mintlee commited on
Commit
c4810eb
·
1 Parent(s): 4d949ca

add choose lang

Browse files
db/__pycache__/mongodb.cpython-310.pyc ADDED
Binary file (5.88 kB). View file
 
excel/__pycache__/excel_translate.cpython-310.pyc ADDED
Binary file (4.03 kB). View file
 
excel/excel_translate.py CHANGED
@@ -7,7 +7,6 @@ import io
7
  import pandas as pd
8
  import pymongo
9
  import gridfs
10
- from io import BytesIO
11
  import tempfile
12
  import os
13
 
 
7
  import pandas as pd
8
  import pymongo
9
  import gridfs
 
10
  import tempfile
11
  import os
12
 
pages/upload.py CHANGED
@@ -25,6 +25,7 @@ st.title("Upload PPTX to MongoDB")
25
  # st.set_option("server.fileUploader.allowMediaFiles", True)
26
  uploaded_file = st.file_uploader("Chọn file")
27
  file_name_input = st.text_input("Tên file để lưu (không cần .pptx)", value="")
 
28
 
29
  final_pptx_id = None # Biến lưu ID file sau khi xử lý
30
 
@@ -39,7 +40,7 @@ if uploaded_file is not None:
39
 
40
  xml_file_id = ppt_to_xml_mongodb(file_id)
41
  text_dict = extract_text_from_xml(file_id=xml_file_id)
42
- translated_dict = translate_text_dict(text_dict, source_lang="VietNamese", target_lang="English", gemini_api=api_key)
43
 
44
  final_xml_id = update_xml_with_translated_text_mongodb(xml_file_id, translated_dict)
45
  st.write(f"Final XML ID: {final_xml_id}")
 
25
  # st.set_option("server.fileUploader.allowMediaFiles", True)
26
  uploaded_file = st.file_uploader("Chọn file")
27
  file_name_input = st.text_input("Tên file để lưu (không cần .pptx)", value="")
28
+ target_lang = st.selectbox("Chọn ngôn ngữ dịch", ["english", "vietnamese"])
29
 
30
  final_pptx_id = None # Biến lưu ID file sau khi xử lý
31
 
 
40
 
41
  xml_file_id = ppt_to_xml_mongodb(file_id)
42
  text_dict = extract_text_from_xml(file_id=xml_file_id)
43
+ translated_dict = translate_text_dict(text_dict, target_lang=target_lang, gemini_api=api_key)
44
 
45
  final_xml_id = update_xml_with_translated_text_mongodb(xml_file_id, translated_dict)
46
  st.write(f"Final XML ID: {final_xml_id}")
powerpoint/__pycache__/__init__.cpython-310.pyc ADDED
Binary file (157 Bytes). View file
 
powerpoint/__pycache__/pptx_object.cpython-310.pyc ADDED
Binary file (8.97 kB). View file
 
powerpoint/__pycache__/xml_handling.cpython-310.pyc ADDED
Binary file (9.66 kB). View file
 
translate/__pycache__/translator.cpython-310.pyc ADDED
Binary file (2.92 kB). View file
 
translate/translator.py CHANGED
@@ -6,8 +6,7 @@ import google.generativeai as genai
6
  def translate_text_dict(text_dict: Dict[str, List[str]], source_lang: str, target_lang: str = "vi", gemini_api: str = "") -> Dict[str, List[str]]:
7
  def translate_batch(batch_dict: Dict[str, List[str]]) -> Dict[str, List[str]]:
8
  """Translates a single batch of text."""
9
- prompt = f"""The following python dictionary contains pieces of text that form a whole document: {json.dumps(batch_dict)}
10
- The text is in {source_lang}, with a chance of there being phrases in other languages as well.
11
 
12
  Read through the entire dictionary, then translate the texts into {target_lang} so that the meaning is as close to the intended context as possible.
13
 
 
6
  def translate_text_dict(text_dict: Dict[str, List[str]], source_lang: str, target_lang: str = "vi", gemini_api: str = "") -> Dict[str, List[str]]:
7
  def translate_batch(batch_dict: Dict[str, List[str]]) -> Dict[str, List[str]]:
8
  """Translates a single batch of text."""
9
+ prompt = f"""The following python dictionary contains pieces of text that form a whole document: {json.dumps(batch_dict)}.
 
10
 
11
  Read through the entire dictionary, then translate the texts into {target_lang} so that the meaning is as close to the intended context as possible.
12
 
word/__pycache__/word_translate.cpython-310.pyc ADDED
Binary file (6.63 kB). View file