app.py

import json
import os

import streamlit as st
import streamlit.components.v1 as components
from langchain.callbacks import get_openai_callback
from langchain.chains import ConversationalRetrievalChain
from langchain.chains.conversation.memory import ConversationBufferMemory
from langchain.embeddings import GPT4AllEmbeddings
from langchain.llms import OpenAI

from chat_history import insert_chat_history, insert_chat_history_articles
from connection import connect
from css import load_css
from message import Message
from vector_store import CustomVectorStore

st.set_page_config(layout="wide")

st.title("Sorbobot - Le futur de la recherche scientifique interactive")

chat_column, doc_column = st.columns([2, 1])

conn = connect()


def initialize_session_state():
    if "history" not in st.session_state:
        st.session_state.history = []
    if "token_count" not in st.session_state:
        st.session_state.token_count = 0
    if "conversation" not in st.session_state:
        embeddings = GPT4AllEmbeddings()

        db = CustomVectorStore(
            embedding_function=embeddings,
            table_name="article",
            column_name="abstract_embedding",
            connection=conn,
        )

        retriever = db.as_retriever()

        llm = OpenAI(
            temperature=0,
            openai_api_key=os.environ["OPENAI_API_KEY"],
            model="text-davinci-003",
        )

        memory = ConversationBufferMemory(
            output_key="answer", memory_key="chat_history", return_messages=True
        )
        st.session_state.conversation = ConversationalRetrievalChain.from_llm(
            llm=llm,
            retriever=retriever,
            verbose=True,
            memory=memory,
            return_source_documents=True,
        )


def send_message_callback():
    with st.spinner("Wait for it..."):
        with get_openai_callback() as cb:
            human_prompt = st.session_state.human_prompt.strip()
            if len(human_prompt) == 0:
                return
            llm_response = st.session_state.conversation(human_prompt)
            st.session_state.history.append(Message("human", human_prompt))
            st.session_state.history.append(
                Message(
                    "ai",
                    llm_response["answer"],
                    documents=llm_response["source_documents"],
                )
            )
            st.session_state.token_count += cb.total_tokens
            if os.environ.get("ENVIRONMENT") == "dev":
                history_id = insert_chat_history(conn, human_prompt, llm_response["answer"])
                insert_chat_history_articles(conn, history_id, llm_response["source_documents"])


def exemple_message_callback_button(args):
    st.session_state.human_prompt = args
    send_message_callback()
    st.session_state.human_prompt = ""


def clear_history():
    st.session_state.history.clear()
    st.session_state.token_count = 0
    st.session_state.conversation.memory.clear()


load_css()
initialize_session_state()

exemples = [
    "Who has published influential research on quantum computing?",
    "List any prominent authors in the field of artificial intelligence ethics?",
    "Who are the leading experts on climate change mitigation strategies?",
]

with chat_column:
    chat_placeholder = st.container()
    prompt_placeholder = st.form("chat-form", clear_on_submit=True)
    information_placeholder = st.container()

    with chat_placeholder:
        for chat in st.session_state.history:
            div = f"""
                <div class="chat-row 
                    {'' if chat.origin == 'ai' else 'row-reverse'}">
                    <img class="chat-icon" src="./app/static/{
                        'ai_icon.png' if chat.origin == 'ai' 
                                    else 'user_icon.png'}"
                        width=32 height=32>
                    <div class="chat-bubble
                    {'ai-bubble' if chat.origin == 'ai' else 'human-bubble'}">
                        &#8203;{chat.message}
                    </div>
                </div>
            """
            st.markdown(div, unsafe_allow_html=True)

        for _ in range(3):
            st.markdown("")

    with prompt_placeholder:
        st.markdown("**Chat**")
        cols = st.columns((6, 1))
        cols[0].text_input(
            "Chat",
            label_visibility="collapsed",
            key="human_prompt",
        )
        cols[1].form_submit_button(
            "Submit",
            type="primary",
            on_click=send_message_callback,
        )

    if st.session_state.token_count == 0:
        information_placeholder.markdown("### Test me !")
        for idx_exemple, exemple in enumerate(exemples):
            information_placeholder.button(
                exemple,
                key=f"{idx_exemple}_button",
                on_click=exemple_message_callback_button,
                args=(exemple,)
            )

    st.button(":new: Start a new conversation", on_click=clear_history, type="secondary")

    information_placeholder.caption(
        f"""
    Used {st.session_state.token_count} tokens \n
    Debug Langchain conversation: 
    {st.session_state.history}
    """
    )

    components.html(
        """
    <script>
    const streamlitDoc = window.parent.document;

    const buttons = Array.from(
        streamlitDoc.querySelectorAll('.stButton > button')
    );
    const submitButton = buttons.find(
        el => el.innerText === 'Submit'
    );

    streamlitDoc.addEventListener('keydown', function(e) {
        switch (e.key) {
            case 'Enter':
                submitButton.click();
                break;
        }
    });
    </script>
    """,
        height=0,
        width=0,
    )

with doc_column:
    st.markdown("**Source documents**")
    if len(st.session_state.history) > 0:
        for doc in st.session_state.history[-1].documents:
            doc_content = json.loads(doc.page_content)

            expander = st.expander(doc_content["title"])
            expander.markdown(f"**HalID** : https://hal.science/{doc_content['hal_id']}")
            expander.markdown(doc_content["abstract"])
            expander.markdown(f"**Authors** : {doc_content['authors']}")
            expander.markdown(f"**Keywords** : {doc_content['keywords']}")
            expander.markdown(f"**Distance** : {doc_content['distance']}")