aboutsummaryrefslogtreecommitdiffstats
path: root/devscripts
diff options
context:
space:
mode:
Diffstat (limited to 'devscripts')
-rw-r--r--devscripts/lazy_load_template.py24
-rw-r--r--devscripts/make_lazy_extractors.py196
-rw-r--r--devscripts/make_supportedsites.py4
3 files changed, 123 insertions, 101 deletions
diff --git a/devscripts/lazy_load_template.py b/devscripts/lazy_load_template.py
index e4b4f5825..6d9b27742 100644
--- a/devscripts/lazy_load_template.py
+++ b/devscripts/lazy_load_template.py
@@ -1,30 +1,28 @@
+import importlib
+import random
import re
-from ..utils import bug_reports_message, write_string
+from ..utils import bug_reports_message, classproperty, write_string
class LazyLoadMetaClass(type):
def __getattr__(cls, name):
- if '_real_class' not in cls.__dict__:
+ # "is_suitable" requires "_TESTS". However, they bloat the lazy_extractors
+ if '_real_class' not in cls.__dict__ and name not in ('is_suitable', 'get_testcases'):
write_string(
'WARNING: Falling back to normal extractor since lazy extractor '
- f'{cls.__name__} does not have attribute {name}{bug_reports_message()}')
- return getattr(cls._get_real_class(), name)
+ f'{cls.__name__} does not have attribute {name}{bug_reports_message()}\n')
+ return getattr(cls.real_class, name)
class LazyLoadExtractor(metaclass=LazyLoadMetaClass):
- _module = None
- _WORKING = True
-
- @classmethod
- def _get_real_class(cls):
+ @classproperty
+ def real_class(cls):
if '_real_class' not in cls.__dict__:
- mod = __import__(cls._module, fromlist=(cls.__name__,))
- cls._real_class = getattr(mod, cls.__name__)
+ cls._real_class = getattr(importlib.import_module(cls._module), cls.__name__)
return cls._real_class
def __new__(cls, *args, **kwargs):
- real_cls = cls._get_real_class()
- instance = real_cls.__new__(real_cls)
+ instance = cls.real_class.__new__(cls.real_class)
instance.__init__(*args, **kwargs)
return instance
diff --git a/devscripts/make_lazy_extractors.py b/devscripts/make_lazy_extractors.py
index 6dc8fed90..8ddc54b9b 100644
--- a/devscripts/make_lazy_extractors.py
+++ b/devscripts/make_lazy_extractors.py
@@ -1,101 +1,125 @@
#!/usr/bin/env python3
import os
+import optparse
import sys
from inspect import getsource
sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
-lazy_extractors_filename = sys.argv[1] if len(sys.argv) > 1 else 'yt_dlp/extractor/lazy_extractors.py'
-if os.path.exists(lazy_extractors_filename):
- os.remove(lazy_extractors_filename)
-# Block plugins from loading
-plugins_dirname = 'ytdlp_plugins'
-plugins_blocked_dirname = 'ytdlp_plugins_blocked'
-if os.path.exists(plugins_dirname):
- os.rename(plugins_dirname, plugins_blocked_dirname)
-
-from yt_dlp.extractor import _ALL_CLASSES
-from yt_dlp.extractor.common import InfoExtractor, SearchInfoExtractor
-
-if os.path.exists(plugins_blocked_dirname):
- os.rename(plugins_blocked_dirname, plugins_dirname)
-
-with open('devscripts/lazy_load_template.py', encoding='utf-8') as f:
- module_template = f.read()
-
-CLASS_PROPERTIES = ['ie_key', 'working', '_match_valid_url', 'suitable', '_match_id', 'get_temp_id']
-module_contents = [
- module_template,
- *[getsource(getattr(InfoExtractor, k)) for k in CLASS_PROPERTIES],
- '\nclass LazyLoadSearchExtractor(LazyLoadExtractor):\n pass\n']
-
-ie_template = '''
+NO_ATTR = object()
+STATIC_CLASS_PROPERTIES = ['IE_NAME', 'IE_DESC', 'SEARCH_KEY', '_WORKING', '_NETRC_MACHINE']
+CLASS_METHODS = [
+ 'ie_key', 'working', 'description', 'suitable', '_match_valid_url', '_match_id', 'get_temp_id',
+]
+IE_TEMPLATE = '''
class {name}({bases}):
- _module = '{module}'
+ _module = {module!r}
'''
-
-
-def get_base_name(base):
- if base is InfoExtractor:
- return 'LazyLoadExtractor'
- elif base is SearchInfoExtractor:
- return 'LazyLoadSearchExtractor'
- else:
- return base.__name__
-
-
-def build_lazy_ie(ie, name):
- s = ie_template.format(
- name=name,
- bases=', '.join(map(get_base_name, ie.__bases__)),
- module=ie.__module__)
+with open('devscripts/lazy_load_template.py', encoding='utf-8') as f:
+ MODULE_TEMPLATE = f.read()
+
+
+def main():
+ parser = optparse.OptionParser(usage='%prog [OUTFILE.py]')
+ args = parser.parse_args()[1] or ['yt_dlp/extractor/lazy_extractors.py']
+ if len(args) != 1:
+ parser.error('Expected only an output filename')
+
+ lazy_extractors_filename = args[0]
+ if os.path.exists(lazy_extractors_filename):
+ os.remove(lazy_extractors_filename)
+
+ _ALL_CLASSES = get_all_ies() # Must be before import
+
+ from yt_dlp.extractor.common import InfoExtractor, SearchInfoExtractor
+
+ DummyInfoExtractor = type('InfoExtractor', (InfoExtractor,), {'IE_NAME': NO_ATTR})
+ module_src = '\n'.join((
+ MODULE_TEMPLATE,
+ ' _module = None',
+ *extra_ie_code(DummyInfoExtractor),
+ '\nclass LazyLoadSearchExtractor(LazyLoadExtractor):\n pass\n',
+ *build_ies(_ALL_CLASSES, (InfoExtractor, SearchInfoExtractor), DummyInfoExtractor),
+ ))
+
+ with open(lazy_extractors_filename, 'wt', encoding='utf-8') as f:
+ f.write(f'{module_src}\n')
+
+
+def get_all_ies():
+ PLUGINS_DIRNAME = 'ytdlp_plugins'
+ BLOCKED_DIRNAME = f'{PLUGINS_DIRNAME}_blocked'
+ if os.path.exists(PLUGINS_DIRNAME):
+ os.rename(PLUGINS_DIRNAME, BLOCKED_DIRNAME)
+ try:
+ from yt_dlp.extractor import _ALL_CLASSES
+ finally:
+ if os.path.exists(BLOCKED_DIRNAME):
+ os.rename(BLOCKED_DIRNAME, PLUGINS_DIRNAME)
+ return _ALL_CLASSES
+
+
+def extra_ie_code(ie, base=None):
+ for var in STATIC_CLASS_PROPERTIES:
+ val = getattr(ie, var)
+ if val != (getattr(base, var) if base else NO_ATTR):
+ yield f' {var} = {val!r}'
+ yield ''
+
+ for name in CLASS_METHODS:
+ f = getattr(ie, name)
+ if not base or f.__func__ != getattr(base, name).__func__:
+ yield getsource(f)
+
+
+def build_ies(ies, bases, attr_base):
+ names = []
+ for ie in sort_ies(ies, bases):
+ yield build_lazy_ie(ie, ie.__name__, attr_base)
+ if ie in ies:
+ names.append(ie.__name__)
+
+ yield f'\n_ALL_CLASSES = [{", ".join(names)}]'
+
+
+def sort_ies(ies, ignored_bases):
+ """find the correct sorting and add the required base classes so that subclasses can be correctly created"""
+ classes, returned_classes = ies[:-1], set()
+ assert ies[-1].__name__ == 'GenericIE', 'Last IE must be GenericIE'
+ while classes:
+ for c in classes[:]:
+ bases = set(c.__bases__) - {object, *ignored_bases}
+ restart = False
+ for b in bases:
+ if b not in classes and b not in returned_classes:
+ assert b.__name__ != 'GenericIE', 'Cannot inherit from GenericIE'
+ classes.insert(0, b)
+ restart = True
+ if restart:
+ break
+ if bases <= returned_classes:
+ yield c
+ returned_classes.add(c)
+ classes.remove(c)
+ break
+ yield ies[-1]
+
+
+def build_lazy_ie(ie, name, attr_base):
+ bases = ', '.join({
+ 'InfoExtractor': 'LazyLoadExtractor',
+ 'SearchInfoExtractor': 'LazyLoadSearchExtractor',
+ }.get(base.__name__, base.__name__) for base in ie.__bases__)
+
+ s = IE_TEMPLATE.format(name=name, module=ie.__module__, bases=bases)
valid_url = getattr(ie, '_VALID_URL', None)
if not valid_url and hasattr(ie, '_make_valid_url'):
valid_url = ie._make_valid_url()
if valid_url:
s += f' _VALID_URL = {valid_url!r}\n'
- if not ie._WORKING:
- s += ' _WORKING = False\n'
- if ie.suitable.__func__ is not InfoExtractor.suitable.__func__:
- s += f'\n{getsource(ie.suitable)}'
- return s
-
-
-# find the correct sorting and add the required base classes so that subclasses
-# can be correctly created
-classes = _ALL_CLASSES[:-1]
-ordered_cls = []
-while classes:
- for c in classes[:]:
- bases = set(c.__bases__) - {object, InfoExtractor, SearchInfoExtractor}
- stop = False
- for b in bases:
- if b not in classes and b not in ordered_cls:
- if b.__name__ == 'GenericIE':
- exit()
- classes.insert(0, b)
- stop = True
- if stop:
- break
- if all(b in ordered_cls for b in bases):
- ordered_cls.append(c)
- classes.remove(c)
- break
-ordered_cls.append(_ALL_CLASSES[-1])
-
-names = []
-for ie in ordered_cls:
- name = ie.__name__
- src = build_lazy_ie(ie, name)
- module_contents.append(src)
- if ie in _ALL_CLASSES:
- names.append(name)
-
-module_contents.append(
- '\n_ALL_CLASSES = [{}]'.format(', '.join(names)))
-
-module_src = '\n'.join(module_contents) + '\n'
-
-with open(lazy_extractors_filename, 'wt', encoding='utf-8') as f:
- f.write(module_src)
+ return s + '\n'.join(extra_ie_code(ie, attr_base))
+
+
+if __name__ == '__main__':
+ main()
diff --git a/devscripts/make_supportedsites.py b/devscripts/make_supportedsites.py
index 5531fec4d..d8c53c5e1 100644
--- a/devscripts/make_supportedsites.py
+++ b/devscripts/make_supportedsites.py
@@ -5,7 +5,7 @@ import sys
sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
-from yt_dlp.extractor import list_extractors
+from yt_dlp.extractor import list_extractor_classes
def main():
@@ -14,7 +14,7 @@ def main():
if len(args) != 1:
parser.error('Expected an output filename')
- out = '\n'.join(ie.description() for ie in list_extractors(None) if ie.IE_DESC is not False)
+ out = '\n'.join(ie.description() for ie in list_extractor_classes() if ie.IE_DESC is not False)
with open(args[0], 'w', encoding='utf-8') as outf:
outf.write(f'# Supported sites\n{out}\n')
generated by cgit v1.2.3 (git 2.25.1) at 2025-06-08 10:06:43 +0000