files are updated

app.py

@@ -1,6 +1,7 @@
 import os
 import streamlit as st
-from langchain.vectorstores import Chroma
+import chromadb
+from langchain.vectorstores import FAISS, Chroma
 from langchain.prompts import ChatPromptTemplate 
 from langchain import HuggingFaceHub
 from langchain.chains import RetrievalQA, RetrievalQAWithSourcesChain, ConversationalRetrievalChain
@@ -10,28 +11,30 @@
 # Streamlit interface for uploading PDF
 st.title("RAG-Based Chat Bot 🤖")
 uploaded_file = st.file_uploader("Upload a PDF", type=["pdf"])
+main_placeholder = st.empty()
 
 if uploaded_file is not None:
     # Step 1: Load PDF content
     pdf_content = load_pdf(uploaded_file)
-    st.write("PDF loaded successfully!")
+    main_placeholder.write("PDF loaded successfully!")
 
     # Step 2: Split the text into chunks
     documents = split_text(pdf_content)
-    st.write(f"Text split into {len(documents)} chunks.")
+    # st.write(f"Text split into {len(documents)} chunks.")
 
     # Step 3: Generate embeddings for the documents
     embeddings = generate_embeddings()  # Pass documents to the function
-    st.write("Embeddings generated for the documents.")
+    # st.write("Embeddings generated for the documents.")
 
     # Step 4: Save the vectorized documents to Weaviate
     vector_db = save_to_vectordb(documents, embeddings)
-    st.success("Documents are vectorized and saved to CromaDB.")
+    # st.success("Documents are vectorized and saved to CromaDB.")
 
     # Step 5: Create the HuggingFace model
-    st.write("Attempting to create the HuggingFace model...")
+    # st.write("Attempting to create the HuggingFace model...")
     model = create_huggingface_model()
-    st.success("Model created successfully!")
+    # st.success("Model created successfully!")
+
 
 
     retriever = vector_db.as_retriever()
@@ -47,52 +50,28 @@
     prompt = ChatPromptTemplate.from_template(template)
 
    # Defining a QnA chain
-    QnA = ConversationalRetrievalChain.from_llm(llm = model,
+    QnA = RetrievalQA.from_chain_type(llm = model,
                                 chain_type = 'stuff',
                                 retriever = retriever,
                                 verbose = False)
-    st.success("RAG pipeline created successfully!")
+
+    # st.success("RAG pipeline created successfully!")
 
     # User Input for Questions
     question = st.text_input("Ask a question")
 
     def get_answers(QnA, query):
-        answer = QnA.run(query)
-        # print(f"\033[1mQuery:\033[0m {query}\n")
-        return (f"\033[1mAnswer:\033[0m ", answer)
+        response = QnA.run(query)
+        return response
+        # if "Answer:" in response:
+        #     answer = response.split("Answer:", 1)[1].strip()
+        #     return answer
 
     if question:
         with st.spinner("Processing your question..."):
             answer = get_answers(QnA, question)
-            st.write("### Answer:")
+            # st.write("### Answer:")
             st.write(answer)
 
 else:
     st.warning("Please upload a PDF file to proceed.")
-
-
-    # def process_rag_response(question):
-
-    #     try:
-    #         # Invoke the RAG pipeline directly with the question
-    #         response = rag_chain.invoke(question)
-
-    #         # Ensure response is a string
-    #         if not isinstance(response, str):
-    #             response = str(response)
-
-    #         # Check if the response contains "Answer:" and extract it
-    #         if "Answer:" in response:
-    #             answer = response.split("Answer:", 1)[1].strip()  # Split and extract the answer
-    #             return answer
-    #         else:
-    #             return "Sorry, I didn’t understand your question. Do you want to connect with a live agent?"
-
-    #     except Exception as e:
-    #         # Handle any exceptions that occur
-    #         import traceback
-    #         st.error(f"An error occurred: {e}")
-    #         print(traceback.format_exc())  # Log the error for debugging
-    #         return "Sorry, there was an error processing your request."
-
-    # Handle user input and display the answer

modules.py

@@ -5,7 +5,8 @@
 from langchain.embeddings import HuggingFaceEmbeddings
 from langchain.llms import CTransformers
 import faiss
-from langchain.vectorstores import FAISS
+import chromadb
+from langchain.vectorstores import FAISS, Chroma
 import tempfile
 
 def load_pdf(uploaded_file):
@@ -37,14 +38,14 @@ def save_to_vectordb(docs, embeddings):
     """
     Saves the vectorized documents to Weaviate vector database.
     """
-    vector_db = FAISS.from_documents(
+    vector_db = Chroma.from_documents(
         docs,
         embeddings
     )
     return vector_db
 
 # Function to initialize the HuggingFace model
-def create_huggingface_model(temperature: float = 1.0, context_length: int = 1000, max_new_tokens= 600):
+def create_huggingface_model(temperature: float = 0, context_length: int = 4096, max_new_tokens= 2048):
 
     try:
         llm = CTransformers(model='TheBloke/Llama-2-7B-Chat-GGML',

requirements.txt

@@ -0,0 +1,12 @@
+chromadb==0.6.1
+ctransformers==0.2.27
+faiss-cpu==1.9.0.post1
+huggingface-hub==0.27.0
+langchain==0.3.13
+langchain-community==0.3.13
+langchain-core==0.3.28
+langchain-text-splitters==0.3.4
+pypdf==5.1.0
+sentence-transformers==3.3.1
+streamlit==1.41.1
+transformers==4.47.1