app.py

import gradio as gr
import os
import PyPDF2
import pandas as pd
import docx
import json
import requests
from docx import Document
from langchain_community.vectorstores import FAISS
from langchain.text_splitter import RecursiveCharacterTextSplitter
from transformers import pipeline

# Configurar Hugging Face API Token
HF_API_TOKEN = os.getenv("HUGGINGFACE_API_TOKEN")

# Carregar o modelo Mistral 7B gratuitamente do Hugging Face
chatbot_pipeline = pipeline("text-generation", model="tiiuae/falcon-7b-instruct", token=HF_API_TOKEN)

def extract_files_from_folder(folder_path):
    """Scans a folder for PDF, TXT, CSV, DOCX, and IPYNB files."""
    extracted_files = {"pdf": [], "txt": [], "csv": [], "docx": [], "ipynb": []}
    
    for root, _, files in os.walk(folder_path):
        for file_name in files:
            file_path = os.path.join(root, file_name)
            if file_name.endswith(".pdf"):
                extracted_files["pdf"].append(file_path)
            elif file_name.endswith(".txt"):
                extracted_files["txt"].append(file_path)
            elif file_name.endswith(".csv"):
                extracted_files["csv"].append(file_path)
            elif file_name.endswith(".docx"):
                extracted_files["docx"].append(file_path)
            elif file_name.endswith(".ipynb"):
                extracted_files["ipynb"].append(file_path)
    return extracted_files

def get_text_from_pdf(pdf_files):
    text = ""
    for pdf_path in pdf_files:
        with open(pdf_path, "rb") as pdf_file:
            reader = PyPDF2.PdfReader(pdf_file)
            for page in reader.pages:
                text += page.extract_text() + "\n"
    return text

def read_text_from_files(file_paths):
    text = ""
    for file_path in file_paths:
        with open(file_path, "r", encoding="utf-8", errors="ignore") as file:
            text += file.read() + "\n"
    return text

def get_text_from_csv(csv_files):
    text = ""
    for csv_path in csv_files:
        df = pd.read_csv(csv_path)
        text += df.to_string() + "\n"
    return text

def get_text_from_docx(docx_files):
    text = ""
    for docx_path in docx_files:
        doc = Document(docx_path)
        for para in doc.paragraphs:
            text += para.text + "\n"
    return text

def get_text_from_ipynb(ipynb_files):
    text = ""
    for ipynb_path in ipynb_files:
        with open(ipynb_path, "r", encoding="utf-8", errors="ignore") as file:
            content = json.load(file)
            for cell in content.get("cells", []):
                if cell.get("cell_type") in ["markdown", "code"]:
                    text += "\n".join(cell.get("source", [])) + "\n"
    return text

def combine_text_from_files(extracted_files):
    text = (
        get_text_from_pdf(extracted_files["pdf"]) +
        read_text_from_files(extracted_files["txt"]) +
        get_text_from_csv(extracted_files["csv"]) +
        get_text_from_docx(extracted_files["docx"]) +
        get_text_from_ipynb(extracted_files["ipynb"])
    )
    return text

def generate_response(question, text):
    """Uses the Mistral 7B model to answer questions based on extracted text."""
    prompt = f"Question: {question}\nBased on the following document content:\n{text[:3000]}"  # Limite de 3000 caracteres
    response = chatbot_pipeline(prompt, max_length=500, truncation=True)[0]['generated_text']
    return response.strip()

def chatbot_interface(question):
    folder_path = "New_Data_Analytics/"
    extracted_files = extract_files_from_folder(folder_path)
    text = combine_text_from_files(extracted_files)
    
    if not text.strip():
        return "No valid files found. Please upload supported file types."
    
    return generate_response(question, text)

demo = gr.Interface(
    fn=chatbot_interface,
    inputs=gr.Textbox(label="Ask a question", placeholder="Type your question here..."),
    outputs=gr.Textbox(label="Answer")
)

demo.launch()