Update app.py

app.py

@@ -1,13 +1,16 @@
 import sys
 import subprocess
 import pkg_resources
+import os
+import tempfile
 
 required_packages = {
     'torch': 'torch',
     'gradio': 'gradio',
     'transformers': 'transformers',
     'decord': 'decord',
-    'numpy': 'numpy'
+    'numpy': 'numpy',
+    'instaloader': 'instaloader'
 }
 
 def install_packages(packages):
@@ -34,8 +37,33 @@ import torch
 from transformers import AutoTokenizer, AutoModelForCausalLM
 from decord import VideoReader, cpu
 import numpy as np
+import instaloader
+
+# Initialize Instaloader
+L = instaloader.Instaloader()
+
+def download_instagram_video(url):
+    try:
+        # Extract post shortcode from URL
+        shortcode = url.split("/")[-2]
+        
+        # Download the post
+        post = instaloader.Post.from_shortcode(L.context, shortcode)
+        
+        # Create a temporary directory
+        with tempfile.TemporaryDirectory() as tmpdirname:
+            # Download the video
+            L.download_post(post, target=tmpdirname)
+            
+            # Find the downloaded video file
+            video_file = next(file for file in os.listdir(tmpdirname) if file.endswith('.mp4'))
+            video_path = os.path.join(tmpdirname, video_file)
+            
+            return video_path
+    except Exception as e:
+        print(f"Error downloading video: {e}")
+        return None
 
-# Define a simple video processing function (placeholder for LLaVA-Video)
 def process_video(video_path, max_frames=64):
     vr = VideoReader(video_path, ctx=cpu(0))
     total_frames = len(vr)
@@ -43,34 +71,38 @@ def process_video(video_path, max_frames=64):
     frames = vr.get_batch(frame_indices).asnumpy()
     return frames
 
-# Define a simple text generation function (placeholder for actual model)
-def generate_response(video_frames, question):
-    # This is a placeholder. In reality, you'd use the LLaVA-Video model here.
+# This is a placeholder for the actual LLaVA-Video model
+def analyze_video(video_frames, question):
+    # In a real implementation, you would use the LLaVA-Video model here
     return f"Analyzed {len(video_frames)} frames. Your question was: {question}"
 
-def analyze_instagram_short(video_file, question):
-    if video_file is None:
-        return "Please upload an Instagram short video."
+def analyze_instagram_video(video_url, question):
+    if not video_url:
+        return "Please enter an Instagram video URL."
+    
+    video_path = download_instagram_video(video_url)
+    if not video_path:
+        return "Failed to download the video. Please check the URL and try again."
     
-    video_frames = process_video(video_file)
-    response = generate_response(video_frames, question)
+    video_frames = process_video(video_path)
+    response = analyze_video(video_frames, question)
     return response
 
 # Create Gradio interface
 with gr.Blocks() as demo:
-    gr.Markdown("# 🎥 Instagram Short Video Analyzer")
-    gr.Markdown("Upload your Instagram short video and ask questions about its content!")
+    gr.Markdown("# 🎥 Instagram Video Analyzer")
+    gr.Markdown("Enter the URL of an Instagram video and ask questions about its content!")
     
     with gr.Row():
         with gr.Column():
-            video_input = gr.Video(label="Upload Instagram Short Video")
-            question_input = gr.Textbox(label="Ask a question about the video", placeholder="What's happening in this Instagram short?")
-            submit_button = gr.Button("Analyze Short Video")
+            video_url_input = gr.Textbox(label="Instagram Video URL", placeholder="https://www.instagram.com/p/...")
+            question_input = gr.Textbox(label="Ask a question about the video", placeholder="What's happening in this Instagram video?")
+            submit_button = gr.Button("Analyze Video")
         output = gr.Textbox(label="Analysis Result")
     
     submit_button.click(
-        fn=analyze_instagram_short,
-        inputs=[video_input, question_input],
+        fn=analyze_instagram_video,
+        inputs=[video_url_input, question_input],
         outputs=output
     )