yt_dlp/cache.py

   1 import contextlib
   2 import errno
   3 import json
   4 import os
   5 import re
   6 import shutil
   7 import traceback
   8 import urllib.parse
   9
  10 from .utils import expand_path, traverse_obj, version_tuple, write_json_file
  11 from .version import __version__
  12
  13
  14 class Cache:
  15     def __init__(self, ydl):
  16         self._ydl = ydl
  17
  18     def _get_root_dir(self):
  19         res = self._ydl.params.get('cachedir')
  20         if res is None:
  21             cache_root = os.getenv('XDG_CACHE_HOME', '~/.cache')
  22             res = os.path.join(cache_root, 'yt-dlp')
  23         return expand_path(res)
  24
  25     def _get_cache_fn(self, section, key, dtype):
  26         assert re.match(r'^[\w.-]+$', section), f'invalid section {section!r}'
  27         key = urllib.parse.quote(key, safe='').replace('%', ',')  # encode non-ascii characters
  28         return os.path.join(self._get_root_dir(), section, f'{key}.{dtype}')
  29
  30     @property
  31     def enabled(self):
  32         return self._ydl.params.get('cachedir') is not False
  33
  34     def store(self, section, key, data, dtype='json'):
  35         assert dtype in ('json',)
  36
  37         if not self.enabled:
  38             return
  39
  40         fn = self._get_cache_fn(section, key, dtype)
  41         try:
  42             os.makedirs(os.path.dirname(fn), exist_ok=True)
  43             self._ydl.write_debug(f'Saving {section}.{key} to cache')
  44             write_json_file({'yt-dlp_version': __version__, 'data': data}, fn)
  45         except Exception:
  46             tb = traceback.format_exc()
  47             self._ydl.report_warning(f'Writing cache to {fn!r} failed: {tb}')
  48
  49     def _validate(self, data, min_ver):
  50         version = traverse_obj(data, 'yt-dlp_version')
  51         if not version:  # Backward compatibility
  52             data, version = {'data': data}, '2022.08.19'
  53         if not min_ver or version_tuple(version) >= version_tuple(min_ver):
  54             return data['data']
  55         self._ydl.write_debug(f'Discarding old cache from version {version} (needs {min_ver})')
  56
  57     def load(self, section, key, dtype='json', default=None, *, min_ver=None):
  58         assert dtype in ('json',)
  59
  60         if not self.enabled:
  61             return default
  62
  63         cache_fn = self._get_cache_fn(section, key, dtype)
  64         with contextlib.suppress(OSError):
  65             try:
  66                 with open(cache_fn, encoding='utf-8') as cachef:
  67                     self._ydl.write_debug(f'Loading {section}.{key} from cache')
  68                     return self._validate(json.load(cachef), min_ver)
  69             except (ValueError, KeyError):
  70                 try:
  71                     file_size = os.path.getsize(cache_fn)
  72                 except OSError as oe:
  73                     file_size = str(oe)
  74                 self._ydl.report_warning(f'Cache retrieval from {cache_fn} failed ({file_size})')
  75
  76         return default
  77
  78     def remove(self):
  79         if not self.enabled:
  80             self._ydl.to_screen('Cache is disabled (Did you combine --no-cache-dir and --rm-cache-dir?)')
  81             return
  82
  83         cachedir = self._get_root_dir()
  84         if not any((term in cachedir) for term in ('cache', 'tmp')):
  85             raise Exception('Not removing directory %s - this does not look like a cache dir' % cachedir)
  86
  87         self._ydl.to_screen(
  88             'Removing cache dir %s .' % cachedir, skip_eol=True)
  89         if os.path.exists(cachedir):
  90             self._ydl.to_screen('.', skip_eol=True)
  91             shutil.rmtree(cachedir)
  92         self._ydl.to_screen('.')