Spaces:
Running
Running
| import streamlit as st | |
| 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 pathlib import Path | |
| import tempfile | |
| from dotenv import load_dotenv | |
| load_dotenv() | |
| import os | |
| API_KEY = os.getenv("GOOGLE_API_KEY") | |
| if API_KEY: | |
| genai.configure(api_key=API_KEY) | |
| # Page configuration | |
| st.set_page_config( | |
| page_title="Multimodal AI Agent- Video Summarizer", | |
| page_icon="🎥", | |
| layout="wide" | |
| ) | |
| st.title("Phidata Video AI Summarizer Agent 🎥🎤🖬") | |
| st.header("Powered by Gemini 2.0 Flash Exp") | |
| def initialize_agent(): | |
| return Agent( | |
| name="Video AI Summarizer", | |
| model=Gemini(id="gemini-2.0-flash-exp"), | |
| tools=[DuckDuckGo()], | |
| markdown=True, | |
| ) | |
| ## Initialize the agent | |
| multimodal_Agent = initialize_agent() | |
| # File uploader | |
| video_file = st.file_uploader( | |
| "Upload a video file", type=['mp4', 'mov', 'avi'], help="Upload a video for AI analysis" | |
| ) | |
| if video_file: | |
| with tempfile.NamedTemporaryFile(delete=False, suffix='.mp4') as temp_video: | |
| temp_video.write(video_file.read()) | |
| video_path = temp_video.name | |
| st.video(video_path, format="video/mp4", start_time=0) | |
| user_query = st.text_area( | |
| "What insights are you seeking from the video?", | |
| value="Generate detailed student-style notes from this video in paragraph format, with thorough explanations of concepts and ideas.", | |
| help="Provide specific questions or insights you want from the video." | |
| ) | |
| if st.button("🔍 Analyze Video", key="analyze_video_button"): | |
| if not user_query: | |
| st.warning("Please enter a question or insight to analyze the video.") | |
| else: | |
| try: | |
| with st.spinner("Processing video and creating detailed notes..."): | |
| # Upload and process video file | |
| processed_video = upload_file(video_path) | |
| while processed_video.state.name == "PROCESSING": | |
| time.sleep(1) | |
| processed_video = get_file(processed_video.name) | |
| # Enhanced prompt for detailed paragraph-style student notes | |
| analysis_prompt = ( | |
| f""" | |
| Analyze the uploaded video thoroughly and create EXTREMELY DETAILED STUDENT NOTES in full paragraph format. | |
| Your notes should: | |
| - Begin with a thorough introduction to the topic | |
| - Use full, detailed paragraphs (NOT bullet points) to explain concepts | |
| - Structure the content with clear headings and subheadings | |
| - Include detailed explanations and elaborations on key concepts | |
| - Connect ideas using transition sentences between paragraphs | |
| - Use complete sentences and proper grammar throughout | |
| - Create a cohesive narrative flow like detailed textbook content | |
| - Include a conclusion paragraph summarizing the main points | |
| IMPORTANT: Do NOT use bullet points or lists. Present all information in well-developed paragraphs. | |
| Additional context/question from user: {user_query} | |
| The notes should read like detailed textbook content that thoroughly explains all concepts from the video. | |
| """ | |
| ) | |
| # AI agent processing | |
| response = multimodal_Agent.run(analysis_prompt, videos=[processed_video]) | |
| # Display the result | |
| st.subheader("📝 Detailed Notes") | |
| st.markdown(response.content) | |
| # Simple download option | |
| st.download_button( | |
| label="Download Notes", | |
| data=response.content, | |
| file_name="video_notes.md", | |
| mime="text/markdown", | |
| ) | |
| except Exception as error: | |
| st.error(f"An error occurred during analysis: {error}") | |
| finally: | |
| # Clean up temporary video file | |
| Path(video_path).unlink(missing_ok=True) | |
| else: | |
| st.info("Upload a video file to begin analysis.") | |