]>
Commit | Line | Data |
---|---|---|
d77c3dfd FV |
1 | import math |
2 | import os | |
3 | import re | |
d77c3dfd FV |
4 | import subprocess |
5 | import sys | |
6 | import time | |
d77c3dfd | 7 | |
76e67c2c PH |
8 | from .utils import ( |
9 | compat_urllib_error, | |
10 | compat_urllib_request, | |
11 | ContentTooShortError, | |
12 | determine_ext, | |
13 | encodeFilename, | |
14 | sanitize_open, | |
15 | timeconvert, | |
16 | ) | |
d77c3dfd FV |
17 | |
18 | ||
19 | class FileDownloader(object): | |
59ae15a5 PH |
20 | """File Downloader class. |
21 | ||
22 | File downloader objects are the ones responsible of downloading the | |
8222d8de | 23 | actual video file and writing it to disk. |
59ae15a5 PH |
24 | |
25 | File downloaders accept a lot of parameters. In order not to saturate | |
26 | the object constructor with arguments, it receives a dictionary of | |
8222d8de | 27 | options instead. |
59ae15a5 PH |
28 | |
29 | Available options: | |
30 | ||
8222d8de | 31 | verbose: Print additional info to stdout. |
59ae15a5 | 32 | quiet: Do not print messages to stdout. |
59ae15a5 | 33 | ratelimit: Download speed limit, in bytes/sec. |
59ae15a5 PH |
34 | retries: Number of times to retry for HTTP error 5xx |
35 | buffersize: Size of download buffer in bytes. | |
36 | noresizebuffer: Do not automatically resize the download buffer. | |
37 | continuedl: Try to continue downloads if possible. | |
38 | noprogress: Do not print the progress bar. | |
59ae15a5 PH |
39 | logtostderr: Log messages to stderr instead of stdout. |
40 | consoletitle: Display progress in console window's titlebar. | |
41 | nopart: Do not use temporary .part files. | |
42 | updatetime: Use the Last-modified header to set output file timestamps. | |
37c8fd48 | 43 | test: Download only first bytes to test the downloader. |
9e982f9e JC |
44 | min_filesize: Skip files smaller than this size |
45 | max_filesize: Skip files larger than this size | |
59ae15a5 PH |
46 | """ |
47 | ||
48 | params = None | |
59ae15a5 | 49 | |
8222d8de | 50 | def __init__(self, ydl, params): |
59ae15a5 | 51 | """Create a FileDownloader object with the given options.""" |
8222d8de | 52 | self.ydl = ydl |
bffbd5f0 | 53 | self._progress_hooks = [] |
59ae15a5 PH |
54 | self.params = params |
55 | ||
59ae15a5 PH |
56 | @staticmethod |
57 | def format_bytes(bytes): | |
58 | if bytes is None: | |
59 | return 'N/A' | |
60 | if type(bytes) is str: | |
61 | bytes = float(bytes) | |
62 | if bytes == 0.0: | |
63 | exponent = 0 | |
64 | else: | |
65 | exponent = int(math.log(bytes, 1024.0)) | |
b0936ef4 | 66 | suffix = ['B','KiB','MiB','GiB','TiB','PiB','EiB','ZiB','YiB'][exponent] |
59ae15a5 PH |
67 | converted = float(bytes) / float(1024 ** exponent) |
68 | return '%.2f%s' % (converted, suffix) | |
69 | ||
af8bd6a8 JMF |
70 | @staticmethod |
71 | def format_seconds(seconds): | |
72 | (mins, secs) = divmod(seconds, 60) | |
061b2889 | 73 | (hours, mins) = divmod(mins, 60) |
af8bd6a8 JMF |
74 | if hours > 99: |
75 | return '--:--:--' | |
76 | if hours == 0: | |
77 | return '%02d:%02d' % (mins, secs) | |
78 | else: | |
79 | return '%02d:%02d:%02d' % (hours, mins, secs) | |
80 | ||
59ae15a5 PH |
81 | @staticmethod |
82 | def calc_percent(byte_counter, data_len): | |
83 | if data_len is None: | |
4ae72004 JMF |
84 | return None |
85 | return float(byte_counter) / float(data_len) * 100.0 | |
86 | ||
87 | @staticmethod | |
88 | def format_percent(percent): | |
89 | if percent is None: | |
59ae15a5 | 90 | return '---.-%' |
4ae72004 | 91 | return '%6s' % ('%3.1f%%' % percent) |
59ae15a5 PH |
92 | |
93 | @staticmethod | |
94 | def calc_eta(start, now, total, current): | |
95 | if total is None: | |
4ae72004 | 96 | return None |
59ae15a5 PH |
97 | dif = now - start |
98 | if current == 0 or dif < 0.001: # One millisecond | |
4ae72004 | 99 | return None |
59ae15a5 | 100 | rate = float(current) / dif |
4ae72004 JMF |
101 | return int((float(total) - float(current)) / rate) |
102 | ||
103 | @staticmethod | |
104 | def format_eta(eta): | |
105 | if eta is None: | |
106 | return '--:--' | |
af8bd6a8 | 107 | return FileDownloader.format_seconds(eta) |
59ae15a5 PH |
108 | |
109 | @staticmethod | |
110 | def calc_speed(start, now, bytes): | |
111 | dif = now - start | |
112 | if bytes == 0 or dif < 0.001: # One millisecond | |
4ae72004 JMF |
113 | return None |
114 | return float(bytes) / dif | |
115 | ||
116 | @staticmethod | |
117 | def format_speed(speed): | |
118 | if speed is None: | |
59ae15a5 | 119 | return '%10s' % '---b/s' |
4ae72004 | 120 | return '%10s' % ('%s/s' % FileDownloader.format_bytes(speed)) |
59ae15a5 PH |
121 | |
122 | @staticmethod | |
123 | def best_block_size(elapsed_time, bytes): | |
124 | new_min = max(bytes / 2.0, 1.0) | |
125 | new_max = min(max(bytes * 2.0, 1.0), 4194304) # Do not surpass 4 MB | |
126 | if elapsed_time < 0.001: | |
127 | return int(new_max) | |
128 | rate = bytes / elapsed_time | |
129 | if rate > new_max: | |
130 | return int(new_max) | |
131 | if rate < new_min: | |
132 | return int(new_min) | |
133 | return int(rate) | |
134 | ||
135 | @staticmethod | |
136 | def parse_bytes(bytestr): | |
137 | """Parse a string indicating a byte quantity into an integer.""" | |
138 | matchobj = re.match(r'(?i)^(\d+(?:\.\d+)?)([kMGTPEZY]?)$', bytestr) | |
139 | if matchobj is None: | |
140 | return None | |
141 | number = float(matchobj.group(1)) | |
142 | multiplier = 1024.0 ** 'bkmgtpezy'.index(matchobj.group(2).lower()) | |
143 | return int(round(number * multiplier)) | |
144 | ||
8222d8de JMF |
145 | def to_screen(self, *args, **kargs): |
146 | self.ydl.to_screen(*args, **kargs) | |
59ae15a5 PH |
147 | |
148 | def to_stderr(self, message): | |
8222d8de | 149 | self.ydl.to_screen(message) |
59ae15a5 | 150 | |
1e5b9a95 PH |
151 | def to_console_title(self, message): |
152 | self.ydl.to_console_title(message) | |
59ae15a5 | 153 | |
8222d8de JMF |
154 | def trouble(self, *args, **kargs): |
155 | self.ydl.trouble(*args, **kargs) | |
156 | ||
157 | def report_warning(self, *args, **kargs): | |
158 | self.ydl.report_warning(*args, **kargs) | |
159 | ||
160 | def report_error(self, *args, **kargs): | |
2e325280 | 161 | self.ydl.report_error(*args, **kargs) |
4e1582f3 | 162 | |
59ae15a5 PH |
163 | def slow_down(self, start_time, byte_counter): |
164 | """Sleep if the download speed is over the rate limit.""" | |
165 | rate_limit = self.params.get('ratelimit', None) | |
166 | if rate_limit is None or byte_counter == 0: | |
167 | return | |
168 | now = time.time() | |
169 | elapsed = now - start_time | |
170 | if elapsed <= 0.0: | |
171 | return | |
172 | speed = float(byte_counter) / elapsed | |
173 | if speed > rate_limit: | |
174 | time.sleep((byte_counter - rate_limit * (now - start_time)) / rate_limit) | |
175 | ||
176 | def temp_name(self, filename): | |
177 | """Returns a temporary filename for the given filename.""" | |
178 | if self.params.get('nopart', False) or filename == u'-' or \ | |
179 | (os.path.exists(encodeFilename(filename)) and not os.path.isfile(encodeFilename(filename))): | |
180 | return filename | |
181 | return filename + u'.part' | |
182 | ||
183 | def undo_temp_name(self, filename): | |
184 | if filename.endswith(u'.part'): | |
185 | return filename[:-len(u'.part')] | |
186 | return filename | |
187 | ||
188 | def try_rename(self, old_filename, new_filename): | |
189 | try: | |
190 | if old_filename == new_filename: | |
191 | return | |
192 | os.rename(encodeFilename(old_filename), encodeFilename(new_filename)) | |
76e67c2c | 193 | except (IOError, OSError): |
6622d22c | 194 | self.report_error(u'unable to rename file') |
59ae15a5 PH |
195 | |
196 | def try_utime(self, filename, last_modified_hdr): | |
197 | """Try to set the last-modified time of the given file.""" | |
198 | if last_modified_hdr is None: | |
199 | return | |
200 | if not os.path.isfile(encodeFilename(filename)): | |
201 | return | |
202 | timestr = last_modified_hdr | |
203 | if timestr is None: | |
204 | return | |
205 | filetime = timeconvert(timestr) | |
206 | if filetime is None: | |
207 | return filetime | |
bb474376 PH |
208 | # Ignore obviously invalid dates |
209 | if filetime == 0: | |
210 | return | |
59ae15a5 PH |
211 | try: |
212 | os.utime(filename, (time.time(), filetime)) | |
213 | except: | |
214 | pass | |
215 | return filetime | |
216 | ||
59ae15a5 PH |
217 | def report_destination(self, filename): |
218 | """Report destination filename.""" | |
219 | self.to_screen(u'[download] Destination: ' + filename) | |
220 | ||
4ae72004 | 221 | def report_progress(self, percent, data_len_str, speed, eta): |
59ae15a5 PH |
222 | """Report download progress.""" |
223 | if self.params.get('noprogress', False): | |
224 | return | |
4ae9e558 | 225 | clear_line = (u'\x1b[K' if sys.stderr.isatty() and os.name != 'nt' else u'') |
4ac5306a JMF |
226 | if eta is not None: |
227 | eta_str = self.format_eta(eta) | |
228 | else: | |
229 | eta_str = 'Unknown ETA' | |
230 | if percent is not None: | |
231 | percent_str = self.format_percent(percent) | |
232 | else: | |
233 | percent_str = 'Unknown %' | |
4ae72004 | 234 | speed_str = self.format_speed(speed) |
5717d91a | 235 | if self.params.get('progress_with_newline', False): |
1528d664 | 236 | self.to_screen(u'[download] %s of %s at %s ETA %s' % |
7311fef8 | 237 | (percent_str, data_len_str, speed_str, eta_str)) |
5717d91a | 238 | else: |
4ae9e558 PH |
239 | self.to_screen(u'\r%s[download] %s of %s at %s ETA %s' % |
240 | (clear_line, percent_str, data_len_str, speed_str, eta_str), skip_eol=True) | |
1e5b9a95 | 241 | self.to_console_title(u'youtube-dl - %s of %s at %s ETA %s' % |
59ae15a5 PH |
242 | (percent_str.strip(), data_len_str.strip(), speed_str.strip(), eta_str.strip())) |
243 | ||
244 | def report_resuming_byte(self, resume_len): | |
245 | """Report attempt to resume at given byte.""" | |
246 | self.to_screen(u'[download] Resuming download at byte %s' % resume_len) | |
247 | ||
248 | def report_retry(self, count, retries): | |
249 | """Report retry in case of HTTP error 5xx""" | |
250 | self.to_screen(u'[download] Got server HTTP error. Retrying (attempt %d of %d)...' % (count, retries)) | |
251 | ||
252 | def report_file_already_downloaded(self, file_name): | |
253 | """Report file has already been fully downloaded.""" | |
254 | try: | |
255 | self.to_screen(u'[download] %s has already been downloaded' % file_name) | |
76e67c2c | 256 | except UnicodeEncodeError: |
59ae15a5 PH |
257 | self.to_screen(u'[download] The file has already been downloaded') |
258 | ||
259 | def report_unable_to_resume(self): | |
260 | """Report it was impossible to resume download.""" | |
261 | self.to_screen(u'[download] Unable to resume') | |
262 | ||
968b5e01 | 263 | def report_finish(self, data_len_str, tot_time): |
59ae15a5 PH |
264 | """Report download finished.""" |
265 | if self.params.get('noprogress', False): | |
266 | self.to_screen(u'[download] Download completed') | |
267 | else: | |
6d38616e | 268 | clear_line = (u'\x1b[K' if sys.stderr.isatty() and os.name != 'nt' else u'') |
af8bd6a8 JMF |
269 | self.to_screen(u'\r%s[download] 100%% of %s in %s' % |
270 | (clear_line, data_len_str, self.format_seconds(tot_time))) | |
59ae15a5 | 271 | |
31366066 | 272 | def _download_with_rtmpdump(self, filename, url, player_url, page_url, play_path, tc_url, live): |
59ae15a5 PH |
273 | self.report_destination(filename) |
274 | tmpfilename = self.temp_name(filename) | |
9026dd38 | 275 | test = self.params.get('test', False) |
59ae15a5 PH |
276 | |
277 | # Check for rtmpdump first | |
278 | try: | |
967897fd | 279 | subprocess.call(['rtmpdump', '-h'], stdout=(open(os.path.devnull, 'w')), stderr=subprocess.STDOUT) |
59ae15a5 | 280 | except (OSError, IOError): |
6622d22c | 281 | self.report_error(u'RTMP download detected but "rtmpdump" could not be run') |
59ae15a5 | 282 | return False |
8cd252f1 | 283 | verbosity_option = '--verbose' if self.params.get('verbose', False) else '--quiet' |
59ae15a5 PH |
284 | |
285 | # Download using rtmpdump. rtmpdump returns exit code 2 when | |
286 | # the connection was interrumpted and resuming appears to be | |
287 | # possible. This is part of rtmpdump's normal usage, AFAIK. | |
8cd252f1 | 288 | basic_args = ['rtmpdump', verbosity_option, '-r', url, '-o', tmpfilename] |
f5ebb614 | 289 | if player_url is not None: |
8cd252f1 | 290 | basic_args += ['--swfVfy', player_url] |
f5ebb614 PH |
291 | if page_url is not None: |
292 | basic_args += ['--pageUrl', page_url] | |
adb029ed | 293 | if play_path is not None: |
8cd252f1 | 294 | basic_args += ['--playpath', play_path] |
de5d66d4 | 295 | if tc_url is not None: |
296 | basic_args += ['--tcUrl', url] | |
9026dd38 | 297 | if test: |
ad7a071a | 298 | basic_args += ['--stop', '1'] |
31366066 | 299 | if live: |
300 | basic_args += ['--live'] | |
8cd252f1 | 301 | args = basic_args + [[], ['--resume', '--skip', '1']][self.params.get('continuedl', False)] |
59ae15a5 PH |
302 | if self.params.get('verbose', False): |
303 | try: | |
304 | import pipes | |
305 | shell_quote = lambda args: ' '.join(map(pipes.quote, args)) | |
306 | except ImportError: | |
307 | shell_quote = repr | |
308 | self.to_screen(u'[debug] rtmpdump command line: ' + shell_quote(args)) | |
309 | retval = subprocess.call(args) | |
9026dd38 | 310 | while (retval == 2 or retval == 1) and not test: |
59ae15a5 PH |
311 | prevsize = os.path.getsize(encodeFilename(tmpfilename)) |
312 | self.to_screen(u'\r[rtmpdump] %s bytes' % prevsize, skip_eol=True) | |
313 | time.sleep(5.0) # This seems to be needed | |
314 | retval = subprocess.call(basic_args + ['-e'] + [[], ['-k', '1']][retval == 1]) | |
315 | cursize = os.path.getsize(encodeFilename(tmpfilename)) | |
316 | if prevsize == cursize and retval == 1: | |
317 | break | |
318 | # Some rtmp streams seem abort after ~ 99.8%. Don't complain for those | |
319 | if prevsize == cursize and retval == 2 and cursize > 1024: | |
320 | self.to_screen(u'\r[rtmpdump] Could not download the whole video. This can happen for some advertisements.') | |
321 | retval = 0 | |
322 | break | |
9026dd38 | 323 | if retval == 0 or (test and retval == 2): |
bffbd5f0 PH |
324 | fsize = os.path.getsize(encodeFilename(tmpfilename)) |
325 | self.to_screen(u'\r[rtmpdump] %s bytes' % fsize) | |
59ae15a5 | 326 | self.try_rename(tmpfilename, filename) |
bffbd5f0 PH |
327 | self._hook_progress({ |
328 | 'downloaded_bytes': fsize, | |
329 | 'total_bytes': fsize, | |
330 | 'filename': filename, | |
331 | 'status': 'finished', | |
332 | }) | |
59ae15a5 PH |
333 | return True |
334 | else: | |
6622d22c JMF |
335 | self.to_stderr(u"\n") |
336 | self.report_error(u'rtmpdump exited with code %d' % retval) | |
59ae15a5 PH |
337 | return False |
338 | ||
f2cd958c | 339 | def _download_with_mplayer(self, filename, url): |
340 | self.report_destination(filename) | |
341 | tmpfilename = self.temp_name(filename) | |
342 | ||
f2cd958c | 343 | args = ['mplayer', '-really-quiet', '-vo', 'null', '-vc', 'dummy', '-dumpstream', '-dumpfile', tmpfilename, url] |
344 | # Check for mplayer first | |
345 | try: | |
3054ff0c | 346 | subprocess.call(['mplayer', '-h'], stdout=(open(os.path.devnull, 'w')), stderr=subprocess.STDOUT) |
f2cd958c | 347 | except (OSError, IOError): |
348 | self.report_error(u'MMS or RTSP download detected but "%s" could not be run' % args[0] ) | |
349 | return False | |
350 | ||
351 | # Download using mplayer. | |
352 | retval = subprocess.call(args) | |
353 | if retval == 0: | |
354 | fsize = os.path.getsize(encodeFilename(tmpfilename)) | |
355 | self.to_screen(u'\r[%s] %s bytes' % (args[0], fsize)) | |
356 | self.try_rename(tmpfilename, filename) | |
357 | self._hook_progress({ | |
358 | 'downloaded_bytes': fsize, | |
359 | 'total_bytes': fsize, | |
360 | 'filename': filename, | |
361 | 'status': 'finished', | |
362 | }) | |
363 | return True | |
364 | else: | |
365 | self.to_stderr(u"\n") | |
3054ff0c | 366 | self.report_error(u'mplayer exited with code %d' % retval) |
f2cd958c | 367 | return False |
368 | ||
b15d4f62 JMF |
369 | def _download_m3u8_with_ffmpeg(self, filename, url): |
370 | self.report_destination(filename) | |
371 | tmpfilename = self.temp_name(filename) | |
372 | ||
801dbbdf JMF |
373 | args = ['-y', '-i', url, '-f', 'mp4', '-c', 'copy', |
374 | '-bsf:a', 'aac_adtstoasc', tmpfilename] | |
b15d4f62 | 375 | |
801dbbdf JMF |
376 | for program in ['avconv', 'ffmpeg']: |
377 | try: | |
378 | subprocess.call([program, '-version'], stdout=(open(os.path.devnull, 'w')), stderr=subprocess.STDOUT) | |
379 | break | |
380 | except (OSError, IOError): | |
381 | pass | |
382 | else: | |
383 | self.report_error(u'm3u8 download detected but ffmpeg or avconv could not be found') | |
384 | cmd = [program] + args | |
385 | ||
386 | retval = subprocess.call(cmd) | |
b15d4f62 JMF |
387 | if retval == 0: |
388 | fsize = os.path.getsize(encodeFilename(tmpfilename)) | |
389 | self.to_screen(u'\r[%s] %s bytes' % (args[0], fsize)) | |
390 | self.try_rename(tmpfilename, filename) | |
391 | self._hook_progress({ | |
392 | 'downloaded_bytes': fsize, | |
393 | 'total_bytes': fsize, | |
394 | 'filename': filename, | |
395 | 'status': 'finished', | |
396 | }) | |
397 | return True | |
398 | else: | |
399 | self.to_stderr(u"\n") | |
400 | self.report_error(u'ffmpeg exited with code %d' % retval) | |
401 | return False | |
402 | ||
f2cd958c | 403 | |
59ae15a5 PH |
404 | def _do_download(self, filename, info_dict): |
405 | url = info_dict['url'] | |
59ae15a5 PH |
406 | |
407 | # Check file already present | |
408 | if self.params.get('continuedl', False) and os.path.isfile(encodeFilename(filename)) and not self.params.get('nopart', False): | |
409 | self.report_file_already_downloaded(filename) | |
bffbd5f0 PH |
410 | self._hook_progress({ |
411 | 'filename': filename, | |
412 | 'status': 'finished', | |
dd5d2eb0 | 413 | 'total_bytes': os.path.getsize(encodeFilename(filename)), |
bffbd5f0 | 414 | }) |
59ae15a5 PH |
415 | return True |
416 | ||
417 | # Attempt to download using rtmpdump | |
418 | if url.startswith('rtmp'): | |
f5ebb614 PH |
419 | return self._download_with_rtmpdump(filename, url, |
420 | info_dict.get('player_url', None), | |
adb029ed | 421 | info_dict.get('page_url', None), |
de5d66d4 | 422 | info_dict.get('play_path', None), |
31366066 | 423 | info_dict.get('tc_url', None), |
0ed05a1d | 424 | info_dict.get('rtmp_live', False)) |
59ae15a5 | 425 | |
f2cd958c | 426 | # Attempt to download using mplayer |
427 | if url.startswith('mms') or url.startswith('rtsp'): | |
428 | return self._download_with_mplayer(filename, url) | |
429 | ||
b15d4f62 JMF |
430 | # m3u8 manifest are downloaded with ffmpeg |
431 | if determine_ext(url) == u'm3u8': | |
432 | return self._download_m3u8_with_ffmpeg(filename, url) | |
433 | ||
59ae15a5 PH |
434 | tmpfilename = self.temp_name(filename) |
435 | stream = None | |
436 | ||
437 | # Do not include the Accept-Encoding header | |
438 | headers = {'Youtubedl-no-compression': 'True'} | |
3446dfb7 PH |
439 | if 'user_agent' in info_dict: |
440 | headers['Youtubedl-user-agent'] = info_dict['user_agent'] | |
59ae15a5 PH |
441 | basic_request = compat_urllib_request.Request(url, None, headers) |
442 | request = compat_urllib_request.Request(url, None, headers) | |
443 | ||
37c8fd48 FV |
444 | if self.params.get('test', False): |
445 | request.add_header('Range','bytes=0-10240') | |
446 | ||
59ae15a5 PH |
447 | # Establish possible resume length |
448 | if os.path.isfile(encodeFilename(tmpfilename)): | |
449 | resume_len = os.path.getsize(encodeFilename(tmpfilename)) | |
450 | else: | |
451 | resume_len = 0 | |
452 | ||
453 | open_mode = 'wb' | |
454 | if resume_len != 0: | |
455 | if self.params.get('continuedl', False): | |
456 | self.report_resuming_byte(resume_len) | |
457 | request.add_header('Range','bytes=%d-' % resume_len) | |
458 | open_mode = 'ab' | |
459 | else: | |
460 | resume_len = 0 | |
461 | ||
462 | count = 0 | |
463 | retries = self.params.get('retries', 0) | |
464 | while count <= retries: | |
465 | # Establish connection | |
466 | try: | |
467 | if count == 0 and 'urlhandle' in info_dict: | |
468 | data = info_dict['urlhandle'] | |
469 | data = compat_urllib_request.urlopen(request) | |
470 | break | |
471 | except (compat_urllib_error.HTTPError, ) as err: | |
472 | if (err.code < 500 or err.code >= 600) and err.code != 416: | |
473 | # Unexpected HTTP error | |
474 | raise | |
475 | elif err.code == 416: | |
476 | # Unable to resume (requested range not satisfiable) | |
477 | try: | |
478 | # Open the connection again without the range header | |
479 | data = compat_urllib_request.urlopen(basic_request) | |
480 | content_length = data.info()['Content-Length'] | |
481 | except (compat_urllib_error.HTTPError, ) as err: | |
482 | if err.code < 500 or err.code >= 600: | |
483 | raise | |
484 | else: | |
485 | # Examine the reported length | |
486 | if (content_length is not None and | |
487 | (resume_len - 100 < int(content_length) < resume_len + 100)): | |
488 | # The file had already been fully downloaded. | |
489 | # Explanation to the above condition: in issue #175 it was revealed that | |
490 | # YouTube sometimes adds or removes a few bytes from the end of the file, | |
491 | # changing the file size slightly and causing problems for some users. So | |
492 | # I decided to implement a suggested change and consider the file | |
493 | # completely downloaded if the file size differs less than 100 bytes from | |
494 | # the one in the hard drive. | |
495 | self.report_file_already_downloaded(filename) | |
496 | self.try_rename(tmpfilename, filename) | |
bffbd5f0 PH |
497 | self._hook_progress({ |
498 | 'filename': filename, | |
499 | 'status': 'finished', | |
500 | }) | |
59ae15a5 PH |
501 | return True |
502 | else: | |
503 | # The length does not match, we start the download over | |
504 | self.report_unable_to_resume() | |
505 | open_mode = 'wb' | |
506 | break | |
507 | # Retry | |
508 | count += 1 | |
509 | if count <= retries: | |
510 | self.report_retry(count, retries) | |
511 | ||
512 | if count > retries: | |
6622d22c | 513 | self.report_error(u'giving up after %s retries' % retries) |
59ae15a5 PH |
514 | return False |
515 | ||
516 | data_len = data.info().get('Content-length', None) | |
517 | if data_len is not None: | |
518 | data_len = int(data_len) + resume_len | |
9e982f9e JC |
519 | min_data_len = self.params.get("min_filesize", None) |
520 | max_data_len = self.params.get("max_filesize", None) | |
521 | if min_data_len is not None and data_len < min_data_len: | |
522 | self.to_screen(u'\r[download] File is smaller than min-filesize (%s bytes < %s bytes). Aborting.' % (data_len, min_data_len)) | |
523 | return False | |
524 | if max_data_len is not None and data_len > max_data_len: | |
525 | self.to_screen(u'\r[download] File is larger than max-filesize (%s bytes > %s bytes). Aborting.' % (data_len, max_data_len)) | |
526 | return False | |
527 | ||
59ae15a5 PH |
528 | data_len_str = self.format_bytes(data_len) |
529 | byte_counter = 0 + resume_len | |
530 | block_size = self.params.get('buffersize', 1024) | |
531 | start = time.time() | |
532 | while True: | |
533 | # Download and write | |
534 | before = time.time() | |
535 | data_block = data.read(block_size) | |
536 | after = time.time() | |
537 | if len(data_block) == 0: | |
538 | break | |
539 | byte_counter += len(data_block) | |
540 | ||
541 | # Open file just in time | |
542 | if stream is None: | |
543 | try: | |
544 | (stream, tmpfilename) = sanitize_open(tmpfilename, open_mode) | |
545 | assert stream is not None | |
546 | filename = self.undo_temp_name(tmpfilename) | |
547 | self.report_destination(filename) | |
548 | except (OSError, IOError) as err: | |
6622d22c | 549 | self.report_error(u'unable to open for writing: %s' % str(err)) |
59ae15a5 PH |
550 | return False |
551 | try: | |
552 | stream.write(data_block) | |
553 | except (IOError, OSError) as err: | |
6622d22c JMF |
554 | self.to_stderr(u"\n") |
555 | self.report_error(u'unable to write data: %s' % str(err)) | |
59ae15a5 PH |
556 | return False |
557 | if not self.params.get('noresizebuffer', False): | |
558 | block_size = self.best_block_size(after - before, len(data_block)) | |
559 | ||
560 | # Progress message | |
4ae72004 | 561 | speed = self.calc_speed(start, time.time(), byte_counter - resume_len) |
59ae15a5 | 562 | if data_len is None: |
4ac5306a | 563 | eta = percent = None |
59ae15a5 | 564 | else: |
4ae72004 JMF |
565 | percent = self.calc_percent(byte_counter, data_len) |
566 | eta = self.calc_eta(start, time.time(), data_len - resume_len, byte_counter - resume_len) | |
4ac5306a | 567 | self.report_progress(percent, data_len_str, speed, eta) |
59ae15a5 | 568 | |
bffbd5f0 PH |
569 | self._hook_progress({ |
570 | 'downloaded_bytes': byte_counter, | |
571 | 'total_bytes': data_len, | |
572 | 'tmpfilename': tmpfilename, | |
573 | 'filename': filename, | |
574 | 'status': 'downloading', | |
4ae72004 JMF |
575 | 'eta': eta, |
576 | 'speed': speed, | |
bffbd5f0 PH |
577 | }) |
578 | ||
59ae15a5 PH |
579 | # Apply rate limit |
580 | self.slow_down(start, byte_counter - resume_len) | |
581 | ||
582 | if stream is None: | |
6622d22c JMF |
583 | self.to_stderr(u"\n") |
584 | self.report_error(u'Did not get any data blocks') | |
59ae15a5 PH |
585 | return False |
586 | stream.close() | |
968b5e01 | 587 | self.report_finish(data_len_str, (time.time() - start)) |
59ae15a5 PH |
588 | if data_len is not None and byte_counter != data_len: |
589 | raise ContentTooShortError(byte_counter, int(data_len)) | |
590 | self.try_rename(tmpfilename, filename) | |
591 | ||
592 | # Update file modification time | |
593 | if self.params.get('updatetime', True): | |
594 | info_dict['filetime'] = self.try_utime(filename, data.info().get('last-modified', None)) | |
595 | ||
bffbd5f0 PH |
596 | self._hook_progress({ |
597 | 'downloaded_bytes': byte_counter, | |
598 | 'total_bytes': byte_counter, | |
599 | 'filename': filename, | |
600 | 'status': 'finished', | |
601 | }) | |
602 | ||
59ae15a5 | 603 | return True |
bffbd5f0 PH |
604 | |
605 | def _hook_progress(self, status): | |
606 | for ph in self._progress_hooks: | |
607 | ph(status) | |
608 | ||
609 | def add_progress_hook(self, ph): | |
610 | """ ph gets called on download progress, with a dictionary with the entries | |
611 | * filename: The final filename | |
612 | * status: One of "downloading" and "finished" | |
613 | ||
614 | It can also have some of the following entries: | |
615 | ||
616 | * downloaded_bytes: Bytes on disks | |
617 | * total_bytes: Total bytes, None if unknown | |
618 | * tmpfilename: The filename we're currently writing to | |
4ae72004 JMF |
619 | * eta: The estimated time in seconds, None if unknown |
620 | * speed: The download speed in bytes/second, None if unknown | |
bffbd5f0 PH |
621 | |
622 | Hooks are guaranteed to be called at least once (with status "finished") | |
623 | if the download is successful. | |
624 | """ | |
625 | self._progress_hooks.append(ph) |