app.py

import time
import openai
import os
import pandas as pd
import numpy as np
import gradio as gr
from llama_index import GPTSimpleVectorIndex
from openai.embeddings_utils import get_embedding, cosine_similarity

openai.api_key = os.getenv("OPENAI_API_KEY")
passwd = os.getenv("PASSWD_SECRET")

title = "Confidential forensics tool with ChatGPT"
examples = ["Who is Phillip Allen?", "What the project in Austin is about?", "Give me more details about the real estate project"]

index = GPTSimpleVectorIndex.load_from_disk('email.json')
df = pd.read_parquet("email_200.parquet")
df2 = df["subject"].replace([None], "<no title>").head(10).to_frame()

def search_emails(opt, message, n=3):
    "Outputs the top n emails that match the most the pattern"
    if len(message.strip()) < 1:
        message = "Oops, it looks like your query was not valid. Please make sure you typed something in your text box and then try again."
    else:
        try:
            embedding = get_embedding(message)
            message = ""
            df['embedding'] = df['embedding'].str.decode('utf-8').astype(float)
            
            df['similarities'] = df.embedding.apply(func=(lambda x: cosine_similarity(x, embedding)))

            message_tmp = df.sort_values('similarities', ascending=False).head(n)
            message_tmp = [(row.file, row.body, row.similarities) for index, row in message_tmp.iterrows()]
            for msg in message_tmp:
                message += f"Mail ID: {msg[0]}\nContent: {msg[1].strip()}\nSimilarity score: {msg[2]}\n\n"
        except Exception as e:
            message = "An error occured when handling your query, please try again."
            print(e)
    return message, ""

def respond_upload(btn_upload, message, chat_history):
    time.sleep(2)
    message = "***File uploaded***"
    bot_message = "Your document has been uploaded and will be accounted for your queries."
    chat_history.append((message, bot_message))
    return btn_upload, "", chat_history

def respond2(message, chat_history, box, btn):
    if len(message.strip()) < 1:
        message = "***Empty***"
        bot_message = "Oops, it looks like your query was not valid. Please make sure you typed something in your text box and then try again."
    else:
        try:
            bot_message = str(index.query(message)).strip()
        except:
            bot_message = "An error occured when handling your query, please try again."
    chat_history.append((message, bot_message))
    return message, chat_history, box

def respond(message, chat_history):
    if len(message.strip()) < 1:
        message = "***Empty***"
        bot_message = "Oops, it looks like your query was not valid. Please make sure you typed something in your text box and then try again."
    else:
        try:
            bot_message = str(index.query(message)).strip()
        except:
            bot_message = "An error occured when handling your query, please try again."
    chat_history.append((message, bot_message))
    return "", chat_history

with gr.Blocks(title=title) as demo:
    gr.Markdown(
    """

    # """ + title + """
    """)
    dat = gr.Dataframe(
            value=df2,
            max_cols=1,
            max_rows=4,
            title="Documents loaded",
            overflow_row_behaviour="paginate",
    )
    btn_upload = gr.UploadButton("Upload a new document...", file_types=["text"])
    gr.Markdown(
    """
    ## Chatbot
    """)
    chatbot = gr.Chatbot().style(height=400)
    with gr.Row():
        with gr.Column(scale=0.85):
            msg = gr.Textbox(
                show_label=False,
                placeholder="Enter text and press enter, or click on Send.",
            ).style(container=False)
        with gr.Column(scale=0.15, min_width=0):
            btn_send = gr.Button("Send your query")
    with gr.Row():
        gr.Markdown(
        """
        Example of queries
        """)
        for ex in examples:
            btn = gr.Button(ex)
            btn.click(respond2, [btn, chatbot, msg], [btn, chatbot, msg])

    msg.submit(respond, [msg, chatbot], [msg, chatbot])
    btn_send.click(respond, [msg, chatbot], [msg, chatbot])
    btn_upload.upload(respond_upload, [btn_upload, msg, chatbot], [btn_upload, msg, chatbot])
    gr.Markdown(
    """

    ## Search the matching document
    """)
    opt = gr.Textbox(
        show_label=False,
        placeholder="The document matching with your query will be shown here.",
        interactive=False,
        lines=8
    )
    with gr.Row():
        with gr.Column(scale=0.85):
            msg2 = gr.Textbox(
                show_label=False,
                placeholder="Enter text and press enter, or click on Send.",
            ).style(container=False)
        with gr.Column(scale=0.15, min_width=0):
            btn_send2 = gr.Button("Send your query")

    btn_send2.click(search_emails, [opt, msg2], [opt, msg2])
    
if __name__ == "__main__":
    demo.launch(auth=("mithril", passwd))