mirror of
https://github.com/explosion/spaCy.git
synced 2024-09-21 19:39:13 +03:00
56 lines
1.5 KiB
Python
56 lines
1.5 KiB
Python
from os import path
|
|
import os
|
|
import tarfile
|
|
import shutil
|
|
import wget
|
|
import plac
|
|
|
|
|
|
ALL_DATA_DIR_URL = 'http://s3-us-west-1.amazonaws.com/media.spacynlp.com/en_data_all-0.4.tgz'
|
|
|
|
SM_DATA_DIR_URL = 'http://s3-us-west-1.amazonaws.com/media.spacynlp.com/en_data_sm-0.4.tgz'
|
|
|
|
PARSER_URL = 'http://s3-us-west-1.amazonaws.com/media.spacynlp.com/en_deps-0.30.tgz'
|
|
|
|
DEP_VECTORS_URL = 'http://s3-us-west-1.amazonaws.com/media.spacynlp.com/vec.bin'
|
|
|
|
DEST_DIR = path.join(path.dirname(__file__), 'data')
|
|
|
|
def download_file(url, out):
|
|
wget.download(url, out=out)
|
|
return url.rsplit('/', 1)[1]
|
|
|
|
|
|
def install_data(url, dest_dir):
|
|
filename = download_file(url, dest_dir)
|
|
t = tarfile.open(path.join(dest_dir, filename))
|
|
t.extractall(dest_dir)
|
|
|
|
def install_parser_model(url, dest_dir):
|
|
filename = download_file(url, dest_dir)
|
|
t = tarfile.open(path.join(dest_dir, filename), mode=":gz")
|
|
t.extractall(path.dirname(__file__))
|
|
|
|
|
|
def install_dep_vectors(url, dest_dir):
|
|
if not os.path.exists(dest_dir):
|
|
os.mkdir(dest_dir)
|
|
|
|
filename = download_file(url, dest_dir)
|
|
|
|
|
|
def main(data_size='all'):
|
|
if data_size == 'all':
|
|
data_url = ALL_DATA_DIR_URL
|
|
elif data_size == 'small':
|
|
data_url = SM_DATA_DIR_URL
|
|
if not path.exists(DEST_DIR):
|
|
install_data(data_url, path.dirname(DEST_DIR))
|
|
else:
|
|
install_parser_model(PARSER_URL, DEST_DIR)
|
|
install_dep_vectors(DEP_VECTORS_URL, path.join(DEST_DIR, 'vocab'))
|
|
|
|
|
|
if __name__ == '__main__':
|
|
plac.call(main)
|