diff --git a/pyproject.toml b/pyproject.toml index b5d77d35b..5e987a6fd 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -55,8 +55,7 @@ default = [ "websockets>=13.0", ] curl-cffi = [ - "curl-cffi==0.5.10; os_name=='nt' and implementation_name=='cpython'", - "curl-cffi>=0.5.10,!=0.6.*,<0.7.2; os_name!='nt' and implementation_name=='cpython'", + "curl-cffi>=0.5.10,!=0.6.*,!=0.7.*,!=0.8.*,!=0.9.*,<0.11; implementation_name=='cpython'", ] secretstorage = [ "cffi", diff --git a/test/test_networking.py b/test/test_networking.py index 63914bc4b..3ab60fe83 100644 --- a/test/test_networking.py +++ b/test/test_networking.py @@ -614,7 +614,6 @@ class TestHTTPRequestHandler(TestRequestHandlerBase): rh, Request(f'http://127.0.0.1:{self.http_port}/source_address')).read().decode() assert source_address == data - # Not supported by CurlCFFI @pytest.mark.skip_handler('CurlCFFI', 'not supported by curl-cffi') def test_gzip_trailing_garbage(self, handler): with handler() as rh: diff --git a/yt_dlp/YoutubeDL.py b/yt_dlp/YoutubeDL.py index 6c0b12e66..63e6e11b2 100644 --- a/yt_dlp/YoutubeDL.py +++ b/yt_dlp/YoutubeDL.py @@ -4152,7 +4152,7 @@ class YoutubeDL: (target, rh.RH_NAME) for rh in self._request_director.handlers.values() if isinstance(rh, ImpersonateRequestHandler) - for target in rh.supported_targets + for target in reversed(rh.supported_targets) ] def _impersonate_target_available(self, target): diff --git a/yt_dlp/__init__.py b/yt_dlp/__init__.py index 7d8f10047..714d9ad5c 100644 --- a/yt_dlp/__init__.py +++ b/yt_dlp/__init__.py @@ -1021,8 +1021,9 @@ def _real_main(argv=None): # List of simplified targets we know are supported, # to help users know what dependencies may be required. (ImpersonateTarget('chrome'), 'curl_cffi'), - (ImpersonateTarget('edge'), 'curl_cffi'), (ImpersonateTarget('safari'), 'curl_cffi'), + (ImpersonateTarget('firefox'), 'curl_cffi>=0.10'), + (ImpersonateTarget('edge'), 'curl_cffi'), ] available_targets = ydl._get_available_impersonate_targets() @@ -1038,12 +1039,12 @@ def _real_main(argv=None): for known_target, known_handler in known_targets: if not any( - known_target in target and handler == known_handler + known_target in target and known_handler.startswith(handler) for target, handler in available_targets ): - rows.append([ + rows.insert(0, [ ydl._format_out(text, ydl.Styles.SUPPRESS) - for text in make_row(known_target, f'{known_handler} (not available)') + for text in make_row(known_target, f'{known_handler} (unavailable)') ]) ydl.to_screen('[info] Available impersonate targets') diff --git a/yt_dlp/networking/_curlcffi.py b/yt_dlp/networking/_curlcffi.py index 0643348e7..c800f2c09 100644 --- a/yt_dlp/networking/_curlcffi.py +++ b/yt_dlp/networking/_curlcffi.py @@ -1,6 +1,7 @@ from __future__ import annotations import io +import itertools import math import re import urllib.parse @@ -31,9 +32,9 @@ if curl_cffi is None: curl_cffi_version = tuple(map(int, re.split(r'[^\d]+', curl_cffi.__version__)[:3])) -if curl_cffi_version != (0, 5, 10) and not ((0, 7, 0) <= curl_cffi_version < (0, 7, 2)): +if curl_cffi_version != (0, 5, 10) and not (0, 10) <= curl_cffi_version: curl_cffi._yt_dlp__version = f'{curl_cffi.__version__} (unsupported)' - raise ImportError('Only curl_cffi versions 0.5.10, 0.7.0 and 0.7.1 are supported') + raise ImportError('Only curl_cffi versions 0.5.10 and 0.10.x are supported') import curl_cffi.requests from curl_cffi.const import CurlECode, CurlOpt @@ -97,7 +98,7 @@ class CurlCFFIResponseAdapter(Response): 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')) + content_length = e.response and 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, @@ -105,6 +106,51 @@ class CurlCFFIResponseAdapter(Response): raise TransportError(cause=e) from e +# See: https://github.com/lexiforest/curl_cffi?tab=readme-ov-file#supported-impersonate-browsers +# https://github.com/lexiforest/curl-impersonate?tab=readme-ov-file#supported-browsers +BROWSER_TARGETS: dict[tuple[int, ...], dict[str, ImpersonateTarget]] = { + (0, 5): { + 'chrome99': ImpersonateTarget('chrome', '99', 'windows', '10'), + 'chrome99_android': ImpersonateTarget('chrome', '99', 'android', '12'), + 'chrome100': ImpersonateTarget('chrome', '100', 'windows', '10'), + 'chrome101': ImpersonateTarget('chrome', '101', 'windows', '10'), + 'chrome104': ImpersonateTarget('chrome', '104', 'windows', '10'), + 'chrome107': ImpersonateTarget('chrome', '107', 'windows', '10'), + 'chrome110': ImpersonateTarget('chrome', '110', 'windows', '10'), + 'edge99': ImpersonateTarget('edge', '99', 'windows', '10'), + 'edge101': ImpersonateTarget('edge', '101', 'windows', '10'), + 'safari15_3': ImpersonateTarget('safari', '15.3', 'macos', '11'), + 'safari15_5': ImpersonateTarget('safari', '15.5', 'macos', '12'), + }, + (0, 7): { + 'chrome116': ImpersonateTarget('chrome', '116', 'windows', '10'), + 'chrome119': ImpersonateTarget('chrome', '119', 'macos', '14'), + 'chrome120': ImpersonateTarget('chrome', '120', 'macos', '14'), + 'chrome123': ImpersonateTarget('chrome', '123', 'macos', '14'), + 'chrome124': ImpersonateTarget('chrome', '124', 'macos', '14'), + 'safari17_0': ImpersonateTarget('safari', '17.0', 'macos', '14'), + 'safari17_2_ios': ImpersonateTarget('safari', '17.2', 'ios', '17.2'), + }, + (0, 9): { + 'safari15_3': ImpersonateTarget('safari', '15.3', 'macos', '14'), + 'safari15_5': ImpersonateTarget('safari', '15.5', 'macos', '14'), + 'chrome119': ImpersonateTarget('chrome', '119', 'macos', '14'), + 'chrome120': ImpersonateTarget('chrome', '120', 'macos', '14'), + 'chrome123': ImpersonateTarget('chrome', '123', 'macos', '14'), + 'chrome124': ImpersonateTarget('chrome', '124', 'macos', '14'), + 'chrome131': ImpersonateTarget('chrome', '131', 'macos', '14'), + 'chrome131_android': ImpersonateTarget('chrome', '131', 'android', '14'), + 'chrome133a': ImpersonateTarget('chrome', '133', 'macos', '15'), + 'firefox133': ImpersonateTarget('firefox', '133', 'macos', '14'), + 'safari18_0': ImpersonateTarget('safari', '18.0', 'macos', '15'), + 'safari18_0_ios': ImpersonateTarget('safari', '18.0', 'ios', '18.0'), + }, + (0, 10): { + 'firefox135': ImpersonateTarget('firefox', '135', 'macos', '14'), + }, +} + + @register_rh class CurlCFFIRH(ImpersonateRequestHandler, InstanceStoreMixin): RH_NAME = 'curl_cffi' @@ -112,30 +158,21 @@ class CurlCFFIRH(ImpersonateRequestHandler, InstanceStoreMixin): _SUPPORTED_FEATURES = (Features.NO_PROXY, Features.ALL_PROXY) _SUPPORTED_PROXY_SCHEMES = ('http', 'https', 'socks4', 'socks4a', 'socks5', 'socks5h') _SUPPORTED_IMPERSONATE_TARGET_MAP = { - **({ - ImpersonateTarget('chrome', '124', 'macos', '14'): curl_cffi.requests.BrowserType.chrome124, - ImpersonateTarget('chrome', '123', 'macos', '14'): curl_cffi.requests.BrowserType.chrome123, - ImpersonateTarget('chrome', '120', 'macos', '14'): curl_cffi.requests.BrowserType.chrome120, - ImpersonateTarget('chrome', '119', 'macos', '14'): curl_cffi.requests.BrowserType.chrome119, - ImpersonateTarget('chrome', '116', 'windows', '10'): curl_cffi.requests.BrowserType.chrome116, - } if curl_cffi_version >= (0, 7, 0) else {}), - 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', '17.0', 'macos', '14'): curl_cffi.requests.BrowserType.safari17_0, - } if curl_cffi_version >= (0, 7, 0) else {}), - 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, - **({ - ImpersonateTarget('safari', '17.2', 'ios', '17.2'): curl_cffi.requests.BrowserType.safari17_2_ios, - } if curl_cffi_version >= (0, 7, 0) else {}), + target: name if curl_cffi_version >= (0, 9) else curl_cffi.requests.BrowserType[name] + for name, target in dict(sorted(itertools.chain.from_iterable( + targets.items() + for version, targets in BROWSER_TARGETS.items() + if curl_cffi_version >= version + ), key=lambda x: ( + # deprioritize mobile targets since they give very different behavior + x[1].os not in ('ios', 'android'), + # prioritize edge < firefox < safari < chrome + ('edge', 'firefox', 'safari', 'chrome').index(x[1].client), + # prioritize newest version + float(x[1].version) if x[1].version else 0, + # group by os name + x[1].os, + ), reverse=True)).items() } def _create_instance(self, cookiejar=None):