def _create_embeddings_from_file(embed_file, embed_dsz, embed_sha1,
data_download_cache, vocab, unif,
keep_unused):
embed_file = EmbeddingDownloader(embed_file, embed_dsz, embed_sha1,
data_download_cache).download()
EmbeddingT = baseline.GloVeModel if mime_type(
embed_file) == 'text/plain' else baseline.Word2VecModel
return EmbeddingT(embed_file,
vocab,
unif_weight=unif,
keep_unused=keep_unused)
def unzip_model(path):
"""If the path for a model file is a zip file, unzip it in /tmp and return the unzipped path"""
if not os.path.exists(path) or not mime_type(path) == "application/zip":
return path
with open(path, 'rb') as f:
sha1 = hashlib.sha1(f.read()).hexdigest()
temp_dir = os.path.join("/tmp/", sha1)
if not os.path.exists(temp_dir):
print("unzipping model before exporting")
with zipfile.ZipFile(path, "r") as zip_ref:
zip_ref.extractall(temp_dir)
temp_dir = os.path.join(temp_dir, os.listdir(temp_dir)[0])
path = os.path.join(temp_dir,
[x[:-6] for x in os.listdir(temp_dir)
if 'index' in x][0])
return path