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 module_contents = [
  34     module_template + '\n' + getsource(InfoExtractor.suitable) + '\n',
  35     'class LazyLoadSearchExtractor(LazyLoadExtractor):\n    pass\n']
  36
  37 ie_template = '''
  38 class {name}({bases}):
  39     _VALID_URL = {valid_url!r}
  40     _module = '{module}'
  41 '''
  42
  43 make_valid_template = '''
  44     @classmethod
  45     def _make_valid_url(cls):
  46         return {valid_url!r}
  47 '''
  48
  49
  50 def get_base_name(base):
  51     if base is InfoExtractor:
  52         return 'LazyLoadExtractor'
  53     elif base is SearchInfoExtractor:
  54         return 'LazyLoadSearchExtractor'
  55     else:
  56         return base.__name__
  57
  58
  59 def build_lazy_ie(ie, name):
  60     valid_url = getattr(ie, '_VALID_URL', None)
  61     s = ie_template.format(
  62         name=name,
  63         bases=', '.join(map(get_base_name, ie.__bases__)),
  64         valid_url=valid_url,
  65         module=ie.__module__)
  66     if ie.suitable.__func__ is not InfoExtractor.suitable.__func__:
  67         s += '\n' + getsource(ie.suitable)
  68     if hasattr(ie, '_make_valid_url'):
  69         # search extractors
  70         s += make_valid_template.format(valid_url=ie._make_valid_url())
  71     return s
  72
  73
  74 # find the correct sorting and add the required base classes so that subclasses
  75 # can be correctly created
  76 classes = _ALL_CLASSES[:-1]
  77 ordered_cls = []
  78 while classes:
  79     for c in classes[:]:
  80         bases = set(c.__bases__) - set((object, InfoExtractor, SearchInfoExtractor))
  81         stop = False
  82         for b in bases:
  83             if b not in classes and b not in ordered_cls:
  84                 if b.__name__ == 'GenericIE':
  85                     exit()
  86                 classes.insert(0, b)
  87                 stop = True
  88         if stop:
  89             break
  90         if all(b in ordered_cls for b in bases):
  91             ordered_cls.append(c)
  92             classes.remove(c)
  93             break
  94 ordered_cls.append(_ALL_CLASSES[-1])
  95
  96 names = []
  97 for ie in ordered_cls:
  98     name = ie.__name__
  99     src = build_lazy_ie(ie, name)
 100     module_contents.append(src)
 101     if ie in _ALL_CLASSES:
 102         names.append(name)
 103
 104 module_contents.append(
 105     '_ALL_CLASSES = [{0}]'.format(', '.join(names)))
 106
 107 module_src = '\n'.join(module_contents) + '\n'
 108
 109 with io.open(lazy_extractors_filename, 'wt', encoding='utf-8') as f:
 110     f.write(module_src)