Update app.py

app.py

@@ -15,7 +15,7 @@ from dotenv import load_dotenv
 from langchain_community.embeddings import HuggingFaceEmbeddings
 import nltk
 from urllib.parse import urljoin, urlparse
-from langchain.memory import ConversationBufferMemory
+import faiss
 
 # Load environment variables (if needed for API keys)
 load_dotenv()
@@ -37,12 +37,6 @@ def preprocess_text(text):
     cleaned_text = " ".join([word for word in tokens if word not in STOPWORDS])  # Remove stopwords
     return cleaned_text
 
-# Function to Save Processed Data to a Document
-def save_data_to_document(data, filename="processed_data.json"):
-    with open(filename, 'w') as f:
-        json.dump(data, f, indent=4)
-    st.success(f"Data has been saved to {filename}")
-
 # Scrape Website with BeautifulSoup
 def scrape_website(url):
     visited_urls = set()
@@ -122,7 +116,7 @@ def create_faiss_with_uuid(text_chunks):
     return unique_id, faiss_directory  # Return the UUID and the directory path
 
 # Build Conversational Chain
-def get_conversational_chain(memory):
+def get_conversational_chain():
     prompt_template = """
     Answer the question as detailed as possible from the provided context. If the answer is not in
     provided context, just say, "answer is not available in the context." Don't provide the wrong answer.\n\n
@@ -131,27 +125,23 @@ def get_conversational_chain(memory):
 
     Answer:
     """
-    model = Ollama(model="phi")  # Initialize LLaMA model
+    model = Ollama(model="qwen2.5:0.5b")  # Initialize LLaMA model
     prompt = PromptTemplate(template=prompt_template, input_variables=["context", "question"])
     
-    # Add memory to the chain
-    chain = load_qa_chain(model, chain_type="stuff", prompt=prompt, memory=memory)
+    chain = load_qa_chain(model, chain_type="stuff", prompt=prompt)
 
     return chain
 
 # Handle User Input and Process Questions with UUID-based FAISS Index
-def user_input(user_question, faiss_directory, memory):
+def user_input(user_question, faiss_directory):
     # Load the FAISS index based on the given directory (UUID-based)
     new_db = FAISS.load_local(faiss_directory, embeddings, allow_dangerous_deserialization=True)
     
     # Perform similarity search and answer the user's question
     docs = new_db.similarity_search(user_question)
-    chain = get_conversational_chain(memory)
+    chain = get_conversational_chain()
     
-    # Update memory with the question and response
     response = chain({"input_documents": docs, "question": user_question}, return_only_outputs=True)
-    memory.save_context({"input": user_question}, {"output": response["output_text"]})
-    
     st.write("Reply: ", response["output_text"])
 
 # Main Function for Streamlit App
@@ -159,14 +149,11 @@ def main():
     st.set_page_config("Chat PDF & URL", layout="wide")
     st.header("Chat with PDF or URL using Ollama 💁")
 
-    # Initialize memory for conversation history
-    memory = ConversationBufferMemory(memory_key="chat_history", return_messages=True)
-
     user_question = st.text_input("Ask a Question from the Processed Data")
 
     if user_question and 'faiss_directory' in st.session_state:
         faiss_directory = st.session_state['faiss_directory']
-        user_input(user_question, faiss_directory, memory)
+        user_input(user_question, faiss_directory)
 
     with st.sidebar:
         st.title("Menu:")
@@ -182,10 +169,6 @@ def main():
                         text_chunks = get_text_chunks(raw_text)
                         unique_id, faiss_directory = create_faiss_with_uuid(text_chunks)
                         st.session_state['faiss_directory'] = faiss_directory
-                        
-                        # Save the cleaned PDF data to a document
-                        save_data_to_document({"pdf_data": raw_text}, f"pdf_data_{unique_id}.json")
-                        
                         st.success("PDF data is ready for queries!")
                     else:
                         st.error("No PDF files were uploaded.")
@@ -206,10 +189,7 @@ def main():
                             text_chunks = get_text_chunks(raw_text)
                             unique_id, faiss_directory = create_faiss_with_uuid(text_chunks)
                             st.session_state['faiss_directory'] = faiss_directory
-                            
-                            # Save the cleaned URL data to a document
-                            save_data_to_document({"url_data": scraped_data}, f"url_data_{unique_id}.json")
-                            
+
                             st.success("Scraped data is ready for queries!")
                         except Exception as e:
                             st.error(f"Failed to scrape or process data: {e}")