Update pages/llm.py

pages/llm.py

@@ -7,7 +7,7 @@ import os
 from PyPDF2 import PdfReader
 from transformers import pipeline
 from transformers import AutoModel
-#from googletrans import Translator
+from googletrans import Translator
 #from transformers import *
 
 
@@ -22,7 +22,7 @@ from transformers import AutoModel
 
 # PDF in String umwandeln
 def get_pdf_text(folder_path):
-    #translator = Translator()
+    translator = Translator()
     text = ""
     # Durchsuche alle Dateien im angegebenen Verzeichnis
     for filename in os.listdir(folder_path):
@@ -34,9 +34,9 @@ def get_pdf_text(folder_path):
             for page in pdf_reader.pages:
                 text += page.extract_text()
             #text += '\n'
-    #text=text.replace("\n", " ")
-    #text=text.replace("- ", "")
-    return text#translator.translate(text, dest ='en').text
+    text=text.replace("\n", " ")
+    text=text.replace("- ", "")
+    return translator.translate(text, dest ='en').text
 
 #Chunks erstellen
 def get_text_chunks(text):
@@ -81,8 +81,8 @@ def get_llm_answer(user_question):
     #user_question = st.text_area("Stell mir eine Frage: ")
             #if os.path.exists("./Store"): #Nutzereingabe nur eingelesen, wenn vectorstore angelegt
     # Retriever sucht passende Textausschnitte in den PDFs (unformatiert)
-    #translator = Translator()
-    #translator.translate(user_question, dest='en')
+    translator = Translator()
+    translator.translate(user_question, dest='en')
     retriever=get_vectorstore().as_retriever()
     retrieved_docs=retriever.invoke(
     user_question
@@ -99,7 +99,7 @@ def get_llm_answer(user_question):
     # Frage beantworten mit Q&A Pipeline
     answer = qa_pipeline(question=user_question, context=context, max_length=200)
     
-    return answer["answer"]#translator.translate(answer["answer"],dest='de')
+    return translator.translate(answer["answer"],dest='de')
    
 def main():
     st.set_page_config(