]>
Commit | Line | Data |
---|---|---|
1d485a1a | 1 | import concurrent.futures |
19a03940 | 2 | import contextlib |
adbc4ec4 THD |
3 | import http.client |
4 | import json | |
5 | import math | |
95d8f7ea S |
6 | import os |
7 | import time | |
8 | ||
9 | from .common import FileDownloader | |
10 | from .http import HttpFD | |
1d3586d0 | 11 | from ..aes import aes_cbc_decrypt_bytes, unpad_pkcs7 |
f8271158 | 12 | from ..compat import compat_os_name, compat_struct_pack, compat_urllib_error |
95d8f7ea | 13 | from ..utils import ( |
4c7853de | 14 | DownloadError, |
95d8f7ea | 15 | encodeFilename, |
f8271158 | 16 | error_to_compat_str, |
69035555 | 17 | sanitized_Request, |
a539f065 | 18 | traverse_obj, |
95d8f7ea S |
19 | ) |
20 | ||
21 | ||
22 | class HttpQuietDownloader(HttpFD): | |
23 | def to_screen(self, *args, **kargs): | |
24 | pass | |
25 | ||
1d485a1a | 26 | console_title = to_screen |
27 | ||
bd93fd5d | 28 | def report_retry(self, err, count, retries): |
29 | super().to_screen( | |
30 | f'[download] Got server HTTP error: {err}. Retrying (attempt {count} of {self.format_retries(retries)}) ...') | |
31 | ||
95d8f7ea S |
32 | |
33 | class FragmentFD(FileDownloader): | |
34 | """ | |
35 | A base file downloader class for fragmented media (e.g. f4m/m3u8 manifests). | |
16a8b798 S |
36 | |
37 | Available options: | |
38 | ||
9603b660 S |
39 | fragment_retries: Number of times to retry a fragment for HTTP error (DASH |
40 | and hlsnative only) | |
41 | skip_unavailable_fragments: | |
42 | Skip unavailable fragments (DASH and hlsnative only) | |
0eee52f3 S |
43 | keep_fragments: Keep downloaded fragments on disk after downloading is |
44 | finished | |
59a7a13e | 45 | concurrent_fragment_downloads: The number of threads to use for native hls and dash downloads |
e8e73840 | 46 | _no_ytdl_file: Don't use .ytdl file |
290f64db | 47 | |
7a5c1cfe | 48 | For each incomplete fragment download yt-dlp keeps on disk a special |
290f64db | 49 | bookkeeping file with download state and metadata (in future such files will |
7a5c1cfe | 50 | be used for any incomplete download handled by yt-dlp). This file is |
290f64db S |
51 | used to properly handle resuming, check download file consistency and detect |
52 | potential errors. The file has a .ytdl extension and represents a standard | |
53 | JSON file of the following format: | |
54 | ||
55 | extractor: | |
56 | Dictionary of extractor related data. TBD. | |
57 | ||
58 | downloader: | |
59 | Dictionary of downloader related data. May contain following data: | |
60 | current_fragment: | |
61 | Dictionary with current (being downloaded) fragment data: | |
85f6de25 | 62 | index: 0-based index of current fragment among all fragments |
290f64db S |
63 | fragment_count: |
64 | Total count of fragments | |
50534b71 | 65 | |
85f6de25 | 66 | This feature is experimental and file format may change in future. |
95d8f7ea S |
67 | """ |
68 | ||
75a24854 | 69 | def report_retry_fragment(self, err, frag_index, count, retries): |
721f26b8 | 70 | self.to_screen( |
4c7853de | 71 | '\r[download] Got server HTTP error: %s. Retrying fragment %d (attempt %d of %s) ...' |
75a24854 | 72 | % (error_to_compat_str(err), frag_index, count, self.format_retries(retries))) |
721f26b8 | 73 | |
b4b855eb | 74 | def report_skip_fragment(self, frag_index, err=None): |
75 | err = f' {err};' if err else '' | |
76 | self.to_screen(f'[download]{err} Skipping fragment {frag_index:d} ...') | |
9603b660 | 77 | |
69035555 S |
78 | def _prepare_url(self, info_dict, url): |
79 | headers = info_dict.get('http_headers') | |
80 | return sanitized_Request(url, None, headers) if headers else url | |
81 | ||
3ba7740d | 82 | def _prepare_and_start_frag_download(self, ctx, info_dict): |
95d8f7ea | 83 | self._prepare_frag_download(ctx) |
3ba7740d | 84 | self._start_frag_download(ctx, info_dict) |
95d8f7ea | 85 | |
e8e73840 | 86 | def __do_ytdl_file(self, ctx): |
adbc4ec4 | 87 | return ctx['live'] is not True and ctx['tmpfilename'] != '-' and not self.params.get('_no_ytdl_file') |
adb4b03c | 88 | |
d3f0687c | 89 | def _read_ytdl_file(self, ctx): |
500a86a5 | 90 | assert 'ytdl_corrupt' not in ctx |
205a0654 | 91 | stream, _ = self.sanitize_open(self.ytdl_filename(ctx['filename']), 'r') |
500a86a5 | 92 | try: |
4d49884c F |
93 | ytdl_data = json.loads(stream.read()) |
94 | ctx['fragment_index'] = ytdl_data['downloader']['current_fragment']['index'] | |
95 | if 'extra_state' in ytdl_data['downloader']: | |
96 | ctx['extra_state'] = ytdl_data['downloader']['extra_state'] | |
500a86a5 S |
97 | except Exception: |
98 | ctx['ytdl_corrupt'] = True | |
99 | finally: | |
100 | stream.close() | |
d3f0687c S |
101 | |
102 | def _write_ytdl_file(self, ctx): | |
205a0654 | 103 | frag_index_stream, _ = self.sanitize_open(self.ytdl_filename(ctx['filename']), 'w') |
ad3dc496 | 104 | try: |
105 | downloader = { | |
106 | 'current_fragment': { | |
107 | 'index': ctx['fragment_index'], | |
108 | }, | |
109 | } | |
110 | if 'extra_state' in ctx: | |
111 | downloader['extra_state'] = ctx['extra_state'] | |
112 | if ctx.get('fragment_count') is not None: | |
113 | downloader['fragment_count'] = ctx['fragment_count'] | |
114 | frag_index_stream.write(json.dumps({'downloader': downloader})) | |
115 | finally: | |
116 | frag_index_stream.close() | |
d3f0687c | 117 | |
273762c8 | 118 | def _download_fragment(self, ctx, frag_url, info_dict, headers=None, request_data=None): |
d3f0687c | 119 | fragment_filename = '%s-Frag%d' % (ctx['tmpfilename'], ctx['fragment_index']) |
38d70284 | 120 | fragment_info_dict = { |
75a24854 RA |
121 | 'url': frag_url, |
122 | 'http_headers': headers or info_dict.get('http_headers'), | |
273762c8 | 123 | 'request_data': request_data, |
bd50a52b | 124 | 'ctx_id': ctx.get('ctx_id'), |
38d70284 | 125 | } |
3fe75fdc | 126 | success, _ = ctx['dl'].download(fragment_filename, fragment_info_dict) |
75a24854 | 127 | if not success: |
d71fd412 | 128 | return False |
38d70284 | 129 | if fragment_info_dict.get('filetime'): |
130 | ctx['fragment_filetime'] = fragment_info_dict.get('filetime') | |
4c7853de | 131 | ctx['fragment_filename_sanitized'] = fragment_filename |
d71fd412 | 132 | return True |
4c7853de | 133 | |
134 | def _read_fragment(self, ctx): | |
e06bd880 | 135 | if not ctx.get('fragment_filename_sanitized'): |
136 | return None | |
d71fd412 LNO |
137 | try: |
138 | down, frag_sanitized = self.sanitize_open(ctx['fragment_filename_sanitized'], 'rb') | |
139 | except FileNotFoundError: | |
140 | if ctx.get('live'): | |
141 | return None | |
142 | raise | |
d3f0687c S |
143 | ctx['fragment_filename_sanitized'] = frag_sanitized |
144 | frag_content = down.read() | |
75a24854 | 145 | down.close() |
4c7853de | 146 | return frag_content |
75a24854 RA |
147 | |
148 | def _append_fragment(self, ctx, frag_content): | |
d3f0687c S |
149 | try: |
150 | ctx['dest_stream'].write(frag_content) | |
593f2f79 | 151 | ctx['dest_stream'].flush() |
d3f0687c | 152 | finally: |
adb4b03c | 153 | if self.__do_ytdl_file(ctx): |
d3f0687c | 154 | self._write_ytdl_file(ctx) |
0eee52f3 | 155 | if not self.params.get('keep_fragments', False): |
45806d44 | 156 | self.try_remove(encodeFilename(ctx['fragment_filename_sanitized'])) |
d3f0687c | 157 | del ctx['fragment_filename_sanitized'] |
75a24854 | 158 | |
95d8f7ea | 159 | def _prepare_frag_download(self, ctx): |
5fa1702c S |
160 | if 'live' not in ctx: |
161 | ctx['live'] = False | |
5efaf43c S |
162 | if not ctx['live']: |
163 | total_frags_str = '%d' % ctx['total_frags'] | |
164 | ad_frags = ctx.get('ad_frags', 0) | |
165 | if ad_frags: | |
166 | total_frags_str += ' (not including %d ad)' % ad_frags | |
167 | else: | |
168 | total_frags_str = 'unknown (live)' | |
86e5f3ed | 169 | self.to_screen(f'[{self.FD_NAME}] Total fragments: {total_frags_str}') |
95d8f7ea S |
170 | self.report_destination(ctx['filename']) |
171 | dl = HttpQuietDownloader( | |
172 | self.ydl, | |
173 | { | |
93c8410d | 174 | 'continuedl': self.params.get('continuedl', True), |
bd93fd5d | 175 | 'quiet': self.params.get('quiet'), |
95d8f7ea | 176 | 'noprogress': True, |
d800609c | 177 | 'ratelimit': self.params.get('ratelimit'), |
6828c809 | 178 | 'retries': self.params.get('retries', 0), |
7097bffb | 179 | 'nopart': self.params.get('nopart', False), |
3b9d9f43 | 180 | 'test': False, |
95d8f7ea S |
181 | } |
182 | ) | |
183 | tmpfilename = self.temp_name(ctx['filename']) | |
75a24854 RA |
184 | open_mode = 'wb' |
185 | resume_len = 0 | |
d3f0687c | 186 | |
75a24854 RA |
187 | # Establish possible resume length |
188 | if os.path.isfile(encodeFilename(tmpfilename)): | |
189 | open_mode = 'ab' | |
190 | resume_len = os.path.getsize(encodeFilename(tmpfilename)) | |
d3f0687c | 191 | |
adb4b03c S |
192 | # Should be initialized before ytdl file check |
193 | ctx.update({ | |
194 | 'tmpfilename': tmpfilename, | |
195 | 'fragment_index': 0, | |
196 | }) | |
d3f0687c | 197 | |
adb4b03c S |
198 | if self.__do_ytdl_file(ctx): |
199 | if os.path.isfile(encodeFilename(self.ytdl_filename(ctx['filename']))): | |
200 | self._read_ytdl_file(ctx) | |
500a86a5 S |
201 | is_corrupt = ctx.get('ytdl_corrupt') is True |
202 | is_inconsistent = ctx['fragment_index'] > 0 and resume_len == 0 | |
203 | if is_corrupt or is_inconsistent: | |
204 | message = ( | |
205 | '.ytdl file is corrupt' if is_corrupt else | |
206 | 'Inconsistent state of incomplete fragment download') | |
6f3b4a98 | 207 | self.report_warning( |
5ef7d9bd | 208 | '%s. Restarting from the beginning ...' % message) |
e7c3e334 | 209 | ctx['fragment_index'] = resume_len = 0 |
500a86a5 S |
210 | if 'ytdl_corrupt' in ctx: |
211 | del ctx['ytdl_corrupt'] | |
e7c3e334 | 212 | self._write_ytdl_file(ctx) |
adb4b03c S |
213 | else: |
214 | self._write_ytdl_file(ctx) | |
e7c3e334 | 215 | assert ctx['fragment_index'] == 0 |
d3f0687c | 216 | |
205a0654 | 217 | dest_stream, tmpfilename = self.sanitize_open(tmpfilename, open_mode) |
75a24854 | 218 | |
95d8f7ea S |
219 | ctx.update({ |
220 | 'dl': dl, | |
221 | 'dest_stream': dest_stream, | |
222 | 'tmpfilename': tmpfilename, | |
75a24854 RA |
223 | # Total complete fragments downloaded so far in bytes |
224 | 'complete_frags_downloaded_bytes': resume_len, | |
95d8f7ea S |
225 | }) |
226 | ||
3ba7740d | 227 | def _start_frag_download(self, ctx, info_dict): |
3bce4ff7 | 228 | resume_len = ctx['complete_frags_downloaded_bytes'] |
95d8f7ea | 229 | total_frags = ctx['total_frags'] |
bd50a52b | 230 | ctx_id = ctx.get('ctx_id') |
95d8f7ea S |
231 | # This dict stores the download progress, it's updated by the progress |
232 | # hook | |
233 | state = { | |
234 | 'status': 'downloading', | |
3bce4ff7 | 235 | 'downloaded_bytes': resume_len, |
3e0304fe RA |
236 | 'fragment_index': ctx['fragment_index'], |
237 | 'fragment_count': total_frags, | |
95d8f7ea S |
238 | 'filename': ctx['filename'], |
239 | 'tmpfilename': ctx['tmpfilename'], | |
b83b782d S |
240 | } |
241 | ||
242 | start = time.time() | |
243 | ctx.update({ | |
244 | 'started': start, | |
bd93fd5d | 245 | 'fragment_started': start, |
709185a2 S |
246 | # Amount of fragment's bytes downloaded by the time of the previous |
247 | # frag progress hook invocation | |
b83b782d S |
248 | 'prev_frag_downloaded_bytes': 0, |
249 | }) | |
95d8f7ea S |
250 | |
251 | def frag_progress_hook(s): | |
252 | if s['status'] not in ('downloading', 'finished'): | |
253 | return | |
254 | ||
bd50a52b THD |
255 | if ctx_id is not None and s.get('ctx_id') != ctx_id: |
256 | return | |
257 | ||
258 | state['max_progress'] = ctx.get('max_progress') | |
259 | state['progress_idx'] = ctx.get('progress_idx') | |
260 | ||
5fa1702c | 261 | time_now = time.time() |
2c2f1efd | 262 | state['elapsed'] = time_now - start |
3c91e416 | 263 | frag_total_bytes = s.get('total_bytes') or 0 |
3ba7740d | 264 | s['fragment_info_dict'] = s.pop('info_dict', {}) |
5fa1702c S |
265 | if not ctx['live']: |
266 | estimated_size = ( | |
3089bc74 S |
267 | (ctx['complete_frags_downloaded_bytes'] + frag_total_bytes) |
268 | / (state['fragment_index'] + 1) * total_frags) | |
5fa1702c | 269 | state['total_bytes_estimate'] = estimated_size |
95d8f7ea | 270 | |
709185a2 | 271 | if s['status'] == 'finished': |
3e0304fe RA |
272 | state['fragment_index'] += 1 |
273 | ctx['fragment_index'] = state['fragment_index'] | |
b83b782d S |
274 | state['downloaded_bytes'] += frag_total_bytes - ctx['prev_frag_downloaded_bytes'] |
275 | ctx['complete_frags_downloaded_bytes'] = state['downloaded_bytes'] | |
bd93fd5d | 276 | ctx['speed'] = state['speed'] = self.calc_speed( |
277 | ctx['fragment_started'], time_now, frag_total_bytes) | |
278 | ctx['fragment_started'] = time.time() | |
b83b782d | 279 | ctx['prev_frag_downloaded_bytes'] = 0 |
709185a2 S |
280 | else: |
281 | frag_downloaded_bytes = s['downloaded_bytes'] | |
b83b782d | 282 | state['downloaded_bytes'] += frag_downloaded_bytes - ctx['prev_frag_downloaded_bytes'] |
5fa1702c S |
283 | if not ctx['live']: |
284 | state['eta'] = self.calc_eta( | |
3bce4ff7 | 285 | start, time_now, estimated_size - resume_len, |
286 | state['downloaded_bytes'] - resume_len) | |
bd93fd5d | 287 | ctx['speed'] = state['speed'] = self.calc_speed( |
288 | ctx['fragment_started'], time_now, frag_downloaded_bytes) | |
b83b782d | 289 | ctx['prev_frag_downloaded_bytes'] = frag_downloaded_bytes |
3ba7740d | 290 | self._hook_progress(state, info_dict) |
95d8f7ea S |
291 | |
292 | ctx['dl'].add_progress_hook(frag_progress_hook) | |
293 | ||
294 | return start | |
295 | ||
3ba7740d | 296 | def _finish_frag_download(self, ctx, info_dict): |
95d8f7ea | 297 | ctx['dest_stream'].close() |
adb4b03c S |
298 | if self.__do_ytdl_file(ctx): |
299 | ytdl_filename = encodeFilename(self.ytdl_filename(ctx['filename'])) | |
300 | if os.path.isfile(ytdl_filename): | |
45806d44 | 301 | self.try_remove(ytdl_filename) |
95d8f7ea | 302 | elapsed = time.time() - ctx['started'] |
0ff2c1ec S |
303 | |
304 | if ctx['tmpfilename'] == '-': | |
305 | downloaded_bytes = ctx['complete_frags_downloaded_bytes'] | |
306 | else: | |
307 | self.try_rename(ctx['tmpfilename'], ctx['filename']) | |
38d70284 | 308 | if self.params.get('updatetime', True): |
309 | filetime = ctx.get('fragment_filetime') | |
310 | if filetime: | |
19a03940 | 311 | with contextlib.suppress(Exception): |
38d70284 | 312 | os.utime(ctx['filename'], (time.time(), filetime)) |
0ff2c1ec | 313 | downloaded_bytes = os.path.getsize(encodeFilename(ctx['filename'])) |
95d8f7ea S |
314 | |
315 | self._hook_progress({ | |
0ff2c1ec S |
316 | 'downloaded_bytes': downloaded_bytes, |
317 | 'total_bytes': downloaded_bytes, | |
95d8f7ea S |
318 | 'filename': ctx['filename'], |
319 | 'status': 'finished', | |
320 | 'elapsed': elapsed, | |
bd50a52b THD |
321 | 'ctx_id': ctx.get('ctx_id'), |
322 | 'max_progress': ctx.get('max_progress'), | |
323 | 'progress_idx': ctx.get('progress_idx'), | |
3ba7740d | 324 | }, info_dict) |
5219cb3e | 325 | |
326 | def _prepare_external_frag_download(self, ctx): | |
327 | if 'live' not in ctx: | |
328 | ctx['live'] = False | |
329 | if not ctx['live']: | |
330 | total_frags_str = '%d' % ctx['total_frags'] | |
331 | ad_frags = ctx.get('ad_frags', 0) | |
332 | if ad_frags: | |
333 | total_frags_str += ' (not including %d ad)' % ad_frags | |
334 | else: | |
335 | total_frags_str = 'unknown (live)' | |
86e5f3ed | 336 | self.to_screen(f'[{self.FD_NAME}] Total fragments: {total_frags_str}') |
5219cb3e | 337 | |
338 | tmpfilename = self.temp_name(ctx['filename']) | |
339 | ||
340 | # Should be initialized before ytdl file check | |
341 | ctx.update({ | |
342 | 'tmpfilename': tmpfilename, | |
343 | 'fragment_index': 0, | |
344 | }) | |
4c7853de | 345 | |
1009f67c | 346 | def decrypter(self, info_dict): |
347 | _key_cache = {} | |
348 | ||
349 | def _get_key(url): | |
350 | if url not in _key_cache: | |
351 | _key_cache[url] = self.ydl.urlopen(self._prepare_url(info_dict, url)).read() | |
352 | return _key_cache[url] | |
353 | ||
354 | def decrypt_fragment(fragment, frag_content): | |
355 | decrypt_info = fragment.get('decrypt_info') | |
356 | if not decrypt_info or decrypt_info['METHOD'] != 'AES-128': | |
357 | return frag_content | |
358 | iv = decrypt_info.get('IV') or compat_struct_pack('>8xq', fragment['media_sequence']) | |
359 | decrypt_info['KEY'] = decrypt_info.get('KEY') or _get_key(info_dict.get('_decryption_key_url') or decrypt_info['URI']) | |
360 | # Don't decrypt the content in tests since the data is explicitly truncated and it's not to a valid block | |
361 | # size (see https://github.com/ytdl-org/youtube-dl/pull/27660). Tests only care that the correct data downloaded, | |
362 | # not what it decrypts to. | |
363 | if self.params.get('test', False): | |
364 | return frag_content | |
1d3586d0 | 365 | return unpad_pkcs7(aes_cbc_decrypt_bytes(frag_content, decrypt_info['KEY'], iv)) |
1009f67c | 366 | |
367 | return decrypt_fragment | |
368 | ||
bd50a52b THD |
369 | def download_and_append_fragments_multiple(self, *args, pack_func=None, finish_func=None): |
370 | ''' | |
371 | @params (ctx1, fragments1, info_dict1), (ctx2, fragments2, info_dict2), ... | |
372 | all args must be either tuple or list | |
373 | ''' | |
adbc4ec4 | 374 | interrupt_trigger = [True] |
bd50a52b THD |
375 | max_progress = len(args) |
376 | if max_progress == 1: | |
377 | return self.download_and_append_fragments(*args[0], pack_func=pack_func, finish_func=finish_func) | |
adbc4ec4 | 378 | max_workers = self.params.get('concurrent_fragment_downloads', 1) |
49a57e70 | 379 | if max_progress > 1: |
380 | self._prepare_multiline_status(max_progress) | |
a539f065 | 381 | is_live = any(traverse_obj(args, (..., 2, 'is_live'), default=[])) |
bd50a52b THD |
382 | |
383 | def thread_func(idx, ctx, fragments, info_dict, tpe): | |
384 | ctx['max_progress'] = max_progress | |
385 | ctx['progress_idx'] = idx | |
adbc4ec4 THD |
386 | return self.download_and_append_fragments( |
387 | ctx, fragments, info_dict, pack_func=pack_func, finish_func=finish_func, | |
388 | tpe=tpe, interrupt_trigger=interrupt_trigger) | |
bd50a52b THD |
389 | |
390 | class FTPE(concurrent.futures.ThreadPoolExecutor): | |
391 | # has to stop this or it's going to wait on the worker thread itself | |
392 | def __exit__(self, exc_type, exc_val, exc_tb): | |
393 | pass | |
394 | ||
adbc4ec4 | 395 | if compat_os_name == 'nt': |
a44ca5a4 | 396 | def future_result(future): |
a539f065 LNO |
397 | while True: |
398 | try: | |
399 | return future.result(0.1) | |
400 | except KeyboardInterrupt: | |
401 | raise | |
402 | except concurrent.futures.TimeoutError: | |
403 | continue | |
404 | else: | |
a44ca5a4 | 405 | def future_result(future): |
a539f065 LNO |
406 | return future.result() |
407 | ||
f0734e11 L |
408 | def interrupt_trigger_iter(fg): |
409 | for f in fg: | |
410 | if not interrupt_trigger[0]: | |
411 | break | |
412 | yield f | |
413 | ||
a539f065 | 414 | spins = [] |
bd50a52b | 415 | for idx, (ctx, fragments, info_dict) in enumerate(args): |
adbc4ec4 | 416 | tpe = FTPE(math.ceil(max_workers / max_progress)) |
f0734e11 | 417 | job = tpe.submit(thread_func, idx, ctx, interrupt_trigger_iter(fragments), info_dict, tpe) |
bd50a52b THD |
418 | spins.append((tpe, job)) |
419 | ||
420 | result = True | |
421 | for tpe, job in spins: | |
422 | try: | |
a44ca5a4 | 423 | result = result and future_result(job) |
adbc4ec4 THD |
424 | except KeyboardInterrupt: |
425 | interrupt_trigger[0] = False | |
bd50a52b THD |
426 | finally: |
427 | tpe.shutdown(wait=True) | |
a539f065 | 428 | if not interrupt_trigger[0] and not is_live: |
adbc4ec4 | 429 | raise KeyboardInterrupt() |
a539f065 LNO |
430 | # we expect the user wants to stop and DO WANT the preceding postprocessors to run; |
431 | # so returning a intermediate result here instead of KeyboardInterrupt on live | |
819e0531 | 432 | return result |
bd50a52b | 433 | |
adbc4ec4 THD |
434 | def download_and_append_fragments( |
435 | self, ctx, fragments, info_dict, *, pack_func=None, finish_func=None, | |
436 | tpe=None, interrupt_trigger=None): | |
437 | if not interrupt_trigger: | |
438 | interrupt_trigger = (True, ) | |
439 | ||
4c7853de | 440 | fragment_retries = self.params.get('fragment_retries', 0) |
adbc4ec4 THD |
441 | is_fatal = ( |
442 | ((lambda _: False) if info_dict.get('is_live') else (lambda idx: idx == 0)) | |
443 | if self.params.get('skip_unavailable_fragments', True) else (lambda _: True)) | |
444 | ||
4c7853de | 445 | if not pack_func: |
446 | pack_func = lambda frag_content, _: frag_content | |
447 | ||
448 | def download_fragment(fragment, ctx): | |
a539f065 | 449 | if not interrupt_trigger[0]: |
d71fd412 | 450 | return |
a539f065 | 451 | |
4c7853de | 452 | frag_index = ctx['fragment_index'] = fragment['frag_index'] |
185bf310 | 453 | ctx['last_error'] = None |
d9d8b857 | 454 | headers = info_dict.get('http_headers', {}).copy() |
4c7853de | 455 | byte_range = fragment.get('byte_range') |
456 | if byte_range: | |
457 | headers['Range'] = 'bytes=%d-%d' % (byte_range['start'], byte_range['end'] - 1) | |
458 | ||
459 | # Never skip the first fragment | |
d71fd412 | 460 | fatal, count = is_fatal(fragment.get('index') or (frag_index - 1)), 0 |
4c7853de | 461 | while count <= fragment_retries: |
462 | try: | |
d71fd412 LNO |
463 | if self._download_fragment(ctx, fragment['url'], info_dict, headers): |
464 | break | |
465 | return | |
adbc4ec4 | 466 | except (compat_urllib_error.HTTPError, http.client.IncompleteRead) as err: |
4c7853de | 467 | # Unavailable (possibly temporary) fragments may be served. |
468 | # First we try to retry then either skip or abort. | |
469 | # See https://github.com/ytdl-org/youtube-dl/issues/10165, | |
470 | # https://github.com/ytdl-org/youtube-dl/issues/10448). | |
471 | count += 1 | |
185bf310 | 472 | ctx['last_error'] = err |
4c7853de | 473 | if count <= fragment_retries: |
474 | self.report_retry_fragment(err, frag_index, count, fragment_retries) | |
475 | except DownloadError: | |
476 | # Don't retry fragment if error occurred during HTTP downloading | |
477 | # itself since it has own retry settings | |
478 | if not fatal: | |
479 | break | |
480 | raise | |
481 | ||
d71fd412 | 482 | if count > fragment_retries and fatal: |
4c7853de | 483 | ctx['dest_stream'].close() |
484 | self.report_error('Giving up after %s fragment retries' % fragment_retries) | |
4c7853de | 485 | |
4c7853de | 486 | def append_fragment(frag_content, frag_index, ctx): |
a44ca5a4 | 487 | if frag_content: |
488 | self._append_fragment(ctx, pack_func(frag_content, frag_index)) | |
489 | elif not is_fatal(frag_index - 1): | |
490 | self.report_skip_fragment(frag_index, 'fragment not found') | |
491 | else: | |
492 | ctx['dest_stream'].close() | |
493 | self.report_error(f'fragment {frag_index} not found, unable to continue') | |
494 | return False | |
4c7853de | 495 | return True |
496 | ||
1009f67c | 497 | decrypt_fragment = self.decrypter(info_dict) |
498 | ||
adbc4ec4 THD |
499 | max_workers = math.ceil( |
500 | self.params.get('concurrent_fragment_downloads', 1) / ctx.get('max_progress', 1)) | |
1d485a1a | 501 | if max_workers > 1: |
4c7853de | 502 | def _download_fragment(fragment): |
723d44b9 | 503 | ctx_copy = ctx.copy() |
d71fd412 LNO |
504 | download_fragment(fragment, ctx_copy) |
505 | return fragment, fragment['frag_index'], ctx_copy.get('fragment_filename_sanitized') | |
4c7853de | 506 | |
507 | self.report_warning('The download speed shown is only of one thread. This is a known issue and patches are welcome') | |
bd50a52b | 508 | with tpe or concurrent.futures.ThreadPoolExecutor(max_workers) as pool: |
d71fd412 | 509 | for fragment, frag_index, frag_filename in pool.map(_download_fragment, fragments): |
4c7853de | 510 | ctx['fragment_filename_sanitized'] = frag_filename |
511 | ctx['fragment_index'] = frag_index | |
d71fd412 | 512 | result = append_fragment(decrypt_fragment(fragment, self._read_fragment(ctx)), frag_index, ctx) |
4c7853de | 513 | if not result: |
514 | return False | |
515 | else: | |
516 | for fragment in fragments: | |
adbc4ec4 THD |
517 | if not interrupt_trigger[0]: |
518 | break | |
c854208c LNO |
519 | try: |
520 | download_fragment(fragment, ctx) | |
19a03940 | 521 | result = append_fragment( |
522 | decrypt_fragment(fragment, self._read_fragment(ctx)), fragment['frag_index'], ctx) | |
c854208c LNO |
523 | except KeyboardInterrupt: |
524 | if info_dict.get('is_live'): | |
525 | break | |
526 | raise | |
4c7853de | 527 | if not result: |
528 | return False | |
529 | ||
25a3f4f5 F |
530 | if finish_func is not None: |
531 | ctx['dest_stream'].write(finish_func()) | |
532 | ctx['dest_stream'].flush() | |
3ba7740d | 533 | self._finish_frag_download(ctx, info_dict) |
8e897ed2 | 534 | return True |