app.py

import streamlit as st
import re
import requests
from youtube_transcript_api import YouTubeTranscriptApi
from phi.agent import Agent
from phi.model.google import Gemini
from phi.tools.duckduckgo import DuckDuckGo
from google.generativeai import upload_file, get_file
import google.generativeai as genai
import time
from dotenv import load_dotenv
import os

load_dotenv()

API_KEY = os.getenv("GOOGLE_API_KEY")
if API_KEY:
    genai.configure(api_key=API_KEY)

# Page configuration
st.set_page_config(
    page_title="Video Summarizer with Transcript Extraction",
    page_icon="🎥",
    layout="wide"
)

st.title("Video AI Summarizer with Transcript 🎥🎤🖬")
st.header("Powered by Gemini 2.0 Flash Exp")

# Initialize the agent
@st.cache_resource
def initialize_agent():
    return Agent(
        name="Video AI Summarizer",
        model=Gemini(id="gemini-2.0-flash-exp"),
        tools=[DuckDuckGo()],
        markdown=True,
    )

multimodal_Agent = initialize_agent()

# YouTube URL input
video_url = st.text_input(
    "Enter a YouTube video URL",
    help="Paste a YouTube video link to extract the transcript and generate a summary"
)

# Function to extract YouTube video ID from URL
def extract_video_id(url):
    match = re.search(r"v=([a-zA-Z0-9_-]+)", url)
    if match:
        return match.group(1)
    else:
        raise ValueError("Invalid YouTube URL")

# Function to get the transcript using the YouTubeTranscriptApi
def get_transcript(video_id):
    transcript_raw = YouTubeTranscriptApi.get_transcript(video_id, languages=['en', 'es', 'ko'])
    transcript_full = ' '.join([i['text'] for i in transcript_raw])
    return transcript_full

# Process the YouTube URL
if video_url:
    try:
        # Extract video ID from the URL
        video_id = extract_video_id(video_url)

        # Get the transcript for the video
        transcript = get_transcript(video_id)

        # Display the transcript
        st.subheader("Video Transcript")
        st.text_area("Transcript", transcript, height=200)

        user_query = st.text_area(
            "What insights or summary would you like from the transcript?",
            placeholder="Ask anything about the transcript or request a summary.",
            help="Provide specific questions or insights you want from the transcript."
        )

        if st.button("🔍 Generate Summary", key="generate_summary_button"):
            if not user_query:
                st.warning("Please enter a query or request a summary of the transcript.")
            else:
                try:
                    with st.spinner("Processing transcript and generating summary..."):
                        # Prompt generation for analysis
                        analysis_prompt = (
                            f"""
                            Here is the transcript of the video:
                            {transcript}

                            The user has requested the following insight/summary:
                            {user_query}

                            Please provide a detailed, user-friendly, and actionable summary based on the provided transcript.
                            """
                        )

                        # AI agent processing
                        response = multimodal_Agent.run(analysis_prompt)

                    # Display the result
                    st.subheader("Summary Result")
                    st.markdown(response.content)

                except Exception as error:
                    st.error(f"An error occurred during analysis: {error}")
    except Exception as e:
        st.error(f"Error processing the YouTube video: {e}")
else:
    st.info("Paste a YouTube video URL to begin transcript extraction and analysis.")