app.py

import gradio as gr
import pandas as pd
from chatbot.bot import QuestionAnsweringBot
import os

current_dir = os.path.dirname(os.path.abspath(__file__))
data_path = os.path.join(current_dir, 'data', 'chunked_data_corpus.csv')

chunked_data_corpus_df = pd.read_csv(data_path)

bot = QuestionAnsweringBot(chunked_data_corpus_df)

def message_respond(message, history):
    answer = bot.form_response(message)
    return answer

gr.ChatInterface(
    fn=message_respond,
    type="messages",
    title="RAG System for 'The Count of Monte Cristo' book",
    description='''
    This project is aimed at developing RAG system for documents corpus. 
    To be more precise, this system operates with the english version of the book 'The Count of Monte Cristo'.

    API key for model requests is already provided.

    Short description:
    1. Core LLM: 'groq/llama3-8b-8192'
    2. Cross-encoder: 'cross-encoder/ms-marco-MiniLM-L-12-v2'
    3. Retriever: bm25 + bi-encoder('all-MiniLM-L6-v2') 

    The examples of the questions are provided below. Also, feel free to ask your own questions.
    
    Full code can be found on github via this link https://github.com/rDrayBen/Neural_networks_RAG
    ''',
    theme=gr.themes.Monochrome(font='Lora', text_size='lg', radius_size='sm'),
    examples=["Who is Monte Cristo?", 
              "What is the title of Chapter 93", 
              "Why Edmond Dantes was in prison?", 
              "How many years does Edmon Dantes spent in prison?",
              "Who said this sentence to whom 'Wait and hope_ (Fac et spera)'?",
              "What is the title of Chapter 64?",
              "Who is the president of the USA?",
              "Who is the author of the book The Count of Monte Cristo?",
              "Tell me about all the main identites in Monte Cristo?"
              ],
    cache_examples=False,
).launch()