename, info_dict): - cmd = [encodeArgument(a) for a in self._make_cmd(tmpfilename, info_dict)] - - self._debug_cmd(cmd) - - # curl writes the progress to stderr so don't capture it. - p = Popen(cmd) - p.communicate_or_kill() - return p.returncode - class AxelFD(ExternalFD): AVAILABLE_OPT = '-V' @@ -216,7 +208,7 @@ class AxelFD(ExternalFD): cmd = [self.exe, '-o', tmpfilename] if info_dict.get('http_headers') is not None: for key, val in info_dict['http_headers'].items(): - cmd += ['-H', '%s: %s' % (key, val)] + cmd += ['-H', f'{key}: {val}'] cmd += self._configuration_args() cmd += ['--', info_dict['url']] return cmd @@ -229,7 +221,7 @@ class WgetFD(ExternalFD): cmd = [self.exe, '-O', tmpfilename, '-nv', '--no-cookies', '--compression=auto'] if info_dict.get('http_headers') is not None: for key, val in info_dict['http_headers'].items(): - cmd += ['--header', '%s: %s' % (key, val)] + cmd += ['--header', f'{key}: {val}'] cmd += self._option('--limit-rate', 'ratelimit') retry = self._option('--tries', 'retries') if len(retry) == 2: @@ -240,7 +232,7 @@ class WgetFD(ExternalFD): proxy = self.params.get('proxy') if proxy: for var in ('http_proxy', 'https_proxy'): - cmd += ['--execute', '%s=%s' % (var, proxy)] + cmd += ['--execute', f'{var}={proxy}'] cmd += self._valueless_option('--no-check-certificate', 'nocheckcertificate') cmd += self._configuration_args() cmd += ['--', info_dict['url']] @@ -260,6 +252,10 @@ class Aria2cFD(ExternalFD): check_results = (not re.search(feature, manifest) for feature in UNSUPPORTED_FEATURES) return all(check_results) + @staticmethod + def _aria2c_filename(fn): + return fn if os.path.isabs(fn) else f'.{os.path.sep}{fn}' + def _make_cmd(self, tmpfilename, info_dict): cmd = [self.exe, '-c', '--console-log-level=warn', '--summary-interval=0', '--download-result=hide', @@ -271,7 +267,7 @@ class Aria2cFD(ExternalFD): if info_dict.get('http_headers') is not None: for key, val in info_dict['http_headers'].items(): - cmd += ['--header', '%s: %s' % (key, val)] + cmd += ['--header', f'{key}: {val}'] cmd += self._option('--max-overall-download-limit', 'ratelimit') cmd += self._option('--interface', 'source_address') cmd += self._option('--all-proxy', 'proxy') @@ -288,11 +284,9 @@ class Aria2cFD(ExternalFD): # https://github.com/aria2/aria2/issues/1373 dn = os.path.dirname(tmpfilename) if dn: - if not os.path.isabs(dn): - dn = '.%s%s' % (os.path.sep, dn) - cmd += ['--dir', dn + os.path.sep] + cmd += ['--dir', self._aria2c_filename(dn) + os.path.sep] if 'fragments' not in info_dict: - cmd += ['--out', '.%s%s' % (os.path.sep, os.path.basename(tmpfilename))] + cmd += ['--out', self._aria2c_filename(os.path.basename(tmpfilename))] cmd += ['--auto-file-renaming=false'] if 'fragments' in info_dict: @@ -301,11 +295,11 @@ class Aria2cFD(ExternalFD): url_list = [] for frag_index, fragment in enumerate(info_dict['fragments']): fragment_filename = '%s-Frag%d' % (os.path.basename(tmpfilename), frag_index) - url_list.append('%s\n\tout=%s' % (fragment['url'], fragment_filename)) + url_list.append('%s\n\tout=%s' % (fragment['url'], self._aria2c_filename(fragment_filename))) stream, _ = self.sanitize_open(url_list_file, 'wb') - stream.write('\n'.join(url_list).encode('utf-8')) + stream.write('\n'.join(url_list).encode()) stream.close() - cmd += ['-i', url_list_file] + cmd += ['-i', self._aria2c_filename(url_list_file)] else: cmd += ['--', info_dict['url']] return cmd @@ -320,13 +314,13 @@ class HttpieFD(ExternalFD): if info_dict.get('http_headers') is not None: for key, val in info_dict['http_headers'].items(): - cmd += ['%s:%s' % (key, val)] + cmd += [f'{key}:{val}'] return cmd class FFmpegFD(ExternalFD): SUPPORTED_PROTOCOLS = ('http', 'https', 'ftp', 'ftps', 'm3u8', 'm3u8_native', 'rtsp', 'rtmp', 'rtmp_ffmpeg', 'mms', 'http_dash_segments') - can_download_to_stdout = True + SUPPORTED_FEATURES = (Features.TO_STDOUT, Features.MULTIPLE_FORMATS) @classmethod def available(cls, path=None): @@ -334,10 +328,6 @@ class FFmpegFD(ExternalFD): # Fixme: This may be wrong when --ffmpeg-location is used return FFmpegPostProcessor().available - @classmethod - def supports(cls, info_dict): - return all(proto in cls.SUPPORTED_PROTOCOLS for proto in info_dict['protocol'].split('+')) - def on_process_started(self, proc, stdin): """ Override this in subclasses """ pass @@ -368,9 +358,11 @@ class FFmpegFD(ExternalFD): if not self.params.get('verbose'): args += ['-hide_banner'] - args += info_dict.get('_ffmpeg_args', []) + args += traverse_obj(info_dict, ('downloader_options', 'ffmpeg_args'), default=[]) - # This option exists only for compatibility. Extractors should use `_ffmpeg_args` instead + # These exists only for compatibility. Extractors should use + # info_dict['downloader_options']['ffmpeg_args'] instead + args += info_dict.get('_ffmpeg_args') or [] seekable = info_dict.get('_seekable') if seekable is not None: # setting -seekable prevents ffmpeg from guessing if the server @@ -380,20 +372,15 @@ class FFmpegFD(ExternalFD): # http://trac.ffmpeg.org/ticket/6125#comment:10 args += ['-seekable', '1' if seekable else '0'] - # start_time = info_dict.get('start_time') or 0 - # if start_time: - # args += ['-ss', compat_str(start_time)] - # end_time = info_dict.get('end_time') - # if end_time: - # args += ['-t', compat_str(end_time - start_time)] - - if info_dict.get('http_headers') is not None and re.match(r'^https?://', urls[0]): - # Trailing \r\n after each HTTP header is important to prevent warning from ffmpeg/avconv: - # [http @ 00000000003d2fa0] No trailing CRLF found in HTTP header. - headers = handle_youtubedl_headers(info_dict['http_headers']) - args += [ + http_headers = None + if info_dict.get('http_headers'): + youtubedl_headers = handle_youtubedl_headers(info_dict['http_headers']) + http_headers = [ + # Trailing \r\n after each HTTP header is important to prevent warning from ffmpeg/avconv: + # [http @ 00000000003d2fa0] No trailing CRLF found in HTTP header. '-headers', - ''.join('%s: %s\r\n' % (key, val) for key, val in headers.items())] + ''.join(f'{key}: {val}\r\n' for key, val in youtubedl_headers.items()) + ] env = None proxy = self.params.get('proxy') @@ -411,8 +398,8 @@ class FFmpegFD(ExternalFD): # We could switch to the following code if we are able to detect version properly # args += ['-http_proxy', proxy] env = os.environ.copy() - compat_setenv('HTTP_PROXY', proxy, env=env) - compat_setenv('http_proxy', proxy, env=env) + env['HTTP_PROXY'] = proxy + env['http_proxy'] = proxy protocol = info_dict.get('protocol') @@ -442,20 +429,31 @@ class FFmpegFD(ExternalFD): if isinstance(conn, list): for entry in conn: args += ['-rtmp_conn', entry] - elif isinstance(conn, compat_str): + elif isinstance(conn, str): args += ['-rtmp_conn', conn] + start_time, end_time = info_dict.get('section_start') or 0, info_dict.get('section_end') + for i, url in enumerate(urls): + if http_headers is not None and re.match(r'^https?://', url): + args += http_headers + if start_time: + args += ['-ss', str(start_time)] + if end_time: + args += ['-t', str(end_time - start_time)] + args += self._configuration_args((f'_i{i + 1}', '_i')) + ['-i', url] - args += ['-c', 'copy'] + if not (start_time or end_time) or not self.params.get('force_keyframes_at_cuts'): + args += ['-c', 'copy'] + if info_dict.get('requested_formats') or protocol == 'http_dash_segments': for (i, fmt) in enumerate(info_dict.get('requested_formats') or [info_dict]): stream_number = fmt.get('manifest_stream_number', 0) args.extend(['-map', f'{i}:{stream_number}']) if self.params.get('test', False): - args += ['-fs', compat_str(self._TEST_FILE_SIZE)] + args += ['-fs', str(self._TEST_FILE_SIZE)] ext = info_dict['ext'] if protocol in ('m3u8', 'm3u8_native'): @@ -490,24 +488,23 @@ class FFmpegFD(ExternalFD): args.append(encodeFilename(ffpp._ffmpeg_filename_argument(tmpfilename), True)) self._debug_cmd(args) - proc = Popen(args, stdin=subprocess.PIPE, env=env) - if url in ('-', 'pipe:'): - self.on_process_started(proc, proc.stdin) - try: - retval = proc.wait() - except BaseException as e: - # subprocces.run would send the SIGKILL signal to ffmpeg and the - # mp4 file couldn't be played, but if we ask ffmpeg to quit it - # produces a file that is playable (this is mostly useful for live - # streams). Note that Windows is not affected and produces playable - # files (see https://github.com/ytdl-org/youtube-dl/issues/8300). - if isinstance(e, KeyboardInterrupt) and sys.platform != 'win32' and url not in ('-', 'pipe:'): - proc.communicate_or_kill(b'q') - else: - proc.kill() - proc.wait() - raise - return retval + with Popen(args, stdin=subprocess.PIPE, env=env) as proc: + if url in ('-', 'pipe:'): + self.on_process_started(proc, proc.stdin) + try: + retval = proc.wait() + except BaseException as e: + # subprocces.run would send the SIGKILL signal to ffmpeg and the + # mp4 file couldn't be played, but if we ask ffmpeg to quit it + # produces a file that is playable (this is mostly useful for live + # streams). Note that Windows is not affected and produces playable + # files (see https://github.com/ytdl-org/youtube-dl/issues/8300). + if isinstance(e, KeyboardInterrupt) and sys.platform != 'win32' and url not in ('-', 'pipe:'): + proc.communicate_or_kill(b'q') + else: + proc.kill(timeout=None) + raise + return retval class AVconvFD(FFmpegFD): @@ -520,16 +517,14 @@ _BY_NAME = { if name.endswith('FD') and name not in ('ExternalFD', 'FragmentFD') } -_BY_EXE = {klass.EXE_NAME: klass for klass in _BY_NAME.values()} - def list_external_downloaders(): return sorted(_BY_NAME.keys()) def get_external_downloader(external_downloader): - """ Given the name of the executable, see whether we support the given - downloader . """ - # Drop .exe extension on Windows + """ Given the name of the executable, see whether we support the given downloader """ bn = os.path.splitext(os.path.basename(external_downloader))[0] - return _BY_NAME.get(bn, _BY_EXE.get(bn)) + return _BY_NAME.get(bn) or next(( + klass for klass in _BY_NAME.values() if klass.EXE_NAME in bn + ), None) diff --git a/hypervideo_dl/downloader/f4m.py b/hypervideo_dl/downloader/f4m.py index 0008b7c..306f921 100644 --- a/hypervideo_dl/downloader/f4m.py +++ b/hypervideo_dl/downloader/f4m.py @@ -1,23 +1,14 @@ -from __future__ import division, unicode_literals - +import base64 import io import itertools +import struct import time +import urllib.error +import urllib.parse from .fragment import FragmentFD -from ..compat import ( - compat_b64decode, - compat_etree_fromstring, - compat_urlparse, - compat_urllib_error, - compat_urllib_parse_urlparse, - compat_struct_pack, - compat_struct_unpack, -) -from ..utils import ( - fix_xml_ampersands, - xpath_text, -) +from ..compat import compat_etree_fromstring +from ..utils import fix_xml_ampersands, xpath_text class DataTruncatedError(Exception): @@ -40,13 +31,13 @@ class FlvReader(io.BytesIO): # Utility functions for reading numbers and strings def read_unsigned_long_long(self): - return compat_struct_unpack('!Q', self.read_bytes(8))[0] + return struct.unpack('!Q', self.read_bytes(8))[0] def read_unsigned_int(self): - return compat_struct_unpack('!I', self.read_bytes(4))[0] + return struct.unpack('!I', self.read_bytes(4))[0] def read_unsigned_char(self): - return compat_struct_unpack('!B', self.read_bytes(1))[0] + return struct.unpack('!B', self.read_bytes(1))[0] def read_string(self): res = b'' @@ -193,7 +184,7 @@ def build_fragments_list(boot_info): first_frag_number = fragment_run_entry_table[0]['first'] fragments_counter = itertools.count(first_frag_number) for segment, fragments_count in segment_run_table['segment_run']: - # In some live HDS streams (for example Rai), `fragments_count` is + # In some live HDS streams (e.g. Rai), `fragments_count` is # abnormal and causing out-of-memory errors. It's OK to change the # number of fragments for live streams as they are updated periodically if fragments_count == 4294967295 and boot_info['live']: @@ -208,11 +199,11 @@ def build_fragments_list(boot_info): def write_unsigned_int(stream, val): - stream.write(compat_struct_pack('!I', val)) + stream.write(struct.pack('!I', val)) def write_unsigned_int_24(stream, val): - stream.write(compat_struct_pack('!I', val)[1:]) + stream.write(struct.pack('!I', val)[1:]) def write_flv_header(stream): @@ -261,8 +252,6 @@ class F4mFD(FragmentFD): A downloader for f4m manifests or AdobeHDS. """ - FD_NAME = 'f4m' - def _get_unencrypted_media(self, doc): media = doc.findall(_add_ns('media')) if not media: @@ -308,12 +297,12 @@ class F4mFD(FragmentFD): # 1. http://live-1-1.rutube.ru/stream/1024/HDS/SD/C2NKsS85HQNckgn5HdEmOQ/1454167650/S-s604419906/move/four/dirs/upper/1024-576p.f4m bootstrap_url = node.get('url') if bootstrap_url: - bootstrap_url = compat_urlparse.urljoin( + bootstrap_url = urllib.parse.urljoin( base_url, bootstrap_url) boot_info = self._get_bootstrap_from_url(bootstrap_url) else: bootstrap_url = None - bootstrap = compat_b64decode(node.text) + bootstrap = base64.b64decode(node.text) boot_info = read_bootstrap_info(bootstrap) return boot_info, bootstrap_url @@ -343,14 +332,14 @@ class F4mFD(FragmentFD): # Prefer baseURL for relative URLs as per 11.2 of F4M 3.0 spec. man_base_url = get_base_url(doc) or man_url - base_url = compat_urlparse.urljoin(man_base_url, media.attrib['url']) + base_url = urllib.parse.urljoin(man_base_url, media.attrib['url']) bootstrap_node = doc.find(_add_ns('bootstrapInfo')) boot_info, bootstrap_url = self._parse_bootstrap_node( bootstrap_node, man_base_url) live = boot_info['live'] metadata_node = media.find(_add_ns('metadata')) if metadata_node is not None: - metadata = compat_b64decode(metadata_node.text) + metadata = base64.b64decode(metadata_node.text) else: metadata = None @@ -378,7 +367,7 @@ class F4mFD(FragmentFD): if not live: write_metadata_tag(dest_stream, metadata) - base_url_parsed = compat_urllib_parse_urlparse(base_url) + base_url_parsed = urllib.parse.urlparse(base_url) self._start_frag_download(ctx, info_dict) @@ -398,9 +387,10 @@ class F4mFD(FragmentFD): query.append(info_dict['extra_param_to_segment_url']) url_parsed = base_url_parsed._replace(path=base_url_parsed.path + name, query='&'.join(query)) try: - success, down_data = self._download_fragment(ctx, url_parsed.geturl(), info_dict) + success = self._download_fragment(ctx, url_parsed.geturl(), info_dict) if not success: return False + down_data = self._read_fragment(ctx) reader = FlvReader(down_data) while True: try: @@ -417,7 +407,7 @@ class F4mFD(FragmentFD): if box_type == b'mdat': self._append_fragment(ctx, box_data) break - except (compat_urllib_error.HTTPError, ) as err: + except urllib.error.HTTPError as err: if live and (err.code == 404 or err.code == 410): # We didn't keep up with the live window. Continue # with the next available fragment. @@ -434,6 +424,4 @@ class F4mFD(FragmentFD): msg = 'Missed %d fragments' % (fragments_list[0][1] - (frag_i + 1)) self.report_warning(msg) - self._finish_frag_download(ctx, info_dict) - - return True + return self._finish_frag_download(ctx, info_dict) diff --git a/hypervideo_dl/downloader/fc2.py b/hypervideo_dl/downloader/fc2.py index 157bcf2..f9763de 100644 --- a/hypervideo_dl/downloader/fc2.py +++ b/hypervideo_dl/downloader/fc2.py @@ -1,5 +1,3 @@ -from __future__ import division, unicode_literals - import threading from .common import FileDownloader @@ -20,6 +18,9 @@ class FC2LiveFD(FileDownloader): heartbeat_state = [None, 1] def heartbeat(): + if heartbeat_state[1] < 0: + return + try: heartbeat_state[1] += 1 ws.send('{"name":"heartbeat","arguments":{},"id":%d}' % heartbeat_state[1]) @@ -38,4 +39,8 @@ class FC2LiveFD(FileDownloader): 'ws': None, 'protocol': 'live_ffmpeg', }) - return FFmpegFD(self.ydl, self.params or {}).download(filename, new_info_dict) + try: + return FFmpegFD(self.ydl, self.params or {}).download(filename, new_info_dict) + finally: + # stop heartbeating + heartbeat_state[1] = -1 diff --git a/hypervideo_dl/downloader/fragment.py b/hypervideo_dl/downloader/fragment.py index a991c6d..e61bd0e 100644 --- a/hypervideo_dl/downloader/fragment.py +++ b/hypervideo_dl/downloader/fragment.py @@ -1,28 +1,20 @@ -from __future__ import division, unicode_literals - +import concurrent.futures +import contextlib import http.client import json import math import os +import struct import time - -try: - import concurrent.futures - can_threaded_download = True -except ImportError: - can_threaded_download = False +import urllib.error from .common import FileDownloader from .http import HttpFD from ..aes import aes_cbc_decrypt_bytes, unpad_pkcs7 -from ..compat import ( - compat_os_name, - compat_urllib_error, - compat_struct_pack, -) +from ..compat import compat_os_name from ..utils import ( DownloadError, - error_to_compat_str, + RetryManager, encodeFilename, sanitized_Request, traverse_obj, @@ -33,9 +25,7 @@ class HttpQuietDownloader(HttpFD): def to_screen(self, *args, **kargs): pass - def report_retry(self, err, count, retries): - super().to_screen( - f'[download] Got server HTTP error: {err}. Retrying (attempt {count} of {self.format_retries(retries)}) ...') + to_console_title = to_screen class FragmentFD(FileDownloader): @@ -75,9 +65,9 @@ class FragmentFD(FileDownloader): """ def report_retry_fragment(self, err, frag_index, count, retries): - self.to_screen( - '\r[download] Got server HTTP error: %s. Retrying fragment %d (attempt %d of %s) ...' - % (error_to_compat_str(err), frag_index, count, self.format_retries(retries))) + self.deprecation_warning('hypervideo_dl.downloader.FragmentFD.report_retry_fragment is deprecated. ' + 'Use hypervideo_dl.downloader.FileDownloader.report_retry instead') + return self.report_retry(err, count, retries, frag_index) def report_skip_fragment(self, frag_index, err=None): err = f' {err};' if err else '' @@ -131,7 +121,7 @@ class FragmentFD(FileDownloader): 'request_data': request_data, 'ctx_id': ctx.get('ctx_id'), } - success = ctx['dl'].download(fragment_filename, fragment_info_dict) + success, _ = ctx['dl'].download(fragment_filename, fragment_info_dict) if not success: return False if fragment_info_dict.get('filetime'): @@ -140,6 +130,8 @@ class FragmentFD(FileDownloader): return True def _read_fragment(self, ctx): + if not ctx.get('fragment_filename_sanitized'): + return None try: down, frag_sanitized = self.sanitize_open(ctx['fragment_filename_sanitized'], 'rb') except FileNotFoundError: @@ -172,21 +164,13 @@ class FragmentFD(FileDownloader): total_frags_str += ' (not including %d ad)' % ad_frags else: total_frags_str = 'unknown (live)' - self.to_screen( - '[%s] Total fragments: %s' % (self.FD_NAME, total_frags_str)) + self.to_screen(f'[{self.FD_NAME}] Total fragments: {total_frags_str}') self.report_destination(ctx['filename']) - dl = HttpQuietDownloader( - self.ydl, - { - 'continuedl': self.params.get('continuedl', True), - 'quiet': self.params.get('quiet'), - 'noprogress': True, - 'ratelimit': self.params.get('ratelimit'), - 'retries': self.params.get('retries', 0), - 'nopart': self.params.get('nopart', False), - 'test': self.params.get('test', False), - } - ) + dl = HttpQuietDownloader(self.ydl, { + **self.params, + 'noprogress': True, + 'test': False, + }) tmpfilename = self.temp_name(ctx['filename']) open_mode = 'wb' resume_len = 0 @@ -259,6 +243,9 @@ class FragmentFD(FileDownloader): if s['status'] not in ('downloading', 'finished'): return + if not total_frags and ctx.get('fragment_count'): + state['fragment_count'] = ctx['fragment_count'] + if ctx_id is not None and s.get('ctx_id') != ctx_id: return @@ -308,18 +295,23 @@ class FragmentFD(FileDownloader): self.try_remove(ytdl_filename) elapsed = time.time() - ctx['started'] - if ctx['tmpfilename'] == '-': - downloaded_bytes = ctx['complete_frags_downloaded_bytes'] + to_file = ctx['tmpfilename'] != '-' + if to_file: + downloaded_bytes = os.path.getsize(encodeFilename(ctx['tmpfilename'])) else: + downloaded_bytes = ctx['complete_frags_downloaded_bytes'] + + if not downloaded_bytes: + if to_file: + self.try_remove(ctx['tmpfilename']) + self.report_error('The downloaded file is empty') + return False + elif to_file: self.try_rename(ctx['tmpfilename'], ctx['filename']) - if self.params.get('updatetime', True): - filetime = ctx.get('fragment_filetime') - if filetime: - try: - os.utime(ctx['filename'], (time.time(), filetime)) - except Exception: - pass - downloaded_bytes = os.path.getsize(encodeFilename(ctx['filename'])) + filetime = ctx.get('fragment_filetime') + if self.params.get('updatetime', True) and filetime: + with contextlib.suppress(Exception): + os.utime(ctx['filename'], (time.time(), filetime)) self._hook_progress({ 'downloaded_bytes': downloaded_bytes, @@ -331,6 +323,7 @@ class FragmentFD(FileDownloader): 'max_progress': ctx.get('max_progress'), 'progress_idx': ctx.get('progress_idx'), }, info_dict) + return True def _prepare_external_frag_download(self, ctx): if 'live' not in ctx: @@ -342,8 +335,7 @@ class FragmentFD(FileDownloader): total_frags_str += ' (not including %d ad)' % ad_frags else: total_frags_str = 'unknown (live)' - self.to_screen( - '[%s] Total fragments: %s' % (self.FD_NAME, total_frags_str)) + self.to_screen(f'[{self.FD_NAME}] Total fragments: {total_frags_str}') tmpfilename = self.temp_name(ctx['filename']) @@ -362,10 +354,12 @@ class FragmentFD(FileDownloader): return _key_cache[url] def decrypt_fragment(fragment, frag_content): + if frag_content is None: + return decrypt_info = fragment.get('decrypt_info') if not decrypt_info or decrypt_info['METHOD'] != 'AES-128': return frag_content - iv = decrypt_info.get('IV') or compat_struct_pack('>8xq', fragment['media_sequence']) + iv = decrypt_info.get('IV') or struct.pack('>8xq', fragment['media_sequence']) decrypt_info['KEY'] = decrypt_info.get('KEY') or _get_key(info_dict.get('_decryption_key_url') or decrypt_info['URI']) # Don't decrypt the content in tests since the data is explicitly truncated and it's not to a valid block # size (see https://github.com/ytdl-org/youtube-dl/pull/27660). Tests only care that the correct data downloaded, @@ -376,7 +370,7 @@ class FragmentFD(FileDownloader): return decrypt_fragment - def download_and_append_fragments_multiple(self, *args, pack_func=None, finish_func=None): + def download_and_append_fragments_multiple(self, *args, **kwargs): ''' @params (ctx1, fragments1, info_dict1), (ctx2, fragments2, info_dict2), ... all args must be either tuple or list @@ -384,7 +378,7 @@ class FragmentFD(FileDownloader): interrupt_trigger = [True] max_progress = len(args) if max_progress == 1: - return self.download_and_append_fragments(*args[0], pack_func=pack_func, finish_func=finish_func) + return self.download_and_append_fragments(*args[0], **kwargs) max_workers = self.params.get('concurrent_fragment_downloads', 1) if max_progress > 1: self._prepare_multiline_status(max_progress) @@ -394,8 +388,7 @@ class FragmentFD(FileDownloader): ctx['max_progress'] = max_progress ctx['progress_idx'] = idx return self.download_and_append_fragments( - ctx, fragments, info_dict, pack_func=pack_func, finish_func=finish_func, - tpe=tpe, interrupt_trigger=interrupt_trigger) + ctx, fragments, info_dict, **kwargs, tpe=tpe, interrupt_trigger=interrupt_trigger) class FTPE(concurrent.futures.ThreadPoolExecutor): # has to stop this or it's going to wait on the worker thread itself @@ -442,18 +435,12 @@ class FragmentFD(FileDownloader): return result def download_and_append_fragments( - self, ctx, fragments, info_dict, *, pack_func=None, finish_func=None, - tpe=None, interrupt_trigger=None): - if not interrupt_trigger: - interrupt_trigger = (True, ) - - fragment_retries = self.params.get('fragment_retries', 0) - is_fatal = ( - ((lambda _: False) if info_dict.get('is_live') else (lambda idx: idx == 0)) - if self.params.get('skip_unavailable_fragments', True) else (lambda _: True)) + self, ctx, fragments, info_dict, *, is_fatal=(lambda idx: False), + pack_func=(lambda content, idx: content), finish_func=None, + tpe=None, interrupt_trigger=(True, )): - if not pack_func: - pack_func = lambda frag_content, _: frag_content + if not self.params.get('skip_unavailable_fragments', True): + is_fatal = lambda _: True def download_fragment(fragment, ctx): if not interrupt_trigger[0]: @@ -467,31 +454,25 @@ class FragmentFD(FileDownloader): headers['Range'] = 'bytes=%d-%d' % (byte_range['start'], byte_range['end'] - 1) # Never skip the first fragment - fatal, count = is_fatal(fragment.get('index') or (frag_index - 1)), 0 - while count <= fragment_retries: - try: - if self._download_fragment(ctx, fragment['url'], info_dict, headers): - break - return - except (compat_urllib_error.HTTPError, http.client.IncompleteRead) as err: - # Unavailable (possibly temporary) fragments may be served. - # First we try to retry then either skip or abort. - # See https://github.com/ytdl-org/youtube-dl/issues/10165, - # https://github.com/ytdl-org/youtube-dl/issues/10448). - count += 1 - ctx['last_error'] = err - if count <= fragment_retries: - self.report_retry_fragment(err, frag_index, count, fragment_retries) - except DownloadError: - # Don't retry fragment if error occurred during HTTP downloading - # itself since it has own retry settings - if not fatal: - break - raise + fatal = is_fatal(fragment.get('index') or (frag_index - 1)) - if count > fragment_retries and fatal: - ctx['dest_stream'].close() - self.report_error('Giving up after %s fragment retries' % fragment_retries) + def error_callback(err, count, retries): + if fatal and count > retries: + ctx['dest_stream'].close() + self.report_retry(err, count, retries, frag_index, fatal) + ctx['last_error'] = err + + for retry in RetryManager(self.params.get('fragment_retries'), error_callback): + try: + ctx['fragment_count'] = fragment.get('fragment_count') + if not self._download_fragment(ctx, fragment['url'], info_dict, headers): + return + except (urllib.error.HTTPError, http.client.IncompleteRead) as err: + retry.error = err + continue + except DownloadError: # has own retry settings + if fatal: + raise def append_fragment(frag_content, frag_index, ctx): if frag_content: @@ -508,8 +489,7 @@ class FragmentFD(FileDownloader): max_workers = math.ceil( self.params.get('concurrent_fragment_downloads', 1) / ctx.get('max_progress', 1)) - if can_threaded_download and max_workers > 1: - + if max_workers > 1: def _download_fragment(fragment): ctx_copy = ctx.copy() download_fragment(fragment, ctx_copy) @@ -517,23 +497,36 @@ class FragmentFD(FileDownloader): self.report_warning('The download speed shown is only of one thread. This is a known issue and patches are welcome') with tpe or concurrent.futures.ThreadPoolExecutor(max_workers) as pool: - for fragment, frag_index, frag_filename in pool.map(_download_fragment, fragments): - ctx['fragment_filename_sanitized'] = frag_filename - ctx['fragment_index'] = frag_index - result = append_fragment(decrypt_fragment(fragment, self._read_fragment(ctx)), frag_index, ctx) - if not result: - return False + try: + for fragment, frag_index, frag_filename in pool.map(_download_fragment, fragments): + ctx.update({ + 'fragment_filename_sanitized': frag_filename, + 'fragment_index': frag_index, + }) + if not append_fragment(decrypt_fragment(fragment, self._read_fragment(ctx)), frag_index, ctx): + return False + except KeyboardInterrupt: + self._finish_multiline_status() + self.report_error( + 'Interrupted by user. Waiting for all threads to shutdown...', is_error=False, tb=False) + pool.shutdown(wait=False) + raise else: for fragment in fragments: if not interrupt_trigger[0]: break - download_fragment(fragment, ctx) - result = append_fragment(decrypt_fragment(fragment, self._read_fragment(ctx)), fragment['frag_index'], ctx) + try: + download_fragment(fragment, ctx) + result = append_fragment( + decrypt_fragment(fragment, self._read_fragment(ctx)), fragment['frag_index'], ctx) + except KeyboardInterrupt: + if info_dict.get('is_live'): + break + raise if not result: return False if finish_func is not None: ctx['dest_stream'].write(finish_func()) ctx['dest_stream'].flush() - self._finish_frag_download(ctx, info_dict) - return True + return self._finish_frag_download(ctx, info_dict) diff --git a/hypervideo_dl/downloader/hls.py b/hypervideo_dl/downloader/hls.py index f3f32b5..4520edc 100644 --- a/hypervideo_dl/downloader/hls.py +++ b/hypervideo_dl/downloader/hls.py @@ -1,23 +1,14 @@ -from __future__ import unicode_literals - -import re -import io import binascii +import io +import re +import urllib.parse -from ..downloader import get_suitable_downloader -from .fragment import FragmentFD +from . import get_suitable_downloader from .external import FFmpegFD - -from ..compat import ( - compat_pycrypto_AES, - compat_urlparse, -) -from ..utils import ( - parse_m3u8_attributes, - update_url_query, - bug_reports_message, -) +from .fragment import FragmentFD from .. import webvtt +from ..dependencies import Cryptodome_AES +from ..utils import bug_reports_message, parse_m3u8_attributes, update_url_query class HlsFD(FragmentFD): @@ -70,12 +61,18 @@ class HlsFD(FragmentFD): s = urlh.read().decode('utf-8', 'ignore') can_download, message = self.can_download(s, info_dict, self.params.get('allow_unplayable_formats')), None - if can_download and not compat_pycrypto_AES and '#EXT-X-KEY:METHOD=AES-128' in s: - if FFmpegFD.available(): + if can_download: + has_ffmpeg = FFmpegFD.available() + no_crypto = not Cryptodome_AES and '#EXT-X-KEY:METHOD=AES-128' in s + if no_crypto and has_ffmpeg: can_download, message = False, 'The stream has AES-128 encryption and pycryptodome is not available' - else: + elif no_crypto: message = ('The stream has AES-128 encryption and neither ffmpeg nor pycryptodome are available; ' 'Decryption will be performed natively, but will be extremely slow') + elif info_dict.get('extractor_key') == 'Generic' and re.search(r'(?m)#EXT-X-MEDIA-SEQUENCE:(?!0$)', s): + install_ffmpeg = '' if has_ffmpeg else 'install ffmpeg and ' + message = ('Live HLS streams are not supported by the native downloader. If this is a livestream, ' + f'please {install_ffmpeg}add "--downloader ffmpeg --hls-use-mpegts" to your command') if not can_download: has_drm = re.search('|'.join([ r'#EXT-X-FAXS-CM:', # Adobe Flash Access @@ -102,8 +99,7 @@ class HlsFD(FragmentFD): if real_downloader and not real_downloader.supports_manifest(s): real_downloader = None if real_downloader: - self.to_screen( - '[%s] Fragment downloads will be delegated to %s' % (self.FD_NAME, real_downloader.get_basename())) + self.to_screen(f'[{self.FD_NAME}] Fragment downloads will be delegated to {real_downloader.get_basename()}') def is_ad_fragment_start(s): return (s.startswith('#ANVATO-SEGMENT-INFO') and 'type=ad' in s @@ -150,7 +146,7 @@ class HlsFD(FragmentFD): extra_query = None extra_param_to_segment_url = info_dict.get('extra_param_to_segment_url') if extra_param_to_segment_url: - extra_query = compat_urlparse.parse_qs(extra_param_to_segment_url) + extra_query = urllib.parse.parse_qs(extra_param_to_segment_url) i = 0 media_sequence = 0 decrypt_info = {'METHOD': 'NONE'} @@ -172,7 +168,7 @@ class HlsFD(FragmentFD): frag_url = ( line if re.match(r'^https?://', line) - else compat_urlparse.urljoin(man_url, line)) + else urllib.parse.urljoin(man_url, line)) if extra_query: frag_url = update_url_query(frag_url, extra_query) @@ -197,10 +193,18 @@ class HlsFD(FragmentFD): frag_url = ( map_info.get('URI') if re.match(r'^https?://', map_info.get('URI')) - else compat_urlparse.urljoin(man_url, map_info.get('URI'))) + else urllib.parse.urljoin(man_url, map_info.get('URI'))) if extra_query: frag_url = update_url_query(frag_url, extra_query) + if map_info.get('BYTERANGE'): + splitted_byte_range = map_info.get('BYTERANGE').split('@') + sub_range_start = int(splitted_byte_range[1]) if len(splitted_byte_range) == 2 else byte_range['end'] + byte_range = { + 'start': sub_range_start, + 'end': sub_range_start + int(splitted_byte_range[0]), + } + fragments.append({ 'frag_index': frag_index, 'url': frag_url, @@ -210,14 +214,6 @@ class HlsFD(FragmentFD): }) media_sequence += 1 - if map_info.get('BYTERANGE'): - splitted_byte_range = map_info.get('BYTERANGE').split('@') - sub_range_start = int(splitted_byte_range[1]) if len(splitted_byte_range) == 2 else byte_range['end'] - byte_range = { - 'start': sub_range_start, - 'end': sub_range_start + int(splitted_byte_range[0]), - } - elif line.startswith('#EXT-X-KEY'): decrypt_url = decrypt_info.get('URI') decrypt_info = parse_m3u8_attributes(line[11:]) @@ -225,7 +221,7 @@ class HlsFD(FragmentFD): if 'IV' in decrypt_info: decrypt_info['IV'] = binascii.unhexlify(decrypt_info['IV'][2:].zfill(32)) if not re.match(r'^https?://', decrypt_info['URI']): - decrypt_info['URI'] = compat_urlparse.urljoin( + decrypt_info['URI'] = urllib.parse.urljoin( man_url, decrypt_info['URI']) if extra_query: decrypt_info['URI'] = update_url_query(decrypt_info['URI'], extra_query) @@ -339,7 +335,7 @@ class HlsFD(FragmentFD): continue block.write_into(output) - return output.getvalue().encode('utf-8') + return output.getvalue().encode() def fin_fragments(): dedup_window = extra_state.get('webvtt_dedup_window') @@ -350,7 +346,7 @@ class HlsFD(FragmentFD): for cue in dedup_window: webvtt.CueBlock.from_json(cue).write_into(output) - return output.getvalue().encode('utf-8') + return output.getvalue().encode() self.download_and_append_fragments( ctx, fragments, info_dict, pack_func=pack_fragment, finish_func=fin_fragments) diff --git a/hypervideo_dl/downloader/http.py b/hypervideo_dl/downloader/http.py index 591a9b0..95c870e 100644 --- a/hypervideo_dl/downloader/http.py +++ b/hypervideo_dl/downloader/http.py @@ -1,29 +1,33 @@ -from __future__ import unicode_literals - +import http.client import os +import random +import socket import ssl import time -import random +import urllib.error from .common import FileDownloader -from ..compat import ( - compat_urllib_error, - compat_http_client -) from ..utils import ( ContentTooShortError, + RetryManager, + ThrottledDownload, + XAttrMetadataError, + XAttrUnavailableError, encodeFilename, int_or_none, parse_http_range, sanitized_Request, - ThrottledDownload, try_call, write_xattr, - XAttrMetadataError, - XAttrUnavailableError, ) -RESPONSE_READ_EXCEPTIONS = (TimeoutError, ConnectionError, ssl.SSLError, compat_http_client.HTTPException) +RESPONSE_READ_EXCEPTIONS = ( + TimeoutError, + socket.timeout, # compat: py < 3.10 + ConnectionError, + ssl.SSLError, + http.client.HTTPException +) class HttpFD(FileDownloader): @@ -69,9 +73,6 @@ class HttpFD(FileDownloader): ctx.is_resume = ctx.resume_len > 0 - count = 0 - retries = self.params.get('retries', 0) - class SucceedDownload(Exception): pass @@ -134,19 +135,18 @@ class HttpFD(FileDownloader): if has_range: content_range = ctx.data.headers.get('Content-Range') content_range_start, content_range_end, content_len = parse_http_range(content_range) - if content_range_start is not None and range_start == content_range_start: - # Content-Range is present and matches requested Range, resume is possible - accept_content_len = ( + # Content-Range is present and matches requested Range, resume is possible + if range_start == content_range_start and ( # Non-chunked download not ctx.chunk_size # Chunked download and requested piece or # its part is promised to be served or content_range_end == range_end - or content_len < range_end) - if accept_content_len: - ctx.content_len = content_len - ctx.data_len = min(content_len, req_end or content_len) - (req_start or 0) - return + or content_len < range_end): + ctx.content_len = content_len + if content_len or req_end: + ctx.data_len = min(content_len or req_end, req_end or content_len) - (req_start or 0) + return # Content-Range is either not present or invalid. Assuming remote webserver is # trying to send the whole file, resume is not possible, so wiping the local file # and performing entire redownload @@ -154,7 +154,7 @@ class HttpFD(FileDownloader): ctx.resume_len = 0 ctx.open_mode = 'wb' ctx.data_len = ctx.content_len = int_or_none(ctx.data.info().get('Content-length', None)) - except (compat_urllib_error.HTTPError, ) as err: + except urllib.error.HTTPError as err: if err.code == 416: # Unable to resume (requested range not satisfiable) try: @@ -162,7 +162,7 @@ class HttpFD(FileDownloader): ctx.data = self.ydl.urlopen( sanitized_Request(url, request_data, headers)) content_length = ctx.data.info()['Content-Length'] - except (compat_urllib_error.HTTPError, ) as err: + except urllib.error.HTTPError as err: if err.code < 500 or err.code >= 600: raise else: @@ -195,7 +195,7 @@ class HttpFD(FileDownloader): # Unexpected HTTP error raise raise RetryDownload(err) - except compat_urllib_error.URLError as err: + except urllib.error.URLError as err: if isinstance(err.reason, ssl.CertificateError): raise raise RetryDownload(err) @@ -204,6 +204,12 @@ class HttpFD(FileDownloader): except RESPONSE_READ_EXCEPTIONS as err: raise RetryDownload(err) + def close_stream(): + if ctx.stream is not None: + if not ctx.tmpfilename == '-': + ctx.stream.close() + ctx.stream = None + def download(): data_len = ctx.data.info().get('Content-length', None) @@ -220,10 +226,12 @@ class HttpFD(FileDownloader): min_data_len = self.params.get('min_filesize') max_data_len = self.params.get('max_filesize') if min_data_len is not None and data_len < min_data_len: - self.to_screen('\r[download] File is smaller than min-filesize (%s bytes < %s bytes). Aborting.' % (data_len, min_data_len)) + self.to_screen( + f'\r[download] File is smaller than min-filesize ({data_len} bytes < {min_data_len} bytes). Aborting.') return False if max_data_len is not None and data_len > max_data_len: - self.to_screen('\r[download] File is larger than max-filesize (%s bytes > %s bytes). Aborting.' % (data_len, max_data_len)) + self.to_screen( + f'\r[download] File is larger than max-filesize ({data_len} bytes > {max_data_len} bytes). Aborting.') return False byte_counter = 0 + ctx.resume_len @@ -235,12 +243,9 @@ class HttpFD(FileDownloader): before = start # start measuring def retry(e): - to_stdout = ctx.tmpfilename == '-' - if ctx.stream is not None: - if not to_stdout: - ctx.stream.close() - ctx.stream = None - ctx.resume_len = byte_counter if to_stdout else os.path.getsize(encodeFilename(ctx.tmpfilename)) + close_stream() + ctx.resume_len = (byte_counter if ctx.tmpfilename == '-' + else os.path.getsize(encodeFilename(ctx.tmpfilename))) raise RetryDownload(e) while True: @@ -264,19 +269,19 @@ class HttpFD(FileDownloader): assert ctx.stream is not None ctx.filename = self.undo_temp_name(ctx.tmpfilename) self.report_destination(ctx.filename) - except (OSError, IOError) as err: + except OSError as err: self.report_error('unable to open for writing: %s' % str(err)) return False if self.params.get('xattr_set_filesize', False) and data_len is not None: try: - write_xattr(ctx.tmpfilename, 'user.ytdl.filesize', str(data_len).encode('utf-8')) + write_xattr(ctx.tmpfilename, 'user.ytdl.filesize', str(data_len).encode()) except (XAttrUnavailableError, XAttrMetadataError) as err: self.report_error('unable to set filesize xattr: %s' % str(err)) try: ctx.stream.write(data_block) - except (IOError, OSError) as err: + except OSError as err: self.to_stderr('\n') self.report_error('unable to write data: %s' % str(err)) return False @@ -342,9 +347,7 @@ class HttpFD(FileDownloader): if data_len is not None and byte_counter != data_len: err = ContentTooShortError(byte_counter, int(data_len)) - if count <= retries: - retry(err) - raise err + retry(err) self.try_rename(ctx.tmpfilename, ctx.filename) @@ -363,21 +366,20 @@ class HttpFD(FileDownloader): return True - while count <= retries: + for retry in RetryManager(self.params.get('retries'), self.report_retry): try: establish_connection() return download() - except RetryDownload as e: - count += 1 - if count <= retries: - self.report_retry(e.source_error, count, retries) - else: - self.to_screen(f'[download] Got server HTTP error: {e.source_error}') + except RetryDownload as err: + retry.error = err.source_error continue except NextFragment: + retry.error = None + retry.attempt -= 1 continue except SucceedDownload: return True - - self.report_error('giving up after %s retries' % retries) + except: # noqa: E722 + close_stream() + raise return False diff --git a/hypervideo_dl/downloader/ism.py b/hypervideo_dl/downloader/ism.py index 4d5618c..a157a8a 100644 --- a/hypervideo_dl/downloader/ism.py +++ b/hypervideo_dl/downloader/ism.py @@ -1,27 +1,23 @@ -from __future__ import unicode_literals - -import time import binascii import io +import struct +import time +import urllib.error from .fragment import FragmentFD -from ..compat import ( - compat_Struct, - compat_urllib_error, -) +from ..utils import RetryManager +u8 = struct.Struct('>B') +u88 = struct.Struct('>Bx') +u16 = struct.Struct('>H') +u1616 = struct.Struct('>Hxx') +u32 = struct.Struct('>I') +u64 = struct.Struct('>Q') -u8 = compat_Struct('>B') -u88 = compat_Struct('>Bx') -u16 = compat_Struct('>H') -u1616 = compat_Struct('>Hxx') -u32 = compat_Struct('>I') -u64 = compat_Struct('>Q') - -s88 = compat_Struct('>bx') -s16 = compat_Struct('>h') -s1616 = compat_Struct('>hxx') -s32 = compat_Struct('>i') +s88 = struct.Struct('>bx') +s16 = struct.Struct('>h') +s1616 = struct.Struct('>hxx') +s32 = struct.Struct('>i') unity_matrix = (s32.pack(0x10000) + s32.pack(0) * 3) * 2 + s32.pack(0x40000000) @@ -142,6 +138,8 @@ def write_piff_header(stream, params): if fourcc == 'AACL': sample_entry_box = box(b'mp4a', sample_entry_payload) + if fourcc == 'EC-3': + sample_entry_box = box(b'ec-3', sample_entry_payload) elif stream_type == 'video': sample_entry_payload += u16.pack(0) # pre defined sample_entry_payload += u16.pack(0) # reserved @@ -156,7 +154,7 @@ def write_piff_header(stream, params): sample_entry_payload += u16.pack(0x18) # depth sample_entry_payload += s16.pack(-1) # pre defined - codec_private_data = binascii.unhexlify(params['codec_private_data'].encode('utf-8')) + codec_private_data = binascii.unhexlify(params['codec_private_data'].encode()) if fourcc in ('H264', 'AVC1'): sps, pps = codec_private_data.split(u32.pack(1))[1:] avcc_payload = u8.pack(1) # configuration version @@ -235,8 +233,6 @@ class IsmFD(FragmentFD): Download segments in a ISM manifest """ - FD_NAME = 'ism' - def real_download(self, filename, info_dict): segments = info_dict['fragments'][:1] if self.params.get( 'test', False) else info_dict['fragments'] @@ -252,7 +248,6 @@ class IsmFD(FragmentFD): 'ism_track_written': False, }) - fragment_retries = self.params.get('fragment_retries', 0) skip_unavailable_fragments = self.params.get('skip_unavailable_fragments', True) frag_index = 0 @@ -260,8 +255,10 @@ class IsmFD(FragmentFD): frag_index += 1 if frag_index <= ctx['fragment_index']: continue - count = 0 - while count <= fragment_retries: + + retry_manager = RetryManager(self.params.get('fragment_retries'), self.report_retry, + frag_index=frag_index, fatal=not skip_unavailable_fragments) + for retry in retry_manager: try: success = self._download_fragment(ctx, segment['url'], info_dict) if not success: @@ -274,18 +271,13 @@ class IsmFD(FragmentFD): write_piff_header(ctx['dest_stream'], info_dict['_download_params']) extra_state['ism_track_written'] = True self._append_fragment(ctx, frag_content) - break - except compat_urllib_error.HTTPError as err: - count += 1 - if count <= fragment_retries: - self.report_retry_fragment(err, frag_index, count, fragment_retries) - if count > fragment_retries: - if skip_unavailable_fragments: - self.report_skip_fragment(frag_index) + except urllib.error.HTTPError as err: + retry.error = err continue - self.report_error('giving up after %s fragment retries' % fragment_retries) - return False - self._finish_frag_download(ctx, info_dict) + if retry_manager.error: + if not skip_unavailable_fragments: + return False + self.report_skip_fragment(frag_index) - return True + return self._finish_frag_download(ctx, info_dict) diff --git a/hypervideo_dl/downloader/mhtml.py b/hypervideo_dl/downloader/mhtml.py index c8332c0..170a78d 100644 --- a/hypervideo_dl/downloader/mhtml.py +++ b/hypervideo_dl/downloader/mhtml.py @@ -1,24 +1,15 @@ -# coding: utf-8 -from __future__ import unicode_literals - import io import quopri import re import uuid from .fragment import FragmentFD -from ..utils import ( - escapeHTML, - formatSeconds, - srt_subtitles_timecode, - urljoin, -) +from ..compat import imghdr +from ..utils import escapeHTML, formatSeconds, srt_subtitles_timecode, urljoin from ..version import __version__ as YT_DLP_VERSION class MhtmlFD(FragmentFD): - FD_NAME = 'mhtml' - _STYLESHEET = """\ html, body { margin: 0; @@ -62,7 +53,7 @@ body > figure > img { def _escape_mime(s): return '=?utf-8?Q?' + (b''.join( bytes((b,)) if b >= 0x20 else b'=%02X' % b - for b in quopri.encodestring(s.encode('utf-8'), header=True) + for b in quopri.encodestring(s.encode(), header=True) )).decode('us-ascii') + '?=' def _gen_cid(self, i, fragment, frag_boundary): @@ -159,7 +150,7 @@ body > figure > img { length=len(stub), title=self._escape_mime(title), stub=stub - ).encode('utf-8')) + ).encode()) extra_state['header_written'] = True for i, fragment in enumerate(fragments): @@ -176,21 +167,13 @@ body > figure > img { continue frag_content = self._read_fragment(ctx) - mime_type = b'image/jpeg' - if frag_content.startswith(b'\x89PNG\r\n\x1a\n'): - mime_type = b'image/png' - if frag_content.startswith((b'GIF87a', b'GIF89a')): - mime_type = b'image/gif' - if frag_content.startswith(b'RIFF') and frag_content[8:12] == 'WEBP': - mime_type = b'image/webp' - frag_header = io.BytesIO() frag_header.write( b'--%b\r\n' % frag_boundary.encode('us-ascii')) frag_header.write( b'Content-ID: <%b>\r\n' % self._gen_cid(i, fragment, frag_boundary).encode('us-ascii')) frag_header.write( - b'Content-type: %b\r\n' % mime_type) + b'Content-type: %b\r\n' % f'image/{imghdr.what(h=frag_content) or "jpeg"}'.encode()) frag_header.write( b'Content-length: %u\r\n' % len(frag_content)) frag_header.write( @@ -203,5 +186,4 @@ body > figure > img { ctx['dest_stream'].write( b'--%b--\r\n\r\n' % frag_boundary.encode('us-ascii')) - self._finish_frag_download(ctx, info_dict) - return True + return self._finish_frag_download(ctx, info_dict) diff --git a/hypervideo_dl/downloader/niconico.py b/hypervideo_dl/downloader/niconico.py index 521dfec..77ed39e 100644 --- a/hypervideo_dl/downloader/niconico.py +++ b/hypervideo_dl/downloader/niconico.py @@ -1,22 +1,17 @@ -# coding: utf-8 -from __future__ import unicode_literals - import threading +from . import get_suitable_downloader from .common import FileDownloader -from ..downloader import get_suitable_downloader -from ..extractor.niconico import NiconicoIE from ..utils import sanitized_Request class NiconicoDmcFD(FileDownloader): """ Downloading niconico douga from DMC with heartbeat """ - FD_NAME = 'niconico_dmc' - def real_download(self, filename, info_dict): - self.to_screen('[%s] Downloading from DMC' % self.FD_NAME) + from ..extractor.niconico import NiconicoIE + self.to_screen('[%s] Downloading from DMC' % self.FD_NAME) ie = NiconicoIE(self.ydl) info_dict, heartbeat_info_dict = ie._get_heartbeat_info(info_dict) @@ -54,4 +49,4 @@ class NiconicoDmcFD(FileDownloader): with heartbeat_lock: timer[0].cancel() download_complete = True - return success + return success diff --git a/hypervideo_dl/downloader/rtmp.py b/hypervideo_dl/downloader/rtmp.py index 90f1acf..0e09525 100644 --- a/hypervideo_dl/downloader/rtmp.py +++ b/hypervideo_dl/downloader/rtmp.py @@ -1,18 +1,15 @@ -from __future__ import unicode_literals - import os import re import subprocess import time from .common import FileDownloader -from ..compat import compat_str from ..utils import ( + Popen, check_executable, - encodeFilename, encodeArgument, + encodeFilename, get_exe_version, - Popen, ) @@ -94,8 +91,7 @@ class RtmpFD(FileDownloader): self.to_screen('') return proc.wait() except BaseException: # Including KeyboardInterrupt - proc.kill() - proc.wait() + proc.kill(timeout=None) raise url = info_dict['url'] @@ -146,7 +142,7 @@ class RtmpFD(FileDownloader): if isinstance(conn, list): for entry in conn: basic_args += ['--conn', entry] - elif isinstance(conn, compat_str): + elif isinstance(conn, str): basic_args += ['--conn', conn] if protocol is not None: basic_args += ['--protocol', protocol] diff --git a/hypervideo_dl/downloader/rtsp.py b/hypervideo_dl/downloader/rtsp.py index 7815d59..e89269f 100644 --- a/hypervideo_dl/downloader/rtsp.py +++ b/hypervideo_dl/downloader/rtsp.py @@ -1,13 +1,8 @@ -from __future__ import unicode_literals - import os import subprocess from .common import FileDownloader -from ..utils import ( - check_executable, - encodeFilename, -) +from ..utils import check_executable, encodeFilename class RtspFD(FileDownloader): @@ -32,7 +27,7 @@ class RtspFD(FileDownloader): retval = subprocess.call(args) if retval == 0: fsize = os.path.getsize(encodeFilename(tmpfilename)) - self.to_screen('\r[%s] %s bytes' % (args[0], fsize)) + self.to_screen(f'\r[{args[0]}] {fsize} bytes') self.try_rename(tmpfilename, filename) self._hook_progress({ 'downloaded_bytes': fsize, diff --git a/hypervideo_dl/downloader/websocket.py b/hypervideo_dl/downloader/websocket.py index 58e2bce..6837ff1 100644 --- a/hypervideo_dl/downloader/websocket.py +++ b/hypervideo_dl/downloader/websocket.py @@ -1,19 +1,12 @@ +import asyncio +import contextlib import os import signal -import asyncio import threading -try: - import websockets -except (ImportError, SyntaxError): - # websockets 3.10 on python 3.6 causes SyntaxError - # See https://github.com/hypervideo/hypervideo/issues/2633 - has_websockets = False -else: - has_websockets = True - from .common import FileDownloader from .external import FFmpegFD +from ..dependencies import websockets class FFmpegSinkFD(FileDownloader): @@ -26,14 +19,12 @@ class FFmpegSinkFD(FileDownloader): async def call_conn(proc, stdin): try: await self.real_connection(stdin, info_dict) - except (BrokenPipeError, OSError): + except OSError: pass finally: - try: + with contextlib.suppress(OSError): stdin.flush() stdin.close() - except OSError: - pass os.kill(os.getpid(), signal.SIGINT) class FFmpegStdinFD(FFmpegFD): diff --git a/hypervideo_dl/downloader/youtube_live_chat.py b/hypervideo_dl/downloader/youtube_live_chat.py index dd21ac8..dfd290a 100644 --- a/hypervideo_dl/downloader/youtube_live_chat.py +++ b/hypervideo_dl/downloader/youtube_live_chat.py @@ -1,24 +1,20 @@ -from __future__ import division, unicode_literals - import json import time +import urllib.error from .fragment import FragmentFD -from ..compat import compat_urllib_error from ..utils import ( - try_get, + RegexNotFoundError, + RetryManager, dict_get, int_or_none, - RegexNotFoundError, + try_get, ) -from ..extractor.youtube import YoutubeBaseInfoExtractor as YT_BaseIE class YoutubeLiveChatFD(FragmentFD): """ Downloads YouTube live chats fragment by fragment """ - FD_NAME = 'youtube_live_chat' - def real_download(self, filename, info_dict): video_id = info_dict['video_id'] self.to_screen('[%s] Downloading live chat' % self.FD_NAME) @@ -26,7 +22,6 @@ class YoutubeLiveChatFD(FragmentFD): self.report_warning('Live chat download runs until the livestream ends. ' 'If you wish to download the video simultaneously, run a separate hypervideo instance') - fragment_retries = self.params.get('fragment_retries', 0) test = self.params.get('test', False) ctx = { @@ -35,7 +30,9 @@ class YoutubeLiveChatFD(FragmentFD): 'total_frags': None, } - ie = YT_BaseIE(self.ydl) + from ..extractor.youtube import YoutubeBaseInfoExtractor + + ie = YoutubeBaseInfoExtractor(self.ydl) start_time = int(time.time() * 1000) @@ -54,7 +51,7 @@ class YoutubeLiveChatFD(FragmentFD): replay_chat_item_action = action['replayChatItemAction'] offset = int(replay_chat_item_action['videoOffsetTimeMsec']) processed_fragment.extend( - json.dumps(action, ensure_ascii=False).encode('utf-8') + b'\n') + json.dumps(action, ensure_ascii=False).encode() + b'\n') if offset is not None: continuation = try_get( live_chat_continuation, @@ -96,7 +93,7 @@ class YoutubeLiveChatFD(FragmentFD): 'isLive': True, } processed_fragment.extend( - json.dumps(pseudo_action, ensure_ascii=False).encode('utf-8') + b'\n') + json.dumps(pseudo_action, ensure_ascii=False).encode() + b'\n') continuation_data_getters = [ lambda x: x['continuations'][0]['invalidationContinuationData'], lambda x: x['continuations'][0]['timedContinuationData'], @@ -112,8 +109,7 @@ class YoutubeLiveChatFD(FragmentFD): return continuation_id, live_offset, click_tracking_params def download_and_parse_fragment(url, frag_index, request_data=None, headers=None): - count = 0 - while count <= fragment_retries: + for retry in RetryManager(self.params.get('fragment_retries'), self.report_retry, frag_index=frag_index): try: success = dl_fragment(url, request_data, headers) if not success: @@ -128,21 +124,15 @@ class YoutubeLiveChatFD(FragmentFD): live_chat_continuation = try_get( data, lambda x: x['continuationContents']['liveChatContinuation'], dict) or {} - if info_dict['protocol'] == 'youtube_live_chat_replay': - if frag_index == 1: - continuation_id, offset, click_tracking_params = try_refresh_replay_beginning(live_chat_continuation) - else: - continuation_id, offset, click_tracking_params = parse_actions_replay(live_chat_continuation) - elif info_dict['protocol'] == 'youtube_live_chat': - continuation_id, offset, click_tracking_params = parse_actions_live(live_chat_continuation) - return True, continuation_id, offset, click_tracking_params - except compat_urllib_error.HTTPError as err: - count += 1 - if count <= fragment_retries: - self.report_retry_fragment(err, frag_index, count, fragment_retries) - if count > fragment_retries: - self.report_error('giving up after %s fragment retries' % fragment_retries) - return False, None, None, None + + func = (info_dict['protocol'] == 'youtube_live_chat' and parse_actions_live + or frag_index == 1 and try_refresh_replay_beginning + or parse_actions_replay) + return (True, *func(live_chat_continuation)) + except urllib.error.HTTPError as err: + retry.error = err + continue + return False, None, None, None self._prepare_and_start_frag_download(ctx, info_dict) @@ -190,7 +180,7 @@ class YoutubeLiveChatFD(FragmentFD): request_data['context']['clickTracking'] = {'clickTrackingParams': click_tracking_params} headers = ie.generate_api_headers(ytcfg=ytcfg, visitor_data=visitor_data) headers.update({'content-type': 'application/json'}) - fragment_request_data = json.dumps(request_data, ensure_ascii=False).encode('utf-8') + b'\n' + fragment_request_data = json.dumps(request_data, ensure_ascii=False).encode() + b'\n' success, continuation_id, offset, click_tracking_params = download_and_parse_fragment( url, frag_index, fragment_request_data, headers) else: @@ -201,8 +191,7 @@ class YoutubeLiveChatFD(FragmentFD): if test: break - self._finish_frag_download(ctx, info_dict) - return True + return self._finish_frag_download(ctx, info_dict) @staticmethod def parse_live_timestamp(action): |