app.py

import os
import streamlit as st
from openai import OpenAI
from llama_index.node_parser import SemanticSplitterNodeParser
from llama_index.embeddings import OpenAIEmbedding
from llama_index.ingestion import IngestionPipeline
from pinecone.grpc import PineconeGRPC
from pinecone import ServerlessSpec
from llama_index.vector_stores import PineconeVectorStore
from llama_index import VectorStoreIndex
from llama_index.retrievers import VectorIndexRetriever
from llama_index.query_engine import RetrieverQueryEngine

# Set OpenAI API key from environment variables
openai_api_key = os.getenv("OPENAI_API_KEY")
pinecone_api_key = os.getenv("PINECONE_API_KEY")
index_name = os.getenv("INDEX_NAME")

# Initialize OpenAI client
client = OpenAI(api_key=openai_api_key)

# Initialize connection to Pinecone
pc = PineconeGRPC(api_key=pinecone_api_key)

# Initialize your index
if index_name not in pc.list_indexes():
    spec = ServerlessSpec(replicas=1, pod_type="p1")
    pc.create_index(name=index_name, dimension=1536, spec=spec)

pinecone_index = pc.Index(index_name)

# Initialize VectorStore
vector_store = PineconeVectorStore(pinecone_index=pinecone_index)
pinecone_index.describe_index_stats()

# Initialize vector index and retriever
vector_index = VectorStoreIndex.from_vector_store(vector_store=vector_store)
retriever = VectorIndexRetriever(index=vector_index, similarity_top_k=5)
query_engine = RetrieverQueryEngine(retriever=retriever)

# Set up LlamaIndex embedding model and pipeline
embed_model = OpenAIEmbedding(api_key=openai_api_key)
pipeline = IngestionPipeline(
    transformations=[
        SemanticSplitterNodeParser(buffer_size=1, breakpoint_percentile_threshold=95, embed_model=embed_model),
        embed_model,
    ],
)

def query_annual_report(query):
    response = query_engine.query(query)
    return response.response

# Streamlit app setup
st.title("ChatGPT-like Clone with Pinecone Integration")

# Initialize chat history
if "messages" not in st.session_state:
    st.session_state.messages = []

# Display chat messages from history
for message in st.session_state.messages:
    with st.chat_message(message["role"]):
        st.markdown(message["content"])

# Accept user input
if prompt := st.chat_input("What is up?"):
    st.session_state.messages.append({"role": "user", "content": prompt})
    with st.chat_message("user"):
        st.markdown(prompt)

    with st.chat_message("assistant"):
        response = query_annual_report(prompt)
        st.markdown(response)
    st.session_state.messages.append({"role": "assistant", "content": response})