Hieucyber2208 commited on
Commit
9963898
·
verified ·
1 Parent(s): 8531f55

Update src/text_processing.py

Browse files
Files changed (1) hide show
  1. src/text_processing.py +1 -1
src/text_processing.py CHANGED
@@ -153,7 +153,7 @@ def text_processing(file_path, chunks, analysis_level='basic', writting_style='a
153
  if sentence: # Kiểm tra nếu câu không rỗng
154
  output_file = os.path.join(output_dir, f"{chunk_idx}_{sentence_idx}.txt") # Tên file dạng "chunkID_sentenceID.txt"
155
  with open(output_file, "w", encoding="utf-8") as f:
156
- f.write(sentence.replace("*","") + ".") # Giữ dấu chấm cuối câu
157
  print(f"Đã lưu: {output_file}")
158
  if __name__ == "__main__":
159
  text_processing("phan-tich-hinh-tuong-nguoi-lai-do-song-da-2.pdf", chunks = 3)
 
153
  if sentence: # Kiểm tra nếu câu không rỗng
154
  output_file = os.path.join(output_dir, f"{chunk_idx}_{sentence_idx}.txt") # Tên file dạng "chunkID_sentenceID.txt"
155
  with open(output_file, "w", encoding="utf-8") as f:
156
+ f.write(sentence.replace("*","").replace("#","") + ".") # Giữ dấu chấm cuối câu
157
  print(f"Đã lưu: {output_file}")
158
  if __name__ == "__main__":
159
  text_processing("phan-tich-hinh-tuong-nguoi-lai-do-song-da-2.pdf", chunks = 3)