devscripts/make_lazy_extractors.py

   1 #!/usr/bin/env python3
   2 from __future__ import unicode_literals, print_function
   3
   4 from inspect import getsource
   5 import io
   6 import os
   7 from os.path import dirname as dirn
   8 import sys
   9
  10 print('WARNING: Lazy loading extractors is an experimental feature that may not always work', file=sys.stderr)
  11
  12 sys.path.insert(0, dirn(dirn((os.path.abspath(__file__)))))
  13
  14 lazy_extractors_filename = sys.argv[1]
  15 if os.path.exists(lazy_extractors_filename):
  16     os.remove(lazy_extractors_filename)
  17
  18 # Block plugins from loading
  19 os.rename('ytdlp_plugins', 'ytdlp_plugins_blocked')
  20
  21 from yt_dlp.extractor import _ALL_CLASSES
  22 from yt_dlp.extractor.common import InfoExtractor, SearchInfoExtractor
  23
  24 os.rename('ytdlp_plugins_blocked', 'ytdlp_plugins')
  25
  26 with open('devscripts/lazy_load_template.py', 'rt') as f:
  27     module_template = f.read()
  28
  29 module_contents = [
  30     module_template + '\n' + getsource(InfoExtractor.suitable) + '\n',
  31     'class LazyLoadSearchExtractor(LazyLoadExtractor):\n    pass\n']
  32
  33 ie_template = '''
  34 class {name}({bases}):
  35     _VALID_URL = {valid_url!r}
  36     _module = '{module}'
  37 '''
  38
  39 make_valid_template = '''
  40     @classmethod
  41     def _make_valid_url(cls):
  42         return {valid_url!r}
  43 '''
  44
  45
  46 def get_base_name(base):
  47     if base is InfoExtractor:
  48         return 'LazyLoadExtractor'
  49     elif base is SearchInfoExtractor:
  50         return 'LazyLoadSearchExtractor'
  51     else:
  52         return base.__name__
  53
  54
  55 def build_lazy_ie(ie, name):
  56     valid_url = getattr(ie, '_VALID_URL', None)
  57     s = ie_template.format(
  58         name=name,
  59         bases=', '.join(map(get_base_name, ie.__bases__)),
  60         valid_url=valid_url,
  61         module=ie.__module__)
  62     if ie.suitable.__func__ is not InfoExtractor.suitable.__func__:
  63         s += '\n' + getsource(ie.suitable)
  64     if hasattr(ie, '_make_valid_url'):
  65         # search extractors
  66         s += make_valid_template.format(valid_url=ie._make_valid_url())
  67     return s
  68
  69
  70 # find the correct sorting and add the required base classes so that subclasses
  71 # can be correctly created
  72 classes = _ALL_CLASSES[:-1]
  73 ordered_cls = []
  74 while classes:
  75     for c in classes[:]:
  76         bases = set(c.__bases__) - set((object, InfoExtractor, SearchInfoExtractor))
  77         stop = False
  78         for b in bases:
  79             if b not in classes and b not in ordered_cls:
  80                 if b.__name__ == 'GenericIE':
  81                     exit()
  82                 classes.insert(0, b)
  83                 stop = True
  84         if stop:
  85             break
  86         if all(b in ordered_cls for b in bases):
  87             ordered_cls.append(c)
  88             classes.remove(c)
  89             break
  90 ordered_cls.append(_ALL_CLASSES[-1])
  91
  92 names = []
  93 for ie in ordered_cls:
  94     name = ie.__name__
  95     src = build_lazy_ie(ie, name)
  96     module_contents.append(src)
  97     if ie in _ALL_CLASSES:
  98         names.append(name)
  99
 100 module_contents.append(
 101     '_ALL_CLASSES = [{0}]'.format(', '.join(names)))
 102
 103 module_src = '\n'.join(module_contents) + '\n'
 104
 105 with io.open(lazy_extractors_filename, 'wt', encoding='utf-8') as f:
 106     f.write(module_src)