Spaces:
Build error
Build error
update utils
Browse files- src/utils/file.py +32 -1
src/utils/file.py
CHANGED
@@ -1,8 +1,39 @@
|
|
1 |
|
2 |
import os
|
3 |
-
|
4 |
|
|
|
|
|
|
|
5 |
|
6 |
def ensure_folder_exists(path):
|
7 |
if not os.path.exists(path):
|
8 |
os.makedirs(path)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
|
2 |
import os
|
3 |
+
import glob
|
4 |
|
5 |
+
from multiprocessing import Pool
|
6 |
+
import timeit
|
7 |
+
from tqdm import tqdm
|
8 |
|
9 |
def ensure_folder_exists(path):
|
10 |
if not os.path.exists(path):
|
11 |
os.makedirs(path)
|
12 |
+
|
13 |
+
def find_files(pattern):
|
14 |
+
files = glob.glob(pattern)
|
15 |
+
return files
|
16 |
+
|
17 |
+
|
18 |
+
def save_to_file(file_name, content):
|
19 |
+
with open(file_name, 'w', encoding='utf-8') as f:
|
20 |
+
f.write(content)
|
21 |
+
# print("Saved: ", file_name)
|
22 |
+
|
23 |
+
|
24 |
+
def multi_process_list(input_list, function, num_workers=20):
|
25 |
+
with Pool(num_workers) as p:
|
26 |
+
result = list(tqdm(p.imap_unordered(function, input_list), total=len(input_list)))
|
27 |
+
return result
|
28 |
+
|
29 |
+
def process_list(input_list, function):
|
30 |
+
# process with process bar
|
31 |
+
result = []
|
32 |
+
for item in tqdm(input_list):
|
33 |
+
result.append(function(item))
|
34 |
+
return result
|
35 |
+
|
36 |
+
def read_txt_content(file):
|
37 |
+
with open(file, 'r', encoding='utf-8') as f:
|
38 |
+
text = f.read()
|
39 |
+
return text
|