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 plugins_dirname = 'ytdlp_plugins'
  20 plugins_blocked_dirname = 'ytdlp_plugins_blocked'
  21 if os.path.exists(plugins_dirname):
  22     os.rename(plugins_dirname, plugins_blocked_dirname)
  23
  24 from yt_dlp.extractor import _ALL_CLASSES
  25 from yt_dlp.extractor.common import InfoExtractor, SearchInfoExtractor
  26
  27 if os.path.exists(plugins_blocked_dirname):
  28     os.rename(plugins_blocked_dirname, plugins_dirname)
  29
  30 with open('devscripts/lazy_load_template.py', 'rt') as f:
  31     module_template = f.read()
  32
  33 CLASS_PROPERTIES = ['ie_key', 'working', '_match_valid_url', 'suitable', '_match_id', 'get_temp_id']
  34 module_contents = [
  35     module_template,
  36     *[getsource(getattr(InfoExtractor, k)) for k in CLASS_PROPERTIES],
  37     '\nclass LazyLoadSearchExtractor(LazyLoadExtractor):\n    pass\n']
  38
  39 ie_template = '''
  40 class {name}({bases}):
  41     _module = '{module}'
  42 '''
  43
  44 make_valid_template = '''
  45     @classmethod
  46     def _make_valid_url(cls):
  47         return {valid_url!r}
  48 '''
  49
  50
  51 def get_base_name(base):
  52     if base is InfoExtractor:
  53         return 'LazyLoadExtractor'
  54     elif base is SearchInfoExtractor:
  55         return 'LazyLoadSearchExtractor'
  56     else:
  57         return base.__name__
  58
  59
  60 def build_lazy_ie(ie, name):
  61     s = ie_template.format(
  62         name=name,
  63         bases=', '.join(map(get_base_name, ie.__bases__)),
  64         module=ie.__module__)
  65     valid_url = getattr(ie, '_VALID_URL', None)
  66     if valid_url:
  67         s += f'    _VALID_URL = {valid_url!r}\n'
  68     if not ie._WORKING:
  69         s += '    _WORKING = False\n'
  70     if ie.suitable.__func__ is not InfoExtractor.suitable.__func__:
  71         s += f'\n{getsource(ie.suitable)}'
  72     if hasattr(ie, '_make_valid_url'):
  73         # search extractors
  74         s += make_valid_template.format(valid_url=ie._make_valid_url())
  75     return s
  76
  77
  78 # find the correct sorting and add the required base classes so that subclasses
  79 # can be correctly created
  80 classes = _ALL_CLASSES[:-1]
  81 ordered_cls = []
  82 while classes:
  83     for c in classes[:]:
  84         bases = set(c.__bases__) - set((object, InfoExtractor, SearchInfoExtractor))
  85         stop = False
  86         for b in bases:
  87             if b not in classes and b not in ordered_cls:
  88                 if b.__name__ == 'GenericIE':
  89                     exit()
  90                 classes.insert(0, b)
  91                 stop = True
  92         if stop:
  93             break
  94         if all(b in ordered_cls for b in bases):
  95             ordered_cls.append(c)
  96             classes.remove(c)
  97             break
  98 ordered_cls.append(_ALL_CLASSES[-1])
  99
 100 names = []
 101 for ie in ordered_cls:
 102     name = ie.__name__
 103     src = build_lazy_ie(ie, name)
 104     module_contents.append(src)
 105     if ie in _ALL_CLASSES:
 106         names.append(name)
 107
 108 module_contents.append(
 109     '\n_ALL_CLASSES = [{0}]'.format(', '.join(names)))
 110
 111 module_src = '\n'.join(module_contents) + '\n'
 112
 113 with io.open(lazy_extractors_filename, 'wt', encoding='utf-8') as f:
 114     f.write(module_src)