update

app.py

@@ -3,11 +3,14 @@ from pydrive2.drive import GoogleDrive
 import os
 import gradio as gr
 from datasets import load_dataset, Dataset, concatenate_datasets
+from huggingface_hub import create_repo, upload_file, login
+from pathlib import Path
 import pandas as pd
 from PIL import Image
 from tqdm import tqdm
 import logging
 import yaml
+import json
 
 # Set up logging
 logging.basicConfig(
@@ -53,6 +56,77 @@ def validate_input(folder_id, naming_convention):
         return False, "Naming convention should only contain letters, numbers, and underscores"
     return True, ""
 
+def initialize_dataset():
+    """Initialize or verify the dataset structure."""
+    try:
+        # Check if the README.md exists, if not create it
+        readme_content = """# Sports Cards Dataset
+
+This dataset contains sports card images with structured metadata. Each image is named using a consistent convention and includes relevant information about the card.
+
+## Dataset Structure
+
+```
+sports_card_{number}.jpg - Card images
+```
+
+## Features
+- file_path: Path to the image file
+- original_name: Original filename of the card
+- new_name: Standardized filename
+- image: Image data
+
+## Usage
+```python
+from datasets import load_dataset
+dataset = load_dataset("GotThatData/sports-cards")
+```
+
+## License
+This dataset is licensed under MIT.
+
+## Creator
+Created by GotThatData
+"""
+        # Create dataset info content
+        dataset_info = {
+            "description": "A collection of sports card images with metadata",
+            "citation": "",
+            "homepage": "https://huggingface.co/datasets/GotThatData/sports-cards",
+            "license": "mit",
+            "features": {
+                "file_path": {"dtype": "string", "_type": "Value"},
+                "original_name": {"dtype": "string", "_type": "Value"},
+                "new_name": {"dtype": "string", "_type": "Value"},
+                "image": {"dtype": "string", "_type": "Value"}
+            },
+            "splits": ["train"]
+        }
+
+        # Write files
+        with open("README.md", "w") as f:
+            f.write(readme_content)
+        with open("dataset-info.json", "w") as f:
+            json.dump(dataset_info, f, indent=2)
+
+        # Upload files to repository
+        upload_file(
+            path_or_fileobj="README.md",
+            path_in_repo="README.md",
+            repo_id="GotThatData/sports-cards",
+            repo_type="dataset"
+        )
+        upload_file(
+            path_or_fileobj="dataset-info.json",
+            path_in_repo="dataset-info.json",
+            repo_id="GotThatData/sports-cards",
+            repo_type="dataset"
+        )
+        
+        return True, "Dataset structure initialized successfully"
+    except Exception as e:
+        return False, f"Failed to initialize dataset: {str(e)}"
+
 # DatasetManager Class
 class DatasetManager:
     def __init__(self, local_images_dir="downloaded_cards"):
@@ -60,6 +134,11 @@ class DatasetManager:
         self.drive = None
         self.dataset_name = "GotThatData/sports-cards"
         os.makedirs(local_images_dir, exist_ok=True)
+        
+        # Initialize dataset structure
+        success, message = initialize_dataset()
+        if not success:
+            logger.warning(f"Dataset initialization warning: {message}")
 
     def authenticate_drive(self):
         """Authenticate with Google Drive."""
@@ -259,4 +338,27 @@ demo = gr.Interface(
 )
 
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch(
+        server_name="0.0.0.0",
+        server_port=7860,
+        # Add security headers
+        headers=[
+            ("Permissions-Policy", 
+             "accelerometer=(), "
+             "ambient-light-sensor=(), "
+             "battery=(), "
+             "camera=(), "
+             "document-domain=(), "
+             "layout-animations=(), "
+             "legacy-image-formats=(), "
+             "microphone=(), "
+             "oversized-images=(), "
+             "sync-xhr=(), "
+             "vr=(), "
+             "wake-lock=()"
+            ),
+            ("Cross-Origin-Embedder-Policy", "require-corp"),
+            ("Cross-Origin-Opener-Policy", "same-origin"),
+            ("Cross-Origin-Resource-Policy", "same-origin")
+        ]
+    )