# -*- coding: utf-8 -*-
import os
import pandas as pd
from progressbar import Bar, ETA, FileTransferSpeed, ProgressBar, Percentage, RotatingMarker
from six.moves.urllib.request import urlretrieve
def load_datasets(path=os.path.join(os.path.dirname(__file__), 'datasets.csv')):
datasets = pd.read_csv(path)
return datasets
def download(number=-1, name="", save_dir='./'):
"""Download pre-trained word vector
:param number: integer, default ``None``
:param save_dir: str, default './'
:return: file path for downloaded file
"""
df = load_datasets()
if number > -1:
row = df.iloc[[number]]
elif name:
row = df.loc[df["Name"] == name]
url = ''.join(row.URL)
if not url:
print('The word vector you specified was not found. Please specify correct name.')
widgets = ['Test: ', Percentage(), ' ', Bar(marker=RotatingMarker()), ' ', ETA(), ' ', FileTransferSpeed()]
pbar = ProgressBar(widgets=widgets)
def dlProgress(count, blockSize, totalSize):
if pbar.max_value is None:
pbar.max_value = totalSize
pbar.start()
pbar.update(min(count * blockSize, totalSize))
file_name = url.split('/')[-1]
if not os.path.exists(save_dir):
os.makedirs(save_dir)
save_path = os.path.join(save_dir, file_name)
path, _ = urlretrieve(url, save_path, reporthook=dlProgress)
pbar.finish()
return path
def search(lang=''):
"""Search pre-trained word vectors by their language
:param lang: str, default ''
:return: None
print search result as pandas DataFrame
"""
df = load_datasets()
if lang == '':
print(df[['Name', 'Dimension', 'Corpus', 'VocabularySize', 'Method', 'Language', 'Author']])
else:
rows = df[df.Language==lang]
print(rows[['Name', 'Dimension', 'Corpus', 'VocabularySize', 'Method', 'Language', 'Author']])
