app.py

"""
Credit to Derek Thomas, [email protected]
"""

import subprocess

# subprocess.run(["pip", "install", "--upgrade", "transformers[torch,sentencepiece]==4.34.1"])

import logging
from pathlib import Path
from time import perf_counter

import gradio as gr
from jinja2 import Environment, FileSystemLoader
import numpy as np
from sentence_transformers import CrossEncoder

from backend.query_llm import generate_hf, generate_openai
from backend.semantic_search import table, retriever

VECTOR_COLUMN_NAME = "vector"
TEXT_COLUMN_NAME = "text"

proj_dir = Path(__file__).parent
# Setting up the logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

# Set up the template environment with the templates directory
env = Environment(loader=FileSystemLoader(proj_dir / 'templates'))

# Load the templates directly from the environment
template = env.get_template('template.j2')
template_html = env.get_template('template_html.j2')

# crossEncoder
#cross_encoder = CrossEncoder('cross-encoder/ms-marco-MiniLM-L-6-v2') 
#cross_encoder = CrossEncoder('BAAI/bge-reranker-base')
# Examples
examples = ['what is social media and what are rules related to it for expenditure monitoring ,
            'how many reports to be submitted by Expenditure observer with annexure names ?',
            ]


def add_text(history, text):
    history = [] if history is None else history
    history = history + [(text, None)]
    return history, gr.Textbox(value="", interactive=False)


def bot(history, cross_encoder):
    top_rerank = 15
    top_k_rank = 10
    query = history[-1][0]

    if not query:
         gr.Warning("Please submit a non-empty string as a prompt")
         raise ValueError("Empty string was submitted")

    logger.warning('Retrieving documents...')
    # Retrieve documents relevant to query
    document_start = perf_counter()

    query_vec = retriever.encode(query)
    logger.warning(f'Finished query vec')
    doc1 = table.search(query_vec, vector_column_name=VECTOR_COLUMN_NAME).limit(top_k_rank)

    

    logger.warning(f'Finished search')
    documents = table.search(query_vec, vector_column_name=VECTOR_COLUMN_NAME).limit(top_rerank).to_list()
    documents = [doc[TEXT_COLUMN_NAME] for doc in documents]
    logger.warning(f'start cross encoder {len(documents)}')
    # Retrieve documents relevant to query
    query_doc_pair = [[query, doc] for doc in documents]
    if cross_encoder=='MiniLM-L6v2' :
           cross_encoder = CrossEncoder('cross-encoder/ms-marco-MiniLM-L-6-v2') 
    else:
           cross_encoder = CrossEncoder('BAAI/bge-reranker-base')
    cross_scores = cross_encoder.predict(query_doc_pair)
    sim_scores_argsort = list(reversed(np.argsort(cross_scores)))
    logger.warning(f'Finished cross encoder {len(documents)}')
    
    documents = [documents[idx] for idx in sim_scores_argsort[:top_k_rank]]
    logger.warning(f'num documents {len(documents)}')

    document_time = perf_counter() - document_start
    logger.warning(f'Finished Retrieving documents in {round(document_time, 2)} seconds...')

    # Create Prompt
    prompt = template.render(documents=documents, query=query)
    prompt_html = template_html.render(documents=documents, query=query)

    generate_fn = generate_hf

    history[-1][1] = ""
    for character in generate_fn(prompt, history[:-1]):
        history[-1][1] = character
        print('Final history is ',history)
        yield history, prompt_html


with gr.Blocks(theme='Insuz/SimpleIndigo') as demo:
    gr.HTML(value="""<div style="display: flex; align-items: center; justify-content: space-between;">
    <h1 style="color: #008000">NIRVACHANA - <span style="color: #008000">Expenditure Observer AI Assistant</span></h1>
    <img src='logo.png' alt="Chatbot" width="50" height="50" />
    </div>""",elem_id='heading')
    gr.HTML(value="""<p style="font-family: sans-serif; font-size: 16px;">A free chat bot assistant for Expenditure Observers on Compendium on Election Expenditure Monitoring using Open source LLMs. <br> The bot can answer questions in natural language, taking relevant extracts from the ECI document which can be accessed <a href="https://www.eci.gov.in/eci-backend/public/api/download?url=LMAhAK6sOPBp%2FNFF0iRfXbEB1EVSLT41NNLRjYNJJP1KivrUxbfqkDatmHy12e%2Fzk1vx4ptJpQsKYHA87guoLjnPUWtHeZgKtEqs%2FyzfTTYIC0newOHHOjl1rl0u3mJBSIq%2Fi7zDsrcP74v%2FKr8UNw%3D%3D" style="color: #008000; text-decoration: none;">here</a>.</p>""",elem_id='Sub-heading')
    

    chatbot = gr.Chatbot(
            [],
            elem_id="chatbot",
            avatar_images=('https://aui.atlassian.com/aui/8.8/docs/images/avatar-person.svg',
                           'https://huggingface.co/datasets/huggingface/brand-assets/resolve/main/hf-logo.svg'),
            bubble_full_width=False,
            show_copy_button=True,
            show_share_button=True,
            )

    with gr.Row():
        txt = gr.Textbox(
                scale=3,
                show_label=False,
                placeholder="Enter text and press enter",
                container=False,
                )
        txt_btn = gr.Button(value="Submit text", scale=1)

    cross_encoder = gr.Radio(choices=['MiniLM-L6v2','BGE reranker'], value='BGE reranker',label="Embeddings", info="Choose MiniLM for Speed, BGE reranker for accuracy")

    prompt_html = gr.HTML()
    # Turn off interactivity while generating if you click
    txt_msg = txt_btn.click(add_text, [chatbot, txt], [chatbot, txt], queue=False).then(
            bot, [chatbot, cross_encoder], [chatbot, prompt_html])

    # Turn it back on
    txt_msg.then(lambda: gr.Textbox(interactive=True), None, [txt], queue=False)

    # Turn off interactivity while generating if you hit enter
    txt_msg = txt.submit(add_text, [chatbot, txt], [chatbot, txt], queue=False).then(
            bot, [chatbot, cross_encoder], [chatbot, prompt_html])

    # Turn it back on
    txt_msg.then(lambda: gr.Textbox(interactive=True), None, [txt], queue=False)

    # Examples
    gr.Examples(examples, txt)

demo.queue()
demo.launch(debug=True)