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