Update app.py

app.py

@@ -2,25 +2,28 @@ import streamlit as st
 from phi.agent import Agent
 from phi.model.google import Gemini
 from phi.tools.duckduckgo import DuckDuckGo
-from phi.tools.youtube_tools import YouTubeTools  # Import YouTubeTools for YouTube summarization
-from google.generativeai import upload_file, get_file
+from phi.tools.youtube_tools import YouTubeTools
 import google.generativeai as genai
-
 import time
 from pathlib import Path
 import tempfile
 from dotenv import load_dotenv
-load_dotenv()
-
 import os
 
+# Load environment variables
+load_dotenv()
+
+# Configure Gemini API
 API_KEY = os.getenv("GOOGLE_API_KEY")
-if API_KEY:
-    genai.configure(api_key=API_KEY)
+if not API_KEY:
+    st.error("Please set the GOOGLE_API_KEY environment variable")
+    st.stop()
+
+genai.configure(api_key=API_KEY)
 
 # Page configuration
 st.set_page_config(
-    page_title="Multimodal AI Agent- Video Summarizer",
+    page_title="Multimodal AI Agent - Video Summarizer",
     page_icon="🎥",
     layout="wide"
 )
@@ -34,7 +37,7 @@ def initialize_agent():
     return Agent(
         name="Video AI Summarizer",
         model=Gemini(id="gemini-2.0-flash-exp"),
-        tools=[DuckDuckGo(), YouTubeTools()],  # Add YouTubeTools here for YouTube video summarization
+        tools=[DuckDuckGo(), YouTubeTools()],
         markdown=True,
     )
 
@@ -47,91 +50,115 @@ user_query = st.text_area(
     help="Provide specific questions or insights you want from the video."
 )
 
-# Check for either video file upload or YouTube URL
-video_file = st.file_uploader(
-    "Upload a video file", type=['mp4', 'mov', 'avi'], help="Upload a video for AI analysis"
-)
+# Create tabs for different input methods
+video_tab, youtube_tab = st.tabs(["Upload Video", "YouTube URL"])
 
-youtube_url = st.text_input("Or, paste a YouTube URL to summarize:")
+with video_tab:
+    video_file = st.file_uploader(
+        "Upload a video file", 
+        type=['mp4', 'mov', 'avi'], 
+        help="Upload a video for AI analysis"
+    )
 
-# Unified analysis prompt
-def generate_analysis_prompt(user_query, video_source):
-    return (
-        f"""
-        Analyze the {video_source} for content and context.
-        Respond to the following query using video insights and supplementary web research:
-        {user_query}
-        Provide a detailed, user-friendly, and actionable response.
-        """
+with youtube_tab:
+    youtube_url = st.text_input(
+        "Paste a YouTube URL:",
+        placeholder="https://www.youtube.com/watch?v=..."
     )
 
+def generate_analysis_prompt(user_query, video_source):
+    return f"""
+    Analyze the {video_source} for content and context.
+    Key areas to address:
+    1. Main topics and themes
+    2. Key points and insights
+    3. Specific answer to user query: {user_query}
+    4. Supporting context from web research
+    
+    Provide a structured, detailed response with clear sections and actionable insights.
+    """
+
 # Process uploaded video
 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)
-
-    if st.button("🔍 Analyze Uploaded Video", key="analyze_uploaded_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 gathering insights..."):
-                    # 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)
-
-                    # Use the unified prompt for analysis (Uploaded video)
-                    analysis_prompt = generate_analysis_prompt(user_query, "uploaded video")
-
-                    # AI agent processing
-                    response = multimodal_Agent.run(analysis_prompt, videos=[processed_video])
-
-                # Display the result
-                st.subheader("Analysis Result")
-                st.markdown(response.content)
-
-            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)
+    with video_tab:
+        st.video(video_file, format="video/mp4", start_time=0)
+        
+        if st.button("🔍 Analyze Uploaded Video", use_container_width=True):
+            if not user_query:
+                st.warning("Please enter a question or insight to analyze the video.")
+            else:
+                try:
+                    with st.spinner("Processing video and gathering insights..."):
+                        # Create temporary file
+                        with tempfile.NamedTemporaryFile(delete=False, suffix='.mp4') as temp_video:
+                            temp_video.write(video_file.getvalue())
+                            video_path = temp_video.name
+
+                        # Upload and process video
+                        processed_video = genai.upload_file(video_path)
+                        
+                        # Wait for processing with timeout
+                        timeout = 60  # 60 seconds timeout
+                        start_time = time.time()
+                        while processed_video.state.name == "PROCESSING":
+                            if time.time() - start_time > timeout:
+                                raise TimeoutError("Video processing timed out")
+                            time.sleep(1)
+                            processed_video = genai.get_file(processed_video.name)
+
+                        # Generate and run analysis
+                        analysis_prompt = generate_analysis_prompt(user_query, "uploaded video")
+                        response = multimodal_Agent.run(analysis_prompt, videos=[processed_video])
+
+                        # Display results
+                        st.success("Analysis complete!")
+                        st.subheader("Analysis Results")
+                        st.markdown(response.content)
+
+                except TimeoutError:
+                    st.error("Video processing timed out. Please try with a shorter video.")
+                except Exception as error:
+                    st.error(f"An error occurred: {str(error)}")
+                finally:
+                    # Clean up temporary file
+                    Path(video_path).unlink(missing_ok=True)
 
 # Process YouTube URL
-elif youtube_url:
-    if st.button("🔍 Summarize YouTube Video", key="summarize_youtube_button"):
-        if not user_query:
-            st.warning("Please enter a question or insight to analyze the YouTube video.")
-        else:
-            try:
-                with st.spinner("Fetching video captions and summarizing..."):
-                    # Use the unified prompt for summarization (YouTube URL)
-                    analysis_prompt = generate_analysis_prompt(user_query, "YouTube URL")
-
-                    # Summarize the YouTube video using the agent
-                    response = multimodal_Agent.run(analysis_prompt, youtube_url")
-                    
-                    # Display the result
-                    st.subheader("Video Summary")
-                    st.markdown(response["content"], unsafe_allow_html=True)
-
-            except Exception as error:
-                st.error(f"An error occurred: {error}")
-
-else:
-    st.info("Upload a video file or paste a YouTube URL to begin analysis.")
-
-# Customize text area height
+if youtube_url:
+    with youtube_tab:
+        if st.button("🔍 Analyze YouTube Video", use_container_width=True):
+            if not user_query:
+                st.warning("Please enter a question or insight to analyze the video.")
+            else:
+                try:
+                    with st.spinner("Fetching video content and analyzing..."):
+                        # Generate and run analysis
+                        analysis_prompt = generate_analysis_prompt(user_query, "YouTube video")
+                        response = multimodal_Agent.run(analysis_prompt, youtube_url=youtube_url)
+                        
+                        # Display results
+                        st.success("Analysis complete!")
+                        st.subheader("Analysis Results")
+                        st.markdown(response.content)
+                        
+                except Exception as error:
+                    st.error(f"An error occurred: {str(error)}")
+
+# Custom styling
 st.markdown(
     """
     <style>
     .stTextArea textarea {
         height: 100px;
     }
+    .stButton button {
+        background-color: #FF4B4B;
+        color: white;
+        font-weight: bold;
+    }
+    .stButton button:hover {
+        background-color: #FF3333;
+    }
     </style>
     """,
     unsafe_allow_html=True