devscripts/make_lazy_extractors.py

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