summaryrefslogtreecommitdiffstats
path: root/yt_dlp/downloader/http.py
blob: 693828b6e5e71491648bbcd0e278ebf0f397542d (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
import os
import random
import time

from .common import FileDownloader
from ..networking import Request
from ..networking.exceptions import (
    CertificateVerifyError,
    HTTPError,
    TransportError,
)
from ..utils import (
    ContentTooShortError,
    RetryManager,
    ThrottledDownload,
    XAttrMetadataError,
    XAttrUnavailableError,
    encodeFilename,
    int_or_none,
    parse_http_range,
    try_call,
    write_xattr,
)
from ..utils.networking import HTTPHeaderDict


class HttpFD(FileDownloader):
    def real_download(self, filename, info_dict):
        url = info_dict['url']
        request_data = info_dict.get('request_data', None)

        class DownloadContext(dict):
            __getattr__ = dict.get
            __setattr__ = dict.__setitem__
            __delattr__ = dict.__delitem__

        ctx = DownloadContext()
        ctx.filename = filename
        ctx.tmpfilename = self.temp_name(filename)
        ctx.stream = None

        # Disable compression
        headers = HTTPHeaderDict({'Accept-Encoding': 'identity'}, info_dict.get('http_headers'))

        is_test = self.params.get('test', False)
        chunk_size = self._TEST_FILE_SIZE if is_test else (
            self.params.get('http_chunk_size')
            or info_dict.get('downloader_options', {}).get('http_chunk_size')
            or 0)

        ctx.open_mode = 'wb'
        ctx.resume_len = 0
        ctx.block_size = self.params.get('buffersize', 1024)
        ctx.start_time = time.time()

        # parse given Range
        req_start, req_end, _ = parse_http_range(headers.get('Range'))

        if self.params.get('continuedl', True):
            # Establish possible resume length
            if os.path.isfile(encodeFilename(ctx.tmpfilename)):
                ctx.resume_len = os.path.getsize(
                    encodeFilename(ctx.tmpfilename))

        ctx.is_resume = ctx.resume_len > 0

        class SucceedDownload(Exception):
            pass

        class RetryDownload(Exception):
            def __init__(self, source_error):
                self.source_error = source_error

        class NextFragment(Exception):
            pass

        def establish_connection():
            ctx.chunk_size = (random.randint(int(chunk_size * 0.95), chunk_size)
                              if not is_test and chunk_size else chunk_size)
            if ctx.resume_len > 0:
                range_start = ctx.resume_len
                if req_start is not None:
                    # offset the beginning of Range to be within request
                    range_start += req_start
                if ctx.is_resume:
                    self.report_resuming_byte(ctx.resume_len)
                ctx.open_mode = 'ab'
            elif req_start is not None:
                range_start = req_start
            elif ctx.chunk_size > 0:
                range_start = 0
            else:
                range_start = None
            ctx.is_resume = False

            if ctx.chunk_size:
                chunk_aware_end = range_start + ctx.chunk_size - 1
                # we're not allowed to download outside Range
                range_end = chunk_aware_end if req_end is None else min(chunk_aware_end, req_end)
            elif req_end is not None:
                # there's no need for chunked downloads, so download until the end of Range
                range_end = req_end
            else:
                range_end = None

            if try_call(lambda: range_start > range_end):
                ctx.resume_len = 0
                ctx.open_mode = 'wb'
                raise RetryDownload(Exception(f'Conflicting range. (start={range_start} > end={range_end})'))

            if try_call(lambda: range_end >= ctx.content_len):
                range_end = ctx.content_len - 1

            request = Request(url, request_data, headers)
            has_range = range_start is not None
            if has_range:
                request.headers['Range'] = f'bytes={int(range_start)}-{int_or_none(range_end) or ""}'
            # Establish connection
            try:
                ctx.data = self.ydl.urlopen(request)
                # When trying to resume, Content-Range HTTP header of response has to be checked
                # to match the value of requested Range HTTP header. This is due to a webservers
                # that don't support resuming and serve a whole file with no Content-Range
                # set in response despite of requested Range (see
                # https://github.com/ytdl-org/youtube-dl/issues/6057#issuecomment-126129799)
                if has_range:
                    content_range = ctx.data.headers.get('Content-Range')
                    content_range_start, content_range_end, content_len = parse_http_range(content_range)
                    # 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):
                        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
                    elif range_start > 0:
                        self.report_unable_to_resume()
                    ctx.resume_len = 0
                    ctx.open_mode = 'wb'
                ctx.data_len = ctx.content_len = int_or_none(ctx.data.headers.get('Content-length', None))
            except HTTPError as err:
                if err.status == 416:
                    # Unable to resume (requested range not satisfiable)
                    try:
                        # Open the connection again without the range header
                        ctx.data = self.ydl.urlopen(
                            Request(url, request_data, headers))
                        content_length = ctx.data.headers['Content-Length']
                    except HTTPError as err:
                        if err.status < 500 or err.status >= 600:
                            raise
                    else:
                        # Examine the reported length
                        if (content_length is not None
                                and (ctx.resume_len - 100 < int(content_length) < ctx.resume_len + 100)):
                            # The file had already been fully downloaded.
                            # Explanation to the above condition: in issue #175 it was revealed that
                            # YouTube sometimes adds or removes a few bytes from the end of the file,
                            # changing the file size slightly and causing problems for some users. So
                            # I decided to implement a suggested change and consider the file
                            # completely downloaded if the file size differs less than 100 bytes from
                            # the one in the hard drive.
                            self.report_file_already_downloaded(ctx.filename)
                            self.try_rename(ctx.tmpfilename, ctx.filename)
                            self._hook_progress({
                                'filename': ctx.filename,
                                'status': 'finished',
                                'downloaded_bytes': ctx.resume_len,
                                'total_bytes': ctx.resume_len,
                            }, info_dict)
                            raise SucceedDownload()
                        else:
                            # The length does not match, we start the download over
                            self.report_unable_to_resume()
                            ctx.resume_len = 0
                            ctx.open_mode = 'wb'
                            return
                elif err.status < 500 or err.status >= 600:
                    # Unexpected HTTP error
                    raise
                raise RetryDownload(err)
            except CertificateVerifyError:
                raise
            except TransportError 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.headers.get('Content-length')

            if ctx.data.headers.get('Content-encoding'):
                # Content-encoding is present, Content-length is not reliable anymore as we are
                # doing auto decompression. (See: https://github.com/yt-dlp/yt-dlp/pull/6176)
                data_len = None

            # Range HTTP header may be ignored/unsupported by a webserver
            # (e.g. extractor/scivee.py, extractor/bambuser.py).
            # However, for a test we still would like to download just a piece of a file.
            # To achieve this we limit data_len to _TEST_FILE_SIZE and manually control
            # block size when downloading a file.
            if is_test and (data_len is None or int(data_len) > self._TEST_FILE_SIZE):
                data_len = self._TEST_FILE_SIZE

            if data_len is not None:
                data_len = int(data_len) + ctx.resume_len
                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(
                        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(
                        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
            block_size = ctx.block_size
            start = time.time()

            # measure time over whole while-loop, so slow_down() and best_block_size() work together properly
            now = None  # needed for slow_down() in the first loop run
            before = start  # start measuring

            def retry(e):
                close_stream()
                if ctx.tmpfilename == '-':
                    ctx.resume_len = byte_counter
                else:
                    try:
                        ctx.resume_len = os.path.getsize(encodeFilename(ctx.tmpfilename))
                    except FileNotFoundError:
                        ctx.resume_len = 0
                raise RetryDownload(e)

            while True:
                try:
                    # Download and write
                    data_block = ctx.data.read(block_size if not is_test else min(block_size, data_len - byte_counter))
                except TransportError as err:
                    retry(err)

                byte_counter += len(data_block)

                # exit loop when download is finished
                if len(data_block) == 0:
                    break

                # Open destination file just in time
                if ctx.stream is None:
                    try:
                        ctx.stream, ctx.tmpfilename = self.sanitize_open(
                            ctx.tmpfilename, ctx.open_mode)
                        assert ctx.stream is not None
                        ctx.filename = self.undo_temp_name(ctx.tmpfilename)
                        self.report_destination(ctx.filename)
                    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())
                        except (XAttrUnavailableError, XAttrMetadataError) as err:
                            self.report_error('unable to set filesize xattr: %s' % str(err))

                try:
                    ctx.stream.write(data_block)
                except OSError as err:
                    self.to_stderr('\n')
                    self.report_error('unable to write data: %s' % str(err))
                    return False

                # Apply rate limit
                self.slow_down(start, now, byte_counter - ctx.resume_len)

                # end measuring of one loop run
                now = time.time()
                after = now

                # Adjust block size
                if not self.params.get('noresizebuffer', False):
                    block_size = self.best_block_size(after - before, len(data_block))

                before = after

                # Progress message
                speed = self.calc_speed(start, now, byte_counter - ctx.resume_len)
                if ctx.data_len is None:
                    eta = None
                else:
                    eta = self.calc_eta(start, time.time(), ctx.data_len - ctx.resume_len, byte_counter - ctx.resume_len)

                self._hook_progress({
                    'status': 'downloading',
                    'downloaded_bytes': byte_counter,
                    'total_bytes': ctx.data_len,
                    'tmpfilename': ctx.tmpfilename,
                    'filename': ctx.filename,
                    'eta': eta,
                    'speed': speed,
                    'elapsed': now - ctx.start_time,
                    'ctx_id': info_dict.get('ctx_id'),
                }, info_dict)

                if data_len is not None and byte_counter == data_len:
                    break

                if speed and speed < (self.params.get('throttledratelimit') or 0):
                    # The speed must stay below the limit for 3 seconds
                    # This prevents raising error when the speed temporarily goes down
                    if ctx.throttle_start is None:
                        ctx.throttle_start = now
                    elif now - ctx.throttle_start > 3:
                        if ctx.stream is not None and ctx.tmpfilename != '-':
                            ctx.stream.close()
                        raise ThrottledDownload()
                elif speed:
                    ctx.throttle_start = None

            if ctx.stream is None:
                self.to_stderr('\n')
                self.report_error('Did not get any data blocks')
                return False

            if not is_test and ctx.chunk_size and ctx.content_len is not None and byte_counter < ctx.content_len:
                ctx.resume_len = byte_counter
                raise NextFragment()

            if ctx.tmpfilename != '-':
                ctx.stream.close()

            if data_len is not None and byte_counter != data_len:
                err = ContentTooShortError(byte_counter, int(data_len))
                retry(err)

            self.try_rename(ctx.tmpfilename, ctx.filename)

            # Update file modification time
            if self.params.get('updatetime', True):
                info_dict['filetime'] = self.try_utime(ctx.filename, ctx.data.headers.get('last-modified', None))

            self._hook_progress({
                'downloaded_bytes': byte_counter,
                'total_bytes': byte_counter,
                'filename': ctx.filename,
                'status': 'finished',
                'elapsed': time.time() - ctx.start_time,
                'ctx_id': info_dict.get('ctx_id'),
            }, info_dict)

            return True

        for retry in RetryManager(self.params.get('retries'), self.report_retry):
            try:
                establish_connection()
                return download()
            except RetryDownload as err:
                retry.error = err.source_error
                continue
            except NextFragment:
                retry.error = None
                retry.attempt -= 1
                continue
            except SucceedDownload:
                return True
            except:  # noqa: E722
                close_stream()
                raise
        return False