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
|
from __future__ import annotations
import io
import math
import urllib.parse
from ._helper import InstanceStoreMixin, select_proxy
from .common import (
Features,
Request,
Response,
register_preference,
register_rh,
)
from .exceptions import (
CertificateVerifyError,
HTTPError,
IncompleteRead,
ProxyError,
SSLError,
TransportError,
)
from .impersonate import ImpersonateRequestHandler, ImpersonateTarget
from ..dependencies import curl_cffi, certifi
from ..utils import int_or_none
if curl_cffi is None:
raise ImportError('curl_cffi is not installed')
curl_cffi_version = tuple(int_or_none(x, default=0) for x in curl_cffi.__version__.split('.'))
if curl_cffi_version != (0, 5, 10):
curl_cffi._yt_dlp__version = f'{curl_cffi.__version__} (unsupported)'
raise ImportError('Only curl_cffi 0.5.10 is supported')
import curl_cffi.requests
from curl_cffi.const import CurlECode, CurlOpt
class CurlCFFIResponseReader(io.IOBase):
def __init__(self, response: curl_cffi.requests.Response):
self._response = response
self._iterator = response.iter_content()
self._buffer = b''
self.bytes_read = 0
def readable(self):
return True
def read(self, size=None):
exception_raised = True
try:
while self._iterator and (size is None or len(self._buffer) < size):
chunk = next(self._iterator, None)
if chunk is None:
self._iterator = None
break
self._buffer += chunk
self.bytes_read += len(chunk)
if size is None:
size = len(self._buffer)
data = self._buffer[:size]
self._buffer = self._buffer[size:]
# "free" the curl instance if the response is fully read.
# curl_cffi doesn't do this automatically and only allows one open response per thread
if not self._iterator and not self._buffer:
self.close()
exception_raised = False
return data
finally:
if exception_raised:
self.close()
def close(self):
if not self.closed:
self._response.close()
self._buffer = b''
super().close()
class CurlCFFIResponseAdapter(Response):
fp: CurlCFFIResponseReader
def __init__(self, response: curl_cffi.requests.Response):
super().__init__(
fp=CurlCFFIResponseReader(response),
headers=response.headers,
url=response.url,
status=response.status_code)
def read(self, amt=None):
try:
return self.fp.read(amt)
except curl_cffi.requests.errors.RequestsError as e:
if e.code == CurlECode.PARTIAL_FILE:
content_length = int_or_none(e.response.headers.get('Content-Length'))
raise IncompleteRead(
partial=self.fp.bytes_read,
expected=content_length - self.fp.bytes_read if content_length is not None else None,
cause=e) from e
raise TransportError(cause=e) from e
@register_rh
class CurlCFFIRH(ImpersonateRequestHandler, InstanceStoreMixin):
RH_NAME = 'curl_cffi'
_SUPPORTED_URL_SCHEMES = ('http', 'https')
_SUPPORTED_FEATURES = (Features.NO_PROXY, Features.ALL_PROXY)
_SUPPORTED_PROXY_SCHEMES = ('http', 'https', 'socks4', 'socks4a', 'socks5', 'socks5h')
_SUPPORTED_IMPERSONATE_TARGET_MAP = {
ImpersonateTarget('chrome', '110', 'windows', '10'): curl_cffi.requests.BrowserType.chrome110,
ImpersonateTarget('chrome', '107', 'windows', '10'): curl_cffi.requests.BrowserType.chrome107,
ImpersonateTarget('chrome', '104', 'windows', '10'): curl_cffi.requests.BrowserType.chrome104,
ImpersonateTarget('chrome', '101', 'windows', '10'): curl_cffi.requests.BrowserType.chrome101,
ImpersonateTarget('chrome', '100', 'windows', '10'): curl_cffi.requests.BrowserType.chrome100,
ImpersonateTarget('chrome', '99', 'windows', '10'): curl_cffi.requests.BrowserType.chrome99,
ImpersonateTarget('edge', '101', 'windows', '10'): curl_cffi.requests.BrowserType.edge101,
ImpersonateTarget('edge', '99', 'windows', '10'): curl_cffi.requests.BrowserType.edge99,
ImpersonateTarget('safari', '15.5', 'macos', '12'): curl_cffi.requests.BrowserType.safari15_5,
ImpersonateTarget('safari', '15.3', 'macos', '11'): curl_cffi.requests.BrowserType.safari15_3,
ImpersonateTarget('chrome', '99', 'android', '12'): curl_cffi.requests.BrowserType.chrome99_android,
}
def _create_instance(self, cookiejar=None):
return curl_cffi.requests.Session(cookies=cookiejar)
def _check_extensions(self, extensions):
super()._check_extensions(extensions)
extensions.pop('impersonate', None)
extensions.pop('cookiejar', None)
extensions.pop('timeout', None)
def send(self, request: Request) -> Response:
target = self._get_request_target(request)
try:
response = super().send(request)
except HTTPError as e:
e.response.extensions['impersonate'] = target
raise
response.extensions['impersonate'] = target
return response
def _send(self, request: Request):
max_redirects_exceeded = False
session: curl_cffi.requests.Session = self._get_instance(
cookiejar=self._get_cookiejar(request) if 'cookie' not in request.headers else None)
if self.verbose:
session.curl.setopt(CurlOpt.VERBOSE, 1)
proxies = self._get_proxies(request)
if 'no' in proxies:
session.curl.setopt(CurlOpt.NOPROXY, proxies['no'])
proxies.pop('no', None)
# curl doesn't support per protocol proxies, so we select the one that matches the request protocol
proxy = select_proxy(request.url, proxies=proxies)
if proxy:
session.curl.setopt(CurlOpt.PROXY, proxy)
scheme = urllib.parse.urlparse(request.url).scheme.lower()
if scheme != 'http':
# Enable HTTP CONNECT for HTTPS urls.
# Don't use CONNECT for http for compatibility with urllib behaviour.
# See: https://curl.se/libcurl/c/CURLOPT_HTTPPROXYTUNNEL.html
session.curl.setopt(CurlOpt.HTTPPROXYTUNNEL, 1)
# curl_cffi does not currently set these for proxies
session.curl.setopt(CurlOpt.PROXY_CAINFO, certifi.where())
if not self.verify:
session.curl.setopt(CurlOpt.PROXY_SSL_VERIFYPEER, 0)
session.curl.setopt(CurlOpt.PROXY_SSL_VERIFYHOST, 0)
headers = self._get_impersonate_headers(request)
if self._client_cert:
session.curl.setopt(CurlOpt.SSLCERT, self._client_cert['client_certificate'])
client_certificate_key = self._client_cert.get('client_certificate_key')
client_certificate_password = self._client_cert.get('client_certificate_password')
if client_certificate_key:
session.curl.setopt(CurlOpt.SSLKEY, client_certificate_key)
if client_certificate_password:
session.curl.setopt(CurlOpt.KEYPASSWD, client_certificate_password)
timeout = self._calculate_timeout(request)
# set CURLOPT_LOW_SPEED_LIMIT and CURLOPT_LOW_SPEED_TIME to act as a read timeout. [1]
# curl_cffi does not currently do this. [2]
# Note: CURLOPT_LOW_SPEED_TIME is in seconds, so we need to round up to the nearest second. [3]
# [1] https://unix.stackexchange.com/a/305311
# [2] https://github.com/yifeikong/curl_cffi/issues/156
# [3] https://curl.se/libcurl/c/CURLOPT_LOW_SPEED_TIME.html
session.curl.setopt(CurlOpt.LOW_SPEED_LIMIT, 1) # 1 byte per second
session.curl.setopt(CurlOpt.LOW_SPEED_TIME, math.ceil(timeout))
try:
curl_response = session.request(
method=request.method,
url=request.url,
headers=headers,
data=request.data,
verify=self.verify,
max_redirects=5,
timeout=timeout,
impersonate=self._SUPPORTED_IMPERSONATE_TARGET_MAP.get(
self._get_request_target(request)),
interface=self.source_address,
stream=True,
)
except curl_cffi.requests.errors.RequestsError as e:
if e.code == CurlECode.PEER_FAILED_VERIFICATION:
raise CertificateVerifyError(cause=e) from e
elif e.code == CurlECode.SSL_CONNECT_ERROR:
raise SSLError(cause=e) from e
elif e.code == CurlECode.TOO_MANY_REDIRECTS:
max_redirects_exceeded = True
curl_response = e.response
elif (
e.code == CurlECode.PROXY
or (e.code == CurlECode.RECV_ERROR and 'Received HTTP code 407 from proxy after CONNECT' in str(e))
):
raise ProxyError(cause=e) from e
else:
raise TransportError(cause=e) from e
response = CurlCFFIResponseAdapter(curl_response)
if not 200 <= response.status < 300:
raise HTTPError(response, redirect_loop=max_redirects_exceeded)
return response
@register_preference(CurlCFFIRH)
def curl_cffi_preference(rh, request):
return -100
|
