Upload 5 files

app.py

@@ -0,0 +1,321 @@
+# version 2: added custom prompts.
+
+import os
+import json
+import sqlite3
+from datetime import datetime
+import streamlit as st
+from langchain_huggingface import HuggingFaceEmbeddings
+from langchain_chroma import Chroma
+from langchain_groq import ChatGroq
+from langchain.memory import ConversationBufferMemory
+from langchain.chains import ConversationalRetrievalChain
+from langchain.prompts import PromptTemplate
+
+from vectorize_documents import embeddings  # If needed elsewhere
+
+# Load config
+working_dir = os.path.dirname(os.path.abspath(__file__))
+config_data = json.load(open(f"{working_dir}/config.json"))
+GROQ_API_KEY = config_data["GROQ_API_KEY"]
+os.environ["GROQ_API_KEY"] = GROQ_API_KEY
+
+# Set up the database
+def setup_db():
+    conn = sqlite3.connect("chat_history.db", check_same_thread=False)
+    cursor = conn.cursor()
+    cursor.execute("""
+        CREATE TABLE IF NOT EXISTS chat_histories (
+            id INTEGER PRIMARY KEY AUTOINCREMENT,
+            username TEXT,
+            timestamp TEXT,
+            day TEXT,
+            user_message TEXT,
+            assistant_response TEXT
+        )
+    """)
+    conn.commit()
+    return conn
+
+# Set up vectorstore
+def setup_vectorstore():
+    embeddings = HuggingFaceEmbeddings()
+    vectorstore = Chroma(persist_directory="Vector_db", embedding_function=embeddings)
+    return vectorstore
+
+# Custom prompt template
+custom_prompt_template = PromptTemplate.from_template("""
+You are a helpful assistant that helps users choose laptops.
+
+1. Analyze the user's query, take information from vectordb and then give top 3 laptops to user from Relevent information that is context.                                                      
+2. Keep suggestions clear and concise with names, specs, and reasons only from relevant information context.                                                     
+
+Relevant Information:
+{context}
+
+Chat History:
+{chat_history}
+
+User Query:
+{question}
+
+Assistant Response:
+""")
+
+
+# Set up the chatbot chain with a specific model
+def chat_chain(vectorstore, model_name):
+    llm = ChatGroq(model=model_name, temperature=0.3)
+    retriever = vectorstore.as_retriever()
+    memory = ConversationBufferMemory(
+        llm=llm,
+        output_key="answer",
+        memory_key="chat_history",
+        return_messages=True
+    )
+    chain = ConversationalRetrievalChain.from_llm(
+        llm=llm,
+        retriever=retriever,
+        memory=memory,
+        combine_docs_chain_kwargs={"prompt": custom_prompt_template},
+        return_source_documents=True,
+        verbose=True
+    )
+    return chain
+
+# Streamlit UI setup
+st.set_page_config(page_title="ByteX-Ai", page_icon="🤖AI", layout="centered")
+st.title("🤖 ByteX-Ai")
+st.subheader("Hey! Get your Laptop!!")
+
+# Initialize DB connection
+if "conn" not in st.session_state:
+    st.session_state.conn = setup_db()
+
+# Prompt user to log in
+if "username" not in st.session_state:
+    username = st.text_input("Enter your name to proceed:")
+    if username:
+        with st.spinner("Loading chatbot interface... Please wait."):
+
+            st.session_state.username = username
+            st.session_state.chat_history = []
+            st.session_state.vectorstore = setup_vectorstore()
+            st.success(f"Welcome, {username}! Now select a model to start chatting.")
+else:
+    username = st.session_state.username
+
+# Model selection options
+model_options = [
+    "gemma2-9b-it",
+    "llama-3.1-8b-instant",
+    "llama3-70b-8192",
+    "llama3-8b-8192"
+]
+
+selected_model = st.selectbox("Choose a model:", model_options)
+
+# Ensure vectorstore exists
+if "vectorstore" not in st.session_state:
+    st.session_state.vectorstore = setup_vectorstore()
+
+# Set or update the selected model
+if "selected_model" not in st.session_state:
+    st.session_state.selected_model = selected_model
+
+# Reset conversational_chain if model changes or not yet initialized
+if ("conversational_chain" not in st.session_state) or (st.session_state.selected_model != selected_model):
+    st.session_state.selected_model = selected_model
+    st.session_state.conversational_chain = chat_chain(st.session_state.vectorstore, selected_model)
+    st.session_state.chat_history = []
+
+# Reset chat manually
+if st.button("🔄 Reset Chat"):
+    st.session_state.chat_history = []
+    st.session_state.conversational_chain = chat_chain(st.session_state.vectorstore, st.session_state.selected_model)
+    st.success("Chat reset!")
+
+# Show chat UI
+if "username" in st.session_state:
+    st.subheader(f"Hello {username}, start your query below!")
+
+    if st.session_state.chat_history:
+        for message in st.session_state.chat_history:
+            if message['role'] == 'user':
+                with st.chat_message("user"):
+                    st.markdown(message["content"])
+            elif message['role'] == 'assistant':
+                with st.chat_message("assistant"):
+                    st.markdown(message["content"])
+
+    user_input = st.chat_input("Ask AI....")
+
+    if user_input:
+        with st.spinner("Processing your query... Please wait."):
+            st.session_state.chat_history.append({"role": "user", "content": user_input})
+
+            with st.chat_message("user"):
+                st.markdown(user_input)
+
+            with st.chat_message("assistant"):
+                response = st.session_state.conversational_chain({"question": user_input})
+                assistant_response = response["answer"]
+                st.markdown(assistant_response)
+
+                st.session_state.chat_history.append({"role": "assistant", "content": assistant_response})
+
+
+
+
+
+
+
+
+
+
+
+# Version 1: working properly but there is no prompt refinement.
+
+# import os
+# import json
+# import sqlite3
+# from datetime import datetime
+# import streamlit as st
+# from langchain_huggingface import HuggingFaceEmbeddings
+# from langchain_chroma import Chroma
+# from langchain_groq import ChatGroq
+# from langchain.memory import ConversationBufferMemory
+# from langchain.chains import ConversationalRetrievalChain
+
+# from vectorize_documents import embeddings  # If needed elsewhere
+
+# # Load config
+# working_dir = os.path.dirname(os.path.abspath(__file__))
+# config_data = json.load(open(f"{working_dir}/config.json"))
+# GROQ_API_KEY = config_data["GROQ_API_KEY"]
+# os.environ["GROQ_API_KEY"] = GROQ_API_KEY
+
+# # Set up the database
+# def setup_db():
+#     conn = sqlite3.connect("chat_history.db", check_same_thread=False)
+#     cursor = conn.cursor()
+#     cursor.execute("""
+#         CREATE TABLE IF NOT EXISTS chat_histories (
+#             id INTEGER PRIMARY KEY AUTOINCREMENT,
+#             username TEXT,
+#             timestamp TEXT,
+#             day TEXT,
+#             user_message TEXT,
+#             assistant_response TEXT
+#         )
+#     """)
+#     conn.commit()
+#     return conn
+
+# # Set up vectorstore
+# def setup_vectorstore():
+#     embeddings = HuggingFaceEmbeddings()
+#     vectorstore = Chroma(persist_directory="Vector_db", embedding_function=embeddings)
+#     return vectorstore
+
+# # Set up the chatbot chain with a specific model
+# def chat_chain(vectorstore, model_name):
+#     llm = ChatGroq(model=model_name, temperature=0)
+#     retriever = vectorstore.as_retriever()
+#     memory = ConversationBufferMemory(
+#         llm=llm,
+#         output_key="answer",
+#         memory_key="chat_history",
+#         return_messages=True
+#     )
+#     chain = ConversationalRetrievalChain.from_llm(
+#         llm=llm,
+#         retriever=retriever,
+#         chain_type="stuff",
+#         memory=memory,
+#         verbose=True,
+#         return_source_documents=True
+#     )
+#     return chain
+
+# # Streamlit UI setup
+# st.set_page_config(page_title="ByteX-Ai", page_icon="🤖AI", layout="centered")
+# st.title("🤖 ByteX-Ai")
+# st.subheader("Hey! Get your Laptop!!")
+
+# # Initialize DB connection
+# if "conn" not in st.session_state:
+#     st.session_state.conn = setup_db()
+
+# # Prompt user to log in
+# if "username" not in st.session_state:
+#     username = st.text_input("Enter your name to proceed:")
+#     if username:
+#         with st.spinner("Loading chatbot interface... Please wait."):
+#             st.session_state.username = username
+#             st.session_state.chat_history = []
+#             st.session_state.vectorstore = setup_vectorstore()
+#             st.success(f"Welcome, {username}! Now select a model to start chatting.")
+# else:
+#     username = st.session_state.username
+
+# # Model selection options
+# model_options = [
+#     "gemma2-9b-it",
+#     "llama-3.1-8b-instant",
+#     "llama3-70b-8192",
+#     "llama3-8b-8192"
+# ]
+
+# # Model dropdown
+# selected_model = st.selectbox("Choose a model:", model_options)
+
+# # Ensure vectorstore exists
+# if "vectorstore" not in st.session_state:
+#     st.session_state.vectorstore = setup_vectorstore()
+
+# # Set or update the selected model
+# if "selected_model" not in st.session_state:
+#     st.session_state.selected_model = selected_model
+
+# # Reset conversational_chain if model changes or not yet initialized
+# if ("conversational_chain" not in st.session_state) or (st.session_state.selected_model != selected_model):
+#     st.session_state.selected_model = selected_model
+#     st.session_state.conversational_chain = chat_chain(st.session_state.vectorstore, selected_model)
+#     st.session_state.chat_history = []
+
+# # Reset chat manually
+# if st.button("🔄 Reset Chat"):
+#     st.session_state.chat_history = []
+#     st.session_state.conversational_chain = chat_chain(st.session_state.vectorstore, st.session_state.selected_model)
+#     st.success("Chat reset!")
+
+
+# # Show chat UI
+# if "username" in st.session_state:
+#     st.subheader(f"Hello {username}, start your query below!")
+
+#     if st.session_state.chat_history:
+#         for message in st.session_state.chat_history:
+#             if message['role'] == 'user':
+#                 with st.chat_message("user"):
+#                     st.markdown(message["content"])
+#             elif message['role'] == 'assistant':
+#                 with st.chat_message("assistant"):
+#                     st.markdown(message["content"])
+
+#     user_input = st.chat_input("Ask AI....")
+
+#     if user_input:
+#         with st.spinner("Processing your query... Please wait."):
+#             st.session_state.chat_history.append({"role": "user", "content": user_input})
+
+#             with st.chat_message("user"):
+#                 st.markdown(user_input)
+
+#             with st.chat_message("assistant"):
+#                 response = st.session_state.conversational_chain({"question": user_input})
+#                 assistant_response = response["answer"]
+#                 st.markdown(assistant_response)
+
+#                 st.session_state.chat_history.append({"role": "assistant", "content": assistant_response})

config.json

@@ -0,0 +1 @@
+{"GROQ_API_KEY": "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"}

requirements.txt

@@ -0,0 +1,13 @@
+fastapi
+uvicorn
+google-generativeai
+PyPDF2
+fpdf
+streamlit
+langchain-community==0.2.16
+langchain-text-splitters==0.2.4
+langchain-chroma==0.1.3
+langchain-huggingface==0.0.3
+langchain-groq==0.1.9
+unstructured==0.15.0
+unstructured[pdf]==0.15.0

vectorize_documents.py

@@ -0,0 +1,71 @@
+
+from langchain_text_splitters import CharacterTextSplitter
+from langchain_huggingface import HuggingFaceEmbeddings
+from langchain_chroma import Chroma
+from langchain.docstore.document import Document
+import pandas as pd
+import os
+import glob
+from PyPDF2 import PdfReader  # Ensure PyPDF2 is installed
+
+# Define a function to process CSV files
+def process_csv_files(csv_files):
+    documents = []
+    for file_path in csv_files:
+        df = pd.read_csv(file_path)
+        for _, row in df.iterrows():
+            row_content = " ".join(row.astype(str))
+            documents.append(Document(page_content=row_content))
+    return documents
+
+# Define a function to process PDF files
+def process_pdf_files(pdf_files):
+    documents = []
+    for file_path in pdf_files:
+        reader = PdfReader(file_path)
+        for page in reader.pages:
+            text = page.extract_text()
+            if text:  # Only add non-empty text
+                documents.append(Document(page_content=text))
+    return documents
+
+# Define a function to perform vectorization for CSV and PDF files
+def vectorize_documents():
+    embeddings = HuggingFaceEmbeddings()
+
+    # Directory containing files
+    data_directory = "Data"  # Replace with your folder name
+    csv_files = glob.glob(os.path.join(data_directory, "*.csv"))
+    pdf_files = glob.glob(os.path.join(data_directory, "*.pdf"))
+
+    # Process CSV and PDF files
+    documents = process_csv_files(csv_files) + process_pdf_files(pdf_files)
+
+    # Splitting the text and creating chunks of these documents
+    text_splitter = CharacterTextSplitter(
+        chunk_size=2000,
+        chunk_overlap=500
+    )
+
+    text_chunks = text_splitter.split_documents(documents)
+
+    # Process text chunks in batches
+    batch_size = 5000  # Chroma's batch size limit is 5461, set a slightly smaller size for safety
+    for i in range(0, len(text_chunks), batch_size):
+        batch = text_chunks[i:i + batch_size]
+
+        # Store the batch in Chroma vector DB
+        vectordb = Chroma.from_documents(
+            documents=batch,
+            embedding=embeddings,
+            persist_directory="Vector_db"
+        )
+
+    print("Documents Vectorized and saved in VectorDB")
+
+# Expose embeddings if needed
+embeddings = HuggingFaceEmbeddings()
+
+# Main guard to prevent execution on import
+if __name__ == "__main__":
+    vectorize_documents()