Streamlit chatbot powered by Azure OpenAI

chatbot.py

"""
MIT License

Copyright (c) 2023 Saverio Proto

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
"""

"""
create a ".env" file with the following:
    OPENAI_API_TYPE = azure
    OPENAI_API_VERSION = 2023-03-15-preview
    OPENAI_API_BASE = 'https://eastus.api.cognitive.microsoft.com/' # Replace with the URL of an Azure OpenAI
    OPENAI_API_KEY = '' # Replace with the corresponding API key

To run the application, use the following command:
streamlit run chatbot.py
"""

import os
import sys
import logging
from langchain.chat_models import AzureChatOpenAI
from langchain.embeddings import OpenAIEmbeddings

import streamlit as st
from llama_index import (
    download_loader,
    SimpleDirectoryReader,
    LLMPredictor,
    GPTVectorStoreIndex,
    PromptHelper,
    ServiceContext,
    StorageContext,
    load_index_from_storage,
    LangchainEmbedding
)
from llama_index.logger import LlamaLogger

from dotenv import load_dotenv, dotenv_values


logging.basicConfig(stream=sys.stdout, level=logging.INFO)
logging.getLogger("llama_index").setLevel(logging.DEBUG)

index = None
doc_path = "./data/"
index_file = "index.json"

if "config" not in st.session_state:
    # Read the environment variables
    load_dotenv()
    config = dotenv_values(".env")
    st.session_state.config = config

if "response" not in st.session_state:
    st.session_state.response = ""

# save current file name to avoid reprocessing document
if "current_file" not in st.session_state:
    st.session_state.current_file = None

def send_click():
    query_engine = index.as_query_engine()
    # answer = query_engine.query(st.session_state.prompt)
    # st.session_state.response = answer.get_formatted_sources()
    st.session_state.response = query_engine.query(st.session_state.prompt)
    st.session_state.lamalogs = service_context.llama_logger.get_logs()

st.title("Azure OpenAI Doc Chatbot")

sidebar_placeholder = st.sidebar.container()

uploaded_file = st.file_uploader("Choose a file")

# Create the chat llm
llm = AzureChatOpenAI(
    deployment_name="gpt-35-turbo",
    model_kwargs={
        "api_key": st.session_state.config["OPENAI_API_KEY"],
        "api_base": st.session_state.config["OPENAI_API_BASE"],
        "api_type": st.session_state.config["OPENAI_API_TYPE"],
        "api_version": st.session_state.config["OPENAI_API_VERSION"],
    },
)

# Create the embedding llm
embedding_llm = LangchainEmbedding(
    OpenAIEmbeddings(
        model="text-embedding-ada-002",
        deployment="text-embedding-ada-002",
        openai_api_key=st.session_state.config["OPENAI_API_KEY"],
        openai_api_base=st.session_state.config["OPENAI_API_BASE"],
        openai_api_type=st.session_state.config["OPENAI_API_TYPE"],
        openai_api_version=st.session_state.config["OPENAI_API_VERSION"],
    ),
    embed_batch_size=1,
)

# Create llama_index LLMPredictor
llm_predictor = LLMPredictor(llm=llm)
max_input_size = 4096
num_output = 256
max_chunk_overlap = 20
prompt_helper = PromptHelper(max_input_size, num_output, max_chunk_overlap)

llama_logger = LlamaLogger()

# Create llama_index ServiceContext
service_context = ServiceContext.from_defaults(
    llm_predictor=llm_predictor,
    prompt_helper=prompt_helper,
    embed_model=embedding_llm,
    chunk_size_limit=1000,
    llama_logger=llama_logger
)

if uploaded_file and uploaded_file.name != st.session_state.current_file:
    # Ingest the document and create the index
    with st.spinner('Ingesting the file..'):
        doc_files = os.listdir(doc_path)
        for doc_file in doc_files:
            os.remove(doc_path + doc_file)

        bytes_data = uploaded_file.read()
        with open(f"{doc_path}{uploaded_file.name}", "wb") as f:
            f.write(bytes_data)

        loader = SimpleDirectoryReader(doc_path, recursive=True, exclude_hidden=True)
        documents = loader.load_data()
        sidebar_placeholder.header("Current Processing Document:")
        sidebar_placeholder.subheader(uploaded_file.name)
        sidebar_placeholder.write(documents[0].get_text()[:500] + "...")

        index = GPTVectorStoreIndex.from_documents(
            documents, service_context=service_context
        )

        index.set_index_id("vector_index")
        index.storage_context.persist(index_file)
        st.session_state.current_file = uploaded_file.name
        st.session_state.response = ""  # clean up the response when new file is uploaded
    st.success('Done!')

elif os.path.exists(index_file):
    # Read from storage context
    storage_context = StorageContext.from_defaults(persist_dir=index_file)
    index = load_index_from_storage(
        storage_context, index_id="vector_index", service_context=service_context
    )

    SimpleDirectoryReader = download_loader("SimpleDirectoryReader")
    loader = SimpleDirectoryReader(doc_path, recursive=True, exclude_hidden=True)
    documents = loader.load_data()
    doc_filename = os.listdir(doc_path)[0]
    sidebar_placeholder.header("Current Processing Document:")
    sidebar_placeholder.subheader(doc_filename)
    sidebar_placeholder.write(documents[0].get_text()[:500] + "...")

if index:
    st.text_input("Ask something: ", key="prompt", on_change=send_click)
    st.button("Send", on_click=send_click)
    if st.session_state.response:
        st.subheader("Response: ")
        st.success(st.session_state.response, icon="🤖")
        st.subheader("Debug information: ")
        st.write("This is the formatted prompt template:")
        st.code(st.session_state.lamalogs[0]['formatted_prompt_template'])
        st.write("This is the initial response:")
        st.code(st.session_state.lamalogs[1]['initial_response'])