[yt-dlp.git] / youtube_dl / downloader / http.py

import os
import re
import subprocess
import sys
import time

from .common import FileDownloader
from ..utils import (
    compat_urllib_request,
    compat_urllib_error,
    ContentTooShortError,

    encodeFilename,
    sanitize_open,
    format_bytes,
)


class HttpFD(FileDownloader):
    def real_download(self, filename, info_dict):
        url = info_dict['url']
        tmpfilename = self.temp_name(filename)
        stream = None

        # Do not include the Accept-Encoding header
        headers = {'Youtubedl-no-compression': 'True'}
        if 'user_agent' in info_dict:
            headers['Youtubedl-user-agent'] = info_dict['user_agent']
        basic_request = compat_urllib_request.Request(url, None, headers)
        request = compat_urllib_request.Request(url, None, headers)

        if self.params.get('test', False):
            request.add_header('Range','bytes=0-10240')

        # Establish possible resume length
        if os.path.isfile(encodeFilename(tmpfilename)):
            resume_len = os.path.getsize(encodeFilename(tmpfilename))
        else:
            resume_len = 0

        open_mode = 'wb'
        if resume_len != 0:
            if self.params.get('continuedl', False):
                self.report_resuming_byte(resume_len)
                request.add_header('Range','bytes=%d-' % resume_len)
                open_mode = 'ab'
            else:
                resume_len = 0

        count = 0
        retries = self.params.get('retries', 0)
        while count <= retries:
            # Establish connection
            try:
                if count == 0 and 'urlhandle' in info_dict:
                    data = info_dict['urlhandle']
                data = compat_urllib_request.urlopen(request)
                break
            except (compat_urllib_error.HTTPError, ) as err:
                if (err.code < 500 or err.code >= 600) and err.code != 416:
                    # Unexpected HTTP error
                    raise
                elif err.code == 416:
                    # Unable to resume (requested range not satisfiable)
                    try:
                        # Open the connection again without the range header
                        data = compat_urllib_request.urlopen(basic_request)
                        content_length = data.info()['Content-Length']
                    except (compat_urllib_error.HTTPError, ) as err:
                        if err.code < 500 or err.code >= 600:
                            raise
                    else:
                        # Examine the reported length
                        if (content_length is not None and
                                (resume_len - 100 < int(content_length) < 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(filename)
                            self.try_rename(tmpfilename, filename)
                            self._hook_progress({
                                'filename': filename,
                                'status': 'finished',
                            })
                            return True
                        else:
                            # The length does not match, we start the download over
                            self.report_unable_to_resume()
                            open_mode = 'wb'
                            break
            # Retry
            count += 1
            if count <= retries:
                self.report_retry(count, retries)

        if count > retries:
            self.report_error(u'giving up after %s retries' % retries)
            return False

        data_len = data.info().get('Content-length', None)
        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)
            if min_data_len is not None and data_len < min_data_len:
                self.to_screen(u'\r[download] File is smaller than min-filesize (%s bytes < %s bytes). Aborting.' % (data_len, min_data_len))
                return False
            if max_data_len is not None and data_len > max_data_len:
                self.to_screen(u'\r[download] File is larger than max-filesize (%s bytes > %s bytes). Aborting.' % (data_len, max_data_len))
                return False

        data_len_str = format_bytes(data_len)
        byte_counter = 0 + resume_len
        block_size = self.params.get('buffersize', 1024)
        start = time.time()
        while True:
            # Download and write
            before = time.time()
            data_block = data.read(block_size)
            after = time.time()
            if len(data_block) == 0:
                break
            byte_counter += len(data_block)

            # Open file just in time
            if stream is None:
                try:
                    (stream, tmpfilename) = sanitize_open(tmpfilename, open_mode)
                    assert stream is not None
                    filename = self.undo_temp_name(tmpfilename)
                    self.report_destination(filename)
                except (OSError, IOError) as err:
                    self.report_error(u'unable to open for writing: %s' % str(err))
                    return False
            try:
                stream.write(data_block)
            except (IOError, OSError):
                self.to_stderr(u"\n")
                self.report_error(u'unable to write data: %s' % str(err))
                return False
            if not self.params.get('noresizebuffer', False):
                block_size = self.best_block_size(after - before, len(data_block))

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

            self._hook_progress({
                'downloaded_bytes': byte_counter,
                'total_bytes': data_len,
                'tmpfilename': tmpfilename,
                'filename': filename,
                'status': 'downloading',
                'eta': eta,
                'speed': speed,
            })

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

        if stream is None:
            self.to_stderr(u"\n")
            self.report_error(u'Did not get any data blocks')
            return False
        stream.close()
        self.report_finish(data_len_str, (time.time() - start))
        if data_len is not None and byte_counter != data_len:
            raise ContentTooShortError(byte_counter, int(data_len))
        self.try_rename(tmpfilename, filename)

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

        self._hook_progress({
            'downloaded_bytes': byte_counter,
            'total_bytes': byte_counter,
            'filename': filename,
            'status': 'finished',
        })

        return True
Commit	Line	Data
3bc2ddcc JMF	1	import os
	2	import re
	3	import subprocess
	4	import sys
	5	import time
	6
	7	from .common import FileDownloader
	8	from ..utils import (
	9	compat_urllib_request,
	10	compat_urllib_error,
	11	ContentTooShortError,
	12
	13	encodeFilename,
	14	sanitize_open,
	15	format_bytes,
	16	)
	17
	18
	19	class HttpFD(FileDownloader):
	20	def real_download(self, filename, info_dict):
	21	url = info_dict['url']
	22	tmpfilename = self.temp_name(filename)
	23	stream = None
	24
	25	# Do not include the Accept-Encoding header
	26	headers = {'Youtubedl-no-compression': 'True'}
	27	if 'user_agent' in info_dict:
	28	headers['Youtubedl-user-agent'] = info_dict['user_agent']
	29	basic_request = compat_urllib_request.Request(url, None, headers)
	30	request = compat_urllib_request.Request(url, None, headers)
	31
	32	if self.params.get('test', False):
	33	request.add_header('Range','bytes=0-10240')
	34
	35	# Establish possible resume length
	36	if os.path.isfile(encodeFilename(tmpfilename)):
	37	resume_len = os.path.getsize(encodeFilename(tmpfilename))
	38	else:
	39	resume_len = 0
	40
	41	open_mode = 'wb'
	42	if resume_len != 0:
	43	if self.params.get('continuedl', False):
	44	self.report_resuming_byte(resume_len)
	45	request.add_header('Range','bytes=%d-' % resume_len)
	46	open_mode = 'ab'
	47	else:
	48	resume_len = 0
	49
	50	count = 0
	51	retries = self.params.get('retries', 0)
	52	while count <= retries:
	53	# Establish connection
	54	try:
	55	if count == 0 and 'urlhandle' in info_dict:
	56	data = info_dict['urlhandle']
	57	data = compat_urllib_request.urlopen(request)
	58	break
	59	except (compat_urllib_error.HTTPError, ) as err:
	60	if (err.code < 500 or err.code >= 600) and err.code != 416:
	61	# Unexpected HTTP error
	62	raise
	63	elif err.code == 416:
	64	# Unable to resume (requested range not satisfiable)
65	try:
66	# Open the connection again without the range header
67	data = compat_urllib_request.urlopen(basic_request)
68	content_length = data.info()['Content-Length']
69	except (compat_urllib_error.HTTPError, ) as err:
70	if err.code < 500 or err.code >= 600:
71	raise
72	else:
73	# Examine the reported length
74	if (content_length is not None and
75	(resume_len - 100 < int(content_length) < resume_len + 100)):
76	# The file had already been fully downloaded.
77	# Explanation to the above condition: in issue #175 it was revealed that
78	# YouTube sometimes adds or removes a few bytes from the end of the file,
79	# changing the file size slightly and causing problems for some users. So
80	# I decided to implement a suggested change and consider the file
81	# completely downloaded if the file size differs less than 100 bytes from
82	# the one in the hard drive.
83	self.report_file_already_downloaded(filename)
84	self.try_rename(tmpfilename, filename)
85	self._hook_progress({
86	'filename': filename,
87	'status': 'finished',
88	})
89	return True
90	else:
91	# The length does not match, we start the download over
92	self.report_unable_to_resume()
93	open_mode = 'wb'
94	break
95	# Retry
96	count += 1
97	if count <= retries:
98	self.report_retry(count, retries)
99
100	if count > retries:
101	self.report_error(u'giving up after %s retries' % retries)
102	return False
103
104	data_len = data.info().get('Content-length', None)
105	if data_len is not None:
106	data_len = int(data_len) + resume_len
107	min_data_len = self.params.get("min_filesize", None)
108	max_data_len = self.params.get("max_filesize", None)
109	if min_data_len is not None and data_len < min_data_len:
110	self.to_screen(u'\r[download] File is smaller than min-filesize (%s bytes < %s bytes). Aborting.' % (data_len, min_data_len))
111	return False
112	if max_data_len is not None and data_len > max_data_len:
113	self.to_screen(u'\r[download] File is larger than max-filesize (%s bytes > %s bytes). Aborting.' % (data_len, max_data_len))
114	return False
115
116	data_len_str = format_bytes(data_len)
117	byte_counter = 0 + resume_len
118	block_size = self.params.get('buffersize', 1024)
119	start = time.time()
120	while True:
121	# Download and write
122	before = time.time()
123	data_block = data.read(block_size)
124	after = time.time()
125	if len(data_block) == 0:
126	break
127	byte_counter += len(data_block)
128
129	# Open file just in time
130	if stream is None:
131	try:
132	(stream, tmpfilename) = sanitize_open(tmpfilename, open_mode)
133	assert stream is not None
134	filename = self.undo_temp_name(tmpfilename)
135	self.report_destination(filename)
136	except (OSError, IOError) as err:
137	self.report_error(u'unable to open for writing: %s' % str(err))
138	return False
139	try:
140	stream.write(data_block)
141	except (IOError, OSError):
142	self.to_stderr(u"\n")
143	self.report_error(u'unable to write data: %s' % str(err))
144	return False
145	if not self.params.get('noresizebuffer', False):
146	block_size = self.best_block_size(after - before, len(data_block))
147
148	# Progress message
149	speed = self.calc_speed(start, time.time(), byte_counter - resume_len)
150	if data_len is None:
151	eta = percent = None
152	else:
153	percent = self.calc_percent(byte_counter, data_len)
154	eta = self.calc_eta(start, time.time(), data_len - resume_len, byte_counter - resume_len)
155	self.report_progress(percent, data_len_str, speed, eta)
156
157	self._hook_progress({
158	'downloaded_bytes': byte_counter,
159	'total_bytes': data_len,
160	'tmpfilename': tmpfilename,
161	'filename': filename,
162	'status': 'downloading',
163	'eta': eta,
164	'speed': speed,
165	})
166
167	# Apply rate limit
168	self.slow_down(start, byte_counter - resume_len)
169
170	if stream is None:
171	self.to_stderr(u"\n")
172	self.report_error(u'Did not get any data blocks')
173	return False
174	stream.close()
175	self.report_finish(data_len_str, (time.time() - start))
176	if data_len is not None and byte_counter != data_len:
177	raise ContentTooShortError(byte_counter, int(data_len))
178	self.try_rename(tmpfilename, filename)
179
180	# Update file modification time
181	if self.params.get('updatetime', True):
182	info_dict['filetime'] = self.try_utime(filename, data.info().get('last-modified', None))
183
184	self._hook_progress({
185	'downloaded_bytes': byte_counter,
186	'total_bytes': byte_counter,
187	'filename': filename,
188	'status': 'finished',
189	})
190
191	return True