Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -500,17 +500,24 @@ def create_new_blank_record(container):
|
|
500 |
|
501 |
# Function to preprocess the pasted content
|
502 |
def preprocess_text(text):
|
503 |
-
# Replace CRLF with
|
504 |
text = text.replace('\r\n', '\\n')
|
505 |
text = text.replace('\r', '\\n')
|
506 |
text = text.replace('\n', '\\n')
|
507 |
-
|
508 |
-
#
|
509 |
-
|
510 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
511 |
return text
|
512 |
|
513 |
-
|
514 |
|
515 |
# 🎭 Main function - "All the world's a stage, and all the code merely players" -Shakespeare, probably
|
516 |
def main():
|
|
|
500 |
|
501 |
# Function to preprocess the pasted content
|
502 |
def preprocess_text(text):
|
503 |
+
# Replace CRLF and other newline variations with the JSON newline escape sequence
|
504 |
text = text.replace('\r\n', '\\n')
|
505 |
text = text.replace('\r', '\\n')
|
506 |
text = text.replace('\n', '\\n')
|
507 |
+
|
508 |
+
# Escape double quotes inside the text
|
509 |
+
text = text.replace('"', '\\"')
|
510 |
+
|
511 |
+
# Optionally remove or handle other special characters that might not be JSON-safe
|
512 |
+
# Here, we remove characters like tabs or non-ASCII characters (as an example)
|
513 |
+
text = re.sub(r'[\t]', ' ', text) # Replace tabs with spaces
|
514 |
+
text = re.sub(r'[^\x00-\x7F]+', '', text) # Remove non-ASCII characters
|
515 |
+
|
516 |
+
# Normalize spaces (strip leading/trailing whitespace)
|
517 |
+
text = text.strip()
|
518 |
return text
|
519 |
|
520 |
+
|
521 |
|
522 |
# 🎭 Main function - "All the world's a stage, and all the code merely players" -Shakespeare, probably
|
523 |
def main():
|