Spaces:
Runtime error
Runtime error
Commit
•
5c0b561
1
Parent(s):
da8a25b
Upload utils.py
Browse files
utils.py
ADDED
|
@@ -0,0 +1,76 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import jsonlines
|
| 2 |
+
import json
|
| 3 |
+
|
| 4 |
+
def read_jsonlines(path):
|
| 5 |
+
with jsonlines.open(path) as reader:
|
| 6 |
+
lines = [obj for obj in reader]
|
| 7 |
+
return lines
|
| 8 |
+
|
| 9 |
+
def write_jsonlines(file, path):
|
| 10 |
+
with jsonlines.open(path, 'w') as writer:
|
| 11 |
+
writer.write_all(file)
|
| 12 |
+
|
| 13 |
+
def dump_json(file, path):
|
| 14 |
+
"""Save json object"""
|
| 15 |
+
|
| 16 |
+
with open(path, 'w', encoding='utf-8') as f:
|
| 17 |
+
json.dump(file, f, indent=4, ensure_ascii=False)
|
| 18 |
+
print("Saved json to path: " + str(path))
|
| 19 |
+
|
| 20 |
+
def load_json(path):
|
| 21 |
+
"""load json object"""
|
| 22 |
+
with open(path, 'rb') as f:
|
| 23 |
+
data = json.load(f)
|
| 24 |
+
print("Loaded json from path: " + str(path))
|
| 25 |
+
return data
|
| 26 |
+
|
| 27 |
+
def split_w_delimiter(line, delimiter):
|
| 28 |
+
len_delimiter = len(delimiter)
|
| 29 |
+
|
| 30 |
+
sents = [sent for sent in line.split(delimiter) if sent.strip()]
|
| 31 |
+
|
| 32 |
+
if len(sents) == 0:
|
| 33 |
+
return []
|
| 34 |
+
|
| 35 |
+
if line[-len_delimiter:] == delimiter:
|
| 36 |
+
out = [sent + delimiter for sent in sents]
|
| 37 |
+
else:
|
| 38 |
+
out = [sent + delimiter for sent in sents[:-1]] + [sents[-1]]
|
| 39 |
+
|
| 40 |
+
return out
|
| 41 |
+
|
| 42 |
+
import difflib
|
| 43 |
+
|
| 44 |
+
def generate_diff_html(text_pairs, output_file=''):
|
| 45 |
+
diff_css = '''
|
| 46 |
+
<style>
|
| 47 |
+
ins {
|
| 48 |
+
color: white;
|
| 49 |
+
background-color: #d4fcbc;
|
| 50 |
+
text-decoration: none;
|
| 51 |
+
}
|
| 52 |
+
del {
|
| 53 |
+
background-color: #fbb6c2;
|
| 54 |
+
text-decoration: none;
|
| 55 |
+
}
|
| 56 |
+
</style>
|
| 57 |
+
'''
|
| 58 |
+
|
| 59 |
+
diff_html = '<!DOCTYPE html>\n<html lang="en">\n<head>\n<meta charset="UTF-8">\n<title>Text Diff</title>\n'
|
| 60 |
+
diff_html += diff_css
|
| 61 |
+
diff_html += '\n</head>\n<body>\n'
|
| 62 |
+
|
| 63 |
+
for i, (text, text_processed) in enumerate(text_pairs, 1):
|
| 64 |
+
diff_html += f'<h3>Pair {i}</h3>\n'
|
| 65 |
+
d = difflib.HtmlDiff()
|
| 66 |
+
diff_table = d.make_table(text.splitlines(), text_processed.splitlines())
|
| 67 |
+
diff_html += diff_table
|
| 68 |
+
diff_html += '<br>\n'
|
| 69 |
+
|
| 70 |
+
diff_html += '</body>\n</html>'
|
| 71 |
+
|
| 72 |
+
if output_file != '':
|
| 73 |
+
with open(output_file, 'w') as f:
|
| 74 |
+
f.write(diff_html)
|
| 75 |
+
|
| 76 |
+
return diff_html
|