attemts to add simpler example loading and image processing

app.py

@@ -3,8 +3,6 @@ import torch
 from PIL import Image
 import json
 import os
-import shutil
-import tempfile
 from transformers import AutoProcessor, AutoModelForImageTextToText
 from typing import List, Dict, Any
 import logging
@@ -148,6 +146,9 @@ class LOperatorDemo:
             elif isinstance(image, str) and os.path.exists(image):
                 # Handle file path (from examples)
                 pil_image = Image.open(image)
+            elif hasattr(image, 'name') and os.path.exists(image.name):
+                # Handle Gradio file object
+                pil_image = Image.open(image.name)
             else:
                 return history + [{"role": "user", "content": message}, {"role": "assistant", "content": "❌ Invalid image format. Please upload a valid image."}]
             
@@ -183,14 +184,10 @@ def load_model():
 
 
 def load_example_episodes():
-    """Load example episodes using shutil to copy files to temp location"""
+    """Load example episodes using PIL to load images directly"""
     examples = []
 
     try:
-        # Create temporary directory for examples
-        temp_dir = tempfile.mkdtemp()
-        logger.info(f"Created temporary directory for examples: {temp_dir}")
-        
         episode_dirs = ["episode_13", "episode_53", "episode_73"]
 
         for episode_dir in episode_dirs:
@@ -205,16 +202,15 @@ def load_example_episodes():
                     with open(metadata_path, "r") as f:
                         metadata = json.load(f)
                     
-                    # Copy image to temp directory
-                    temp_image_path = os.path.join(temp_dir, f"{episode_dir}_screenshot.png")
-                    shutil.copy2(image_path, temp_image_path)
+                    # Load image directly with PIL
+                    pil_image = Image.open(image_path)
                     
                     episode_num = episode_dir.split('_')[1]
                     goal_text = metadata.get('goal', f'Episode {episode_num} example')
                     
                     examples.append([
-                        temp_image_path,  # Use temp file path
-                        goal_text  # Just the goal text, no additional formatting
+                        pil_image,  # Use PIL Image object directly
+                        goal_text  # Use the goal text from metadata
                     ])
                     logger.info(f"Successfully loaded example for Episode {episode_num}")
                     
@@ -226,7 +222,7 @@ def load_example_episodes():
         logger.error(f"Error loading examples: {str(e)}")
         examples = []
 
-    logger.info(f"Loaded {len(examples)} examples using shutil")
+    logger.info(f"Loaded {len(examples)} examples using PIL")
     return examples
 
 # Create Gradio interface
@@ -306,7 +302,8 @@ def create_demo():
                     textbox=gr.Textbox(
                         label="Goal",
                         placeholder="e.g., Open the Settings app and navigate to Display settings",
-                        lines=2
+                        lines=2,
+                        show_label=True
                     )
                 )