diff options
author | Rogério Brito <rbrito@ime.usp.br> | 2016-02-24 17:23:49 -0300 |
---|---|---|
committer | Rogério Brito <rbrito@ime.usp.br> | 2016-02-24 17:23:49 -0300 |
commit | 9f2b33881274af98a9145c533a1d295fad71521a (patch) | |
tree | b7e8f8ef288b2ec35a41261bb4774f9044cfce83 /youtube_dl/downloader | |
parent | 9ed7fe4fe4c445eb7d9f3197bb300d0db8f1807a (diff) | |
download | youtube-dl-9f2b33881274af98a9145c533a1d295fad71521a.zip youtube-dl-9f2b33881274af98a9145c533a1d295fad71521a.tar.gz youtube-dl-9f2b33881274af98a9145c533a1d295fad71521a.tar.bz2 |
Imported Upstream version 2016.02.22
Diffstat (limited to 'youtube_dl/downloader')
-rw-r--r-- | youtube_dl/downloader/common.py | 9 | ||||
-rw-r--r-- | youtube_dl/downloader/dash.py | 79 | ||||
-rw-r--r-- | youtube_dl/downloader/f4m.py | 24 | ||||
-rw-r--r-- | youtube_dl/downloader/fragment.py | 51 | ||||
-rw-r--r-- | youtube_dl/downloader/hls.py | 24 | ||||
-rw-r--r-- | youtube_dl/downloader/http.py | 4 | ||||
-rw-r--r-- | youtube_dl/downloader/rtmp.py | 16 |
7 files changed, 121 insertions, 86 deletions
diff --git a/youtube_dl/downloader/common.py b/youtube_dl/downloader/common.py index b8bf8da..2d51540 100644 --- a/youtube_dl/downloader/common.py +++ b/youtube_dl/downloader/common.py @@ -5,9 +5,9 @@ import re import sys import time -from ..compat import compat_str from ..utils import ( encodeFilename, + error_to_compat_str, decodeArgument, format_bytes, timeconvert, @@ -45,6 +45,7 @@ class FileDownloader(object): (experimental) external_downloader_args: A list of additional command-line arguments for the external downloader. + hls_use_mpegts: Use the mpegts container for HLS videos. Subclasses of this one must re-define the real_download method. """ @@ -156,7 +157,7 @@ class FileDownloader(object): 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', None) + rate_limit = self.params.get('ratelimit') if rate_limit is None or byte_counter == 0: return if now is None: @@ -186,7 +187,7 @@ class FileDownloader(object): return os.rename(encodeFilename(old_filename), encodeFilename(new_filename)) except (IOError, OSError) as err: - self.report_error('unable to rename file: %s' % compat_str(err)) + self.report_error('unable to rename file: %s' % error_to_compat_str(err)) def try_utime(self, filename, last_modified_hdr): """Try to set the last-modified time of the given file.""" @@ -295,7 +296,7 @@ class FileDownloader(object): def report_retry(self, count, retries): """Report retry in case of HTTP error 5xx""" - self.to_screen('[download] Got server HTTP error. Retrying (attempt %d of %d)...' % (count, retries)) + self.to_screen('[download] Got server HTTP error. Retrying (attempt %d of %.0f)...' % (count, retries)) def report_file_already_downloaded(self, file_name): """Report file has already been fully downloaded.""" diff --git a/youtube_dl/downloader/dash.py b/youtube_dl/downloader/dash.py index 535f2a7..8b1b17c 100644 --- a/youtube_dl/downloader/dash.py +++ b/youtube_dl/downloader/dash.py @@ -1,66 +1,59 @@ from __future__ import unicode_literals +import os import re -from .common import FileDownloader -from ..utils import sanitized_Request +from .fragment import FragmentFD +from ..utils import ( + sanitize_open, + encodeFilename, +) -class DashSegmentsFD(FileDownloader): +class DashSegmentsFD(FragmentFD): """ Download segments in a DASH manifest """ - def real_download(self, filename, info_dict): - self.report_destination(filename) - tmpfilename = self.temp_name(filename) - base_url = info_dict['url'] - segment_urls = info_dict['segment_urls'] - - is_test = self.params.get('test', False) - remaining_bytes = self._TEST_FILE_SIZE if is_test else None - byte_counter = 0 - def append_url_to_file(outf, target_url, target_name, remaining_bytes=None): - self.to_screen('[DashSegments] %s: Downloading %s' % (info_dict['id'], target_name)) - req = sanitized_Request(target_url) - if remaining_bytes is not None: - req.add_header('Range', 'bytes=0-%d' % (remaining_bytes - 1)) + FD_NAME = 'dashsegments' - data = self.ydl.urlopen(req).read() + def real_download(self, filename, info_dict): + base_url = info_dict['url'] + segment_urls = [info_dict['segment_urls'][0]] if self.params.get('test', False) else info_dict['segment_urls'] + initialization_url = info_dict.get('initialization_url') - if remaining_bytes is not None: - data = data[:remaining_bytes] + ctx = { + 'filename': filename, + 'total_frags': len(segment_urls) + (1 if initialization_url else 0), + } - outf.write(data) - return len(data) + self._prepare_and_start_frag_download(ctx) def combine_url(base_url, target_url): if re.match(r'^https?://', target_url): return target_url return '%s%s%s' % (base_url, '' if base_url.endswith('/') else '/', target_url) - with open(tmpfilename, 'wb') as outf: - append_url_to_file( - outf, combine_url(base_url, info_dict['initialization_url']), - 'initialization segment') - for i, segment_url in enumerate(segment_urls): - segment_len = append_url_to_file( - outf, combine_url(base_url, segment_url), - 'segment %d / %d' % (i + 1, len(segment_urls)), - remaining_bytes) - byte_counter += segment_len - if remaining_bytes is not None: - remaining_bytes -= segment_len - if remaining_bytes <= 0: - break + segments_filenames = [] - self.try_rename(tmpfilename, filename) + def append_url_to_file(target_url, target_filename): + success = ctx['dl'].download(target_filename, {'url': combine_url(base_url, target_url)}) + if not success: + return False + down, target_sanitized = sanitize_open(target_filename, 'rb') + ctx['dest_stream'].write(down.read()) + down.close() + segments_filenames.append(target_sanitized) - self._hook_progress({ - 'downloaded_bytes': byte_counter, - 'total_bytes': byte_counter, - 'filename': filename, - 'status': 'finished', - }) + if initialization_url: + append_url_to_file(initialization_url, ctx['tmpfilename'] + '-Init') + for i, segment_url in enumerate(segment_urls): + segment_filename = '%s-Seg%d' % (ctx['tmpfilename'], i) + append_url_to_file(segment_url, segment_filename) + + self._finish_frag_download(ctx) + + for segment_file in segments_filenames: + os.remove(encodeFilename(segment_file)) return True diff --git a/youtube_dl/downloader/f4m.py b/youtube_dl/downloader/f4m.py index 6170cc1..fc96429 100644 --- a/youtube_dl/downloader/f4m.py +++ b/youtube_dl/downloader/f4m.py @@ -15,6 +15,7 @@ from ..compat import ( ) from ..utils import ( encodeFilename, + fix_xml_ampersands, sanitize_open, struct_pack, struct_unpack, @@ -272,15 +273,21 @@ class F4mFD(FragmentFD): return fragments_list def _parse_bootstrap_node(self, node, base_url): - if node.text is None: + # Sometimes non empty inline bootstrap info can be specified along + # with bootstrap url attribute (e.g. dummy inline bootstrap info + # contains whitespace characters in [1]). We will prefer bootstrap + # url over inline bootstrap info when present. + # 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( - base_url, node.attrib['url']) + base_url, bootstrap_url) boot_info = self._get_bootstrap_from_url(bootstrap_url) else: bootstrap_url = None bootstrap = base64.b64decode(node.text.encode('ascii')) boot_info = read_bootstrap_info(bootstrap) - return (boot_info, bootstrap_url) + return boot_info, bootstrap_url def real_download(self, filename, info_dict): man_url = info_dict['url'] @@ -288,7 +295,10 @@ class F4mFD(FragmentFD): self.to_screen('[%s] Downloading f4m manifest' % self.FD_NAME) urlh = self.ydl.urlopen(man_url) man_url = urlh.geturl() - manifest = urlh.read() + # Some manifests may be malformed, e.g. prosiebensat1 generated manifests + # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244 + # and https://github.com/rg3/youtube-dl/issues/7823) + manifest = fix_xml_ampersands(urlh.read().decode('utf-8', 'ignore')).strip() doc = compat_etree_fromstring(manifest) formats = [(int(f.attrib.get('bitrate', -1)), f) @@ -312,7 +322,8 @@ class F4mFD(FragmentFD): metadata = None fragments_list = build_fragments_list(boot_info) - if self.params.get('test', False): + test = self.params.get('test', False) + if test: # We only download the first fragment fragments_list = fragments_list[:1] total_frags = len(fragments_list) @@ -322,6 +333,7 @@ class F4mFD(FragmentFD): ctx = { 'filename': filename, 'total_frags': total_frags, + 'live': live, } self._prepare_frag_download(ctx) @@ -376,7 +388,7 @@ class F4mFD(FragmentFD): else: raise - if not fragments_list and live and bootstrap_url: + if not fragments_list and not test and live and bootstrap_url: fragments_list = self._update_live_fragments(bootstrap_url, frag_i) total_frags += len(fragments_list) if fragments_list and (fragments_list[0][1] > frag_i + 1): diff --git a/youtube_dl/downloader/fragment.py b/youtube_dl/downloader/fragment.py index 5a64b29..5bc9949 100644 --- a/youtube_dl/downloader/fragment.py +++ b/youtube_dl/downloader/fragment.py @@ -26,7 +26,11 @@ class FragmentFD(FileDownloader): self._start_frag_download(ctx) def _prepare_frag_download(self, ctx): - self.to_screen('[%s] Total fragments: %d' % (self.FD_NAME, ctx['total_frags'])) + if 'live' not in ctx: + ctx['live'] = False + self.to_screen( + '[%s] Total fragments: %s' + % (self.FD_NAME, ctx['total_frags'] if not ctx['live'] else 'unknown (live)')) self.report_destination(ctx['filename']) dl = HttpQuietDownloader( self.ydl, @@ -34,7 +38,7 @@ class FragmentFD(FileDownloader): 'continuedl': True, 'quiet': True, 'noprogress': True, - 'ratelimit': self.params.get('ratelimit', None), + 'ratelimit': self.params.get('ratelimit'), 'retries': self.params.get('retries', 0), 'test': self.params.get('test', False), } @@ -59,37 +63,44 @@ class FragmentFD(FileDownloader): 'filename': ctx['filename'], 'tmpfilename': ctx['tmpfilename'], } + start = time.time() - ctx['started'] = start + ctx.update({ + 'started': start, + # Total complete fragments downloaded so far in bytes + 'complete_frags_downloaded_bytes': 0, + # Amount of fragment's bytes downloaded by the time of the previous + # frag progress hook invocation + 'prev_frag_downloaded_bytes': 0, + }) def frag_progress_hook(s): if s['status'] not in ('downloading', 'finished'): return - frag_total_bytes = s.get('total_bytes', 0) - if s['status'] == 'finished': - state['downloaded_bytes'] += frag_total_bytes - state['frag_index'] += 1 - - estimated_size = ( - (state['downloaded_bytes'] + frag_total_bytes) / - (state['frag_index'] + 1) * total_frags) time_now = time.time() - state['total_bytes_estimate'] = estimated_size state['elapsed'] = time_now - start + frag_total_bytes = s.get('total_bytes') or 0 + if not ctx['live']: + estimated_size = ( + (ctx['complete_frags_downloaded_bytes'] + frag_total_bytes) / + (state['frag_index'] + 1) * total_frags) + state['total_bytes_estimate'] = estimated_size if s['status'] == 'finished': - progress = self.calc_percent(state['frag_index'], total_frags) + state['frag_index'] += 1 + state['downloaded_bytes'] += frag_total_bytes - ctx['prev_frag_downloaded_bytes'] + ctx['complete_frags_downloaded_bytes'] = state['downloaded_bytes'] + ctx['prev_frag_downloaded_bytes'] = 0 else: frag_downloaded_bytes = s['downloaded_bytes'] - frag_progress = self.calc_percent(frag_downloaded_bytes, - frag_total_bytes) - progress = self.calc_percent(state['frag_index'], total_frags) - progress += frag_progress / float(total_frags) - - state['eta'] = self.calc_eta( - start, time_now, estimated_size, state['downloaded_bytes'] + frag_downloaded_bytes) + state['downloaded_bytes'] += frag_downloaded_bytes - ctx['prev_frag_downloaded_bytes'] + if not ctx['live']: + state['eta'] = self.calc_eta( + start, time_now, estimated_size, + state['downloaded_bytes']) state['speed'] = s.get('speed') + ctx['prev_frag_downloaded_bytes'] = frag_downloaded_bytes self._hook_progress(state) ctx['dl'].add_progress_hook(frag_progress_hook) diff --git a/youtube_dl/downloader/hls.py b/youtube_dl/downloader/hls.py index 92765a3..2a775bf 100644 --- a/youtube_dl/downloader/hls.py +++ b/youtube_dl/downloader/hls.py @@ -3,6 +3,7 @@ from __future__ import unicode_literals import os import re import subprocess +import sys from .common import FileDownloader from .fragment import FragmentFD @@ -13,6 +14,7 @@ from ..utils import ( encodeArgument, encodeFilename, sanitize_open, + handle_youtubedl_headers, ) @@ -33,18 +35,34 @@ class HlsFD(FileDownloader): if info_dict['http_headers'] and re.match(r'^https?://', url): # 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 += [ '-headers', - ''.join('%s: %s\r\n' % (key, val) for key, val in info_dict['http_headers'].items() if key.lower() != 'accept-encoding')] + ''.join('%s: %s\r\n' % (key, val) for key, val in headers.items())] - args += ['-i', url, '-f', 'mp4', '-c', 'copy', '-bsf:a', 'aac_adtstoasc'] + args += ['-i', url, '-c', 'copy'] + if self.params.get('hls_use_mpegts', False): + args += ['-f', 'mpegts'] + else: + args += ['-f', 'mp4', '-bsf:a', 'aac_adtstoasc'] args = [encodeArgument(opt) for opt in args] args.append(encodeFilename(ffpp._ffmpeg_filename_argument(tmpfilename), True)) self._debug_cmd(args) - retval = subprocess.call(args) + proc = subprocess.Popen(args, stdin=subprocess.PIPE) + try: + retval = proc.wait() + except KeyboardInterrupt: + # 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/rg3/youtube-dl/issues/8300). + if sys.platform != 'win32': + proc.communicate(b'q') + raise if retval == 0: fsize = os.path.getsize(encodeFilename(tmpfilename)) self.to_screen('\r[%s] %s bytes' % (args[0], fsize)) diff --git a/youtube_dl/downloader/http.py b/youtube_dl/downloader/http.py index 56840e0..f8b69d1 100644 --- a/youtube_dl/downloader/http.py +++ b/youtube_dl/downloader/http.py @@ -140,8 +140,8 @@ class HttpFD(FileDownloader): if data_len is not None: data_len = int(data_len) + resume_len - min_data_len = self.params.get("min_filesize", None) - max_data_len = self.params.get("max_filesize", None) + 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)) return False diff --git a/youtube_dl/downloader/rtmp.py b/youtube_dl/downloader/rtmp.py index 14d56db..9de6e70 100644 --- a/youtube_dl/downloader/rtmp.py +++ b/youtube_dl/downloader/rtmp.py @@ -94,15 +94,15 @@ class RtmpFD(FileDownloader): return proc.returncode url = info_dict['url'] - player_url = info_dict.get('player_url', None) - page_url = info_dict.get('page_url', None) - app = info_dict.get('app', None) - play_path = info_dict.get('play_path', None) - tc_url = info_dict.get('tc_url', None) - flash_version = info_dict.get('flash_version', None) + player_url = info_dict.get('player_url') + page_url = info_dict.get('page_url') + app = info_dict.get('app') + play_path = info_dict.get('play_path') + tc_url = info_dict.get('tc_url') + flash_version = info_dict.get('flash_version') live = info_dict.get('rtmp_live', False) - conn = info_dict.get('rtmp_conn', None) - protocol = info_dict.get('rtmp_protocol', None) + conn = info_dict.get('rtmp_conn') + protocol = info_dict.get('rtmp_protocol') real_time = info_dict.get('rtmp_real_time', False) no_resume = info_dict.get('no_resume', False) continue_dl = self.params.get('continuedl', True) |