app.py

import os
import docx
import requests
import gradio as gr
from io import BytesIO

# ✅ API Key from Hugging Face Secrets
API_KEY = os.getenv("GEMINI_API_KEY")  # Set this in Hugging Face secrets!

# ✅ Function to Download File from Link
def download_file_from_link(file_link):
    """Download a file from the provided link."""
    try:
        response = requests.get(file_link)
        response.raise_for_status()  # Raise an error for bad status codes
        return BytesIO(response.content)  # Return file content as a BytesIO object
    except Exception as e:
        return f"❌ Error downloading file: {e}"

# ✅ Function to Load and Read File Contents
def load_data(file_link):
    """Read content from the downloaded file."""
    file_content = download_file_from_link(file_link)
    if isinstance(file_content, str):  # If an error message is returned
        return file_content

    try:
        doc = docx.Document(file_content)
        return "\n".join([para.text for para in doc.paragraphs if para.text.strip()])
    except Exception as e:
        return f"❌ Error reading file: {e}"

# ✅ Function to Call Google AI Gemini API
def call_gemini_api(file_link, user_input):
    """Send user query + document content to Google AI."""
    if not API_KEY:
        return "❌ Error: API Key not found! Add it as GEMINI_API_KEY in Hugging Face secrets."

    # Read document content
    document_content = load_data(file_link)
    if document_content.startswith("❌"):
        return document_content  # Return error message if file loading failed

    # Truncate document content to avoid exceeding API limits
    max_context_length = 4000  # Adjust based on API limits
    truncated_content = document_content[:max_context_length]

    # Construct API request
    url = f"https://generativelanguage.googleapis.com/v1beta/models/gemini-1.5-flash:generateContent?key={API_KEY}"
    headers = {"Content-Type": "application/json"}
    payload = {
        "contents": [
            {
                "parts": [
                    {"text": f"Document Context: {truncated_content}\nUser Query: {user_input}"}
                ]
            }
        ]
    }

    # Make API call
    try:
        response = requests.post(url, json=payload, headers=headers)
        response.raise_for_status()  # Raise an error for bad status codes
        return response.json().get("candidates", [{}])[0].get("content", "No response from API")
    except Exception as e:
        return f"❌ API Error: {e}"

# ✅ Gradio Interface
iface = gr.Interface(
    fn=call_gemini_api,
    inputs=[
        gr.Textbox(label="File Link", placeholder="Paste the file link here (e.g., Google Drive link)"),
        gr.Textbox(label="Ask a question")
    ],
    outputs="text",
    title="📄 AI Chatbot with Real-Time Document Context",
    description="This chatbot answers questions based on the document fetched from the provided link."
)

iface.launch()