app.py

import gradio as gr
from transformers import AutoModelForCausalLM, AutoTokenizer
from transformers import TapexTokenizer, BartForConditionalGeneration
import pandas as pd
import torch

import numpy as np
import time
import os
#import pkg_resources

'''
# Get a list of installed packages and their versions
installed_packages = {pkg.key: pkg.version for pkg in pkg_resources.working_set}

# Print the list of packages
for package, version in installed_packages.items():
    print(f"{package}=={version}")
'''

# Load the chatbot model
chatbot_model_name = "microsoft/DialoGPT-medium" #"gpt2"
chatbot_tokenizer = AutoTokenizer.from_pretrained(chatbot_model_name)
chatbot_model = AutoModelForCausalLM.from_pretrained(chatbot_model_name)


# Load the SQL Model
#wikisql take longer to process
#model_name = "microsoft/tapex-large-finetuned-wikisql"  # You can change this to any other model from the list above
#model_name = "microsoft/tapex-base-finetuned-wikisql"
#model_name = "microsoft/tapex-base-finetuned-wtq"
model_name = "microsoft/tapex-large-finetuned-wtq"
#model_name = "google/tapas-base-finetuned-wtq"
sql_tokenizer = TapexTokenizer.from_pretrained(model_name)
sql_model = BartForConditionalGeneration.from_pretrained(model_name)

data = {
    "year": [1896, 1900, 1904, 2004, 2008, 2012],
    "city": ["athens", "paris", "st. louis", "athens", "beijing", "london"]
}
table = pd.DataFrame.from_dict(data)

chat_history_ids = None
bot_input_ids = None


def chatbot_response(user_message):

    global chat_history_ids
    global bot_input_ids
    # Check if the user input is a question
    is_question = "?" in user_message

    if is_question:  
        # If the user input is a question, use TAPEx for question-answering
        #inputs = user_query
        encoding = sql_tokenizer(table=table, query=user_message, return_tensors="pt")
        outputs = sql_model.generate(**encoding)
        response = sql_tokenizer.batch_decode(outputs, skip_special_tokens=True)
    else:
        # Generate chatbot response using the chatbot model
        '''
        inputs = chatbot_tokenizer.encode("User: " + user_message, return_tensors="pt")
        outputs = chatbot_model.generate(inputs, max_length=100, num_return_sequences=1)
        response = chatbot_tokenizer.decode(outputs[0], skip_special_tokens=True)
        '''
        # encode the new user input, add the eos_token and return a tensor in Pytorch
        new_user_input_ids = chatbot_tokenizer.encode(user_message + chatbot_tokenizer.eos_token, return_tensors='pt')
    
        # append the new user input tokens to the chat history
        if chat_history_ids is not None:
            bot_input_ids = torch.cat([chat_history_ids, new_user_input_ids], dim=-1)
        else:
            bot_input_ids = new_user_input_ids
        
        # generated a response while limiting the total chat history to 1000 tokens, 
        chat_history_ids = chatbot_model.generate(bot_input_ids, max_length=1000, pad_token_id=chatbot_tokenizer.eos_token_id)

        response = chatbot_tokenizer.decode(chat_history_ids[:, bot_input_ids.shape[-1]:][0], skip_special_tokens=True)
    
    return response

# Define the chatbot and SQL execution interfaces using Gradio
chatbot_interface = gr.Interface(
    fn=chatbot_response,
    inputs=gr.Textbox(prompt="You:"),
    outputs=gr.Textbox(),
    live=True,
    capture_session=True,
    title="ST Chatbot",
    description="Type your message in the box above, and the chatbot will respond.",
)

# Launch the Gradio interface
if __name__ == "__main__":
    chatbot_interface.launch()