app.py

import os
from dotenv import load_dotenv
import fitz  # PyMuPDF
from docx import Document
from sentence_transformers import SentenceTransformer
import faiss
import numpy as np
import pickle
from langchain_community.llms import HuggingFaceEndpoint
from langchain_community.vectorstores import FAISS
from langchain_community.embeddings import HuggingFaceEmbeddings
import gradio as gr

# Load environment variables from .env
load_dotenv()

# Initialize the embedding model
embedding_model = SentenceTransformer('all-MiniLM-L6-v2')

# Hugging Face API token
api_token = os.getenv('HUGGINGFACEHUB_API_TOKEN')
if not api_token:
    raise ValueError("HUGGINGFACEHUB_API_TOKEN environment variable is not set")
print(f"API Token: {api_token[:5]}...")

# Initialize the HuggingFace LLM
llm = HuggingFaceEndpoint(
    endpoint_url="https://api-inference.huggingface.co/models/gpt2",
    model_kwargs={"api_key": api_token}
)

# Initialize the HuggingFace embedding
embedding = HuggingFaceEmbeddings()

# Load or create FAISS index
index_path = "faiss_index.pkl"
if os.path.exists(index_path):
    with open(index_path, "rb") as f:
        index = pickle.load(f)
else:
    # Create a new FAISS index if it doesn't exist
    index = faiss.IndexFlatL2(embedding_model.get_sentence_embedding_dimension())
    with open(index_path, "wb") as f:
        pickle.dump(index, f)


# Function to extract text from a PDF file
def extract_text_from_pdf(pdf_path):
    text = ""
    doc = fitz.open(pdf_path)
    for page_num in range(len(doc)):
        page = doc.load_page(page_num)
        text += page.get_text()
    return text


# Function to extract text from a Word document
def extract_text_from_docx(docx_path):
    doc = Document(docx_path)
    text = "\n".join([para.text for para in doc.paragraphs])
    return text


def process_and_query(text):
    # Process the text and update FAISS index (similar to the previous code)
    sentences = text.split("\n")
    embeddings = embedding_model.encode(sentences)
    index.add(np.array(embeddings))

    # Search the FAISS index
    query_embedding = embedding_model.encode([text])
    D, I = index.search(np.array(query_embedding), k=5)

    top_documents = []
    for idx in I[0]:
        if idx != -1:  # Ensure that a valid index is found
            top_documents.append(f"Document {idx}")

    # Generate response using LLM (optional)
    # You can replace this with your desired LLM interaction logic
    response = llm.run(inputs=text, max_length=100, temperature=0.7)["generated_text"]

    return {"top_documents": top_documents, "response": response}


# Define the Gradio interface
interface = gr.Interface(
    fn=process_and_query,
    inputs="textbox",
    outputs=["list", "text"],
    title="Chatbot with Text Processing and Retrieval",
    description="Upload a document (PDF or Word) or enter text to process. The chatbot will retrieve relevant documents and generate a response (optional).",
)

# Launch the Gradio interface
interface.launch()