rain1024 commited on
Commit
de04f3f
·
1 Parent(s): 3d05b38
Files changed (2) hide show
  1. data/UTS_Text_v1.txt +0 -0
  2. generate_dataset.py +48 -0
data/UTS_Text_v1.txt ADDED
The diff for this file is too large to render. See raw diff
 
generate_dataset.py ADDED
@@ -0,0 +1,48 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from os.path import join, dirname
2
+ from underthesea.file_utils import DATASETS_FOLDER
3
+ import random
4
+
5
+ random.seed(10)
6
+ # sampling data
7
+ text_file = join(DATASETS_FOLDER, "VNESES", "VNESEScorpus.txt")
8
+ with open(text_file) as f:
9
+ lines = f.read().splitlines()
10
+ NUM_LONG_TOKENS = 50
11
+ NUM_SHORT_TOKENS = 20
12
+ long_lines = [
13
+ line
14
+ for line in lines
15
+ if len(line) >= NUM_LONG_TOKENS and line[0].isupper() and line[-1] == "."
16
+ ]
17
+ # get random 1000 lines
18
+ random_long_lines = random.sample(long_lines, 5000)
19
+ for line in random_long_lines[:20]:
20
+ print(line)
21
+
22
+
23
+ def shortline_conditions(line):
24
+ if len(line) < NUM_SHORT_TOKENS:
25
+ return False
26
+ if len(line) > NUM_LONG_TOKENS:
27
+ return False
28
+ if not line[0].isupper():
29
+ return False
30
+ return True
31
+
32
+
33
+ short_lines = [line for line in lines if shortline_conditions(line)]
34
+ random_short_lines = random.sample(short_lines, 5000)
35
+ for line in random_short_lines[:20]:
36
+ print(line)
37
+
38
+ print("Long lines", len(random_long_lines))
39
+ print("Short lines", len(random_short_lines))
40
+
41
+ pwd = dirname(__file__)
42
+ tmp = join(pwd, "data")
43
+ corpus_file = join(tmp, "UTS_Text_v1.txt")
44
+
45
+ with open(corpus_file, "w") as f:
46
+ lines = random_long_lines + random_short_lines
47
+ content = "\n".join(lines)
48
+ f.write(content)