How to use the sister.download.get_cache_directory function in sister

def test_fails_to_make_directory(self, f: Callable):
        f.side_effect = OSError()
        with self.assertRaises(OSError):
            download.get_cache_directory('/sister_test_cache', True)

def test_get_cache_directory(self):
        root = download.get_cache_root()
        path = download.get_cache_directory('test', False)
        self.assertEqual(path, os.path.join(root, 'test'))

}
    path = download.cached_download(urls[lang])
    path = Path(path)

    filename = "word2vec.gensim.model"

    print("Loading model...")

    if lang == "ja":
        dirpath = Path(download.get_cache_directory(str(Path("word2vec"))))
        download.cached_unzip(path, dirpath / lang)
        model_path = dirpath / lang / filename
        model = gensim.models.Word2Vec.load(str(model_path))

    if lang == "en":
        dirpath = Path(download.get_cache_directory(str(Path("word2vec") / "en")))
        model_path = dirpath / filename
        download.cached_decompress_gzip(path, model_path)
        model = gensim.models.KeyedVectors.load_word2vec_format(str(model_path), binary=True)

    return model

def get_word2vec(lang: str = "en"):
    # Download.
    urls = {
            "en": "https://s3.amazonaws.com/dl4j-distribution/GoogleNews-vectors-negative300.bin.gz",
            "ja": "http://public.shiroyagi.s3.amazonaws.com/latest-ja-word2vec-gensim-model.zip"
            }
    path = download.cached_download(urls[lang])
    path = Path(path)

    filename = "word2vec.gensim.model"

    print("Loading model...")

    if lang == "ja":
        dirpath = Path(download.get_cache_directory(str(Path("word2vec"))))
        download.cached_unzip(path, dirpath / lang)
        model_path = dirpath / lang / filename
        model = gensim.models.Word2Vec.load(str(model_path))

    if lang == "en":
        dirpath = Path(download.get_cache_directory(str(Path("word2vec") / "en")))
        model_path = dirpath / filename
        download.cached_decompress_gzip(path, model_path)
        model = gensim.models.KeyedVectors.load_word2vec_format(str(model_path), binary=True)

    return model