backup

app.py

@@ -4,10 +4,52 @@ from PIL import Image
 import io
 import time
 import os
+from datetime import datetime, timedelta
 
 
 access_token = os.environ.get("HUGGINGFACE_TOKEN")
 
+# Global variables
+dataset = None
+dataset_size = "Unknown"
+last_refresh_time = None
+REFRESH_INTERVAL = timedelta(hours=24)
+
+
+def load_and_prepare_dataset():
+    global dataset, dataset_size, last_refresh_time
+
+    dataset = load_dataset(
+        "taesiri/PhotoshopRequest-DailyDump",
+        split="train",
+        streaming=True,
+        token=access_token,
+    )
+
+    # Get dataset info
+    dataset_info = dataset.info
+    dataset_size = (
+        dataset_info.splits["train"].num_examples
+        if dataset_info.splits.get("train")
+        else "Unknown"
+    )
+
+    last_refresh_time = datetime.now()
+
+
+def check_and_refresh_dataset():
+    global last_refresh_time
+    current_time = datetime.now()
+    if (
+        last_refresh_time is None
+        or (current_time - last_refresh_time) >= REFRESH_INTERVAL
+    ):
+        load_and_prepare_dataset()
+
+
+# Initial dataset load
+load_and_prepare_dataset()
+
 # Load and prepare the dataset
 dataset = load_dataset(
     "taesiri/PhotoshopRequest-DailyDump",
@@ -16,6 +58,14 @@ dataset = load_dataset(
     token=access_token,
 )
 
+# Get dataset info
+dataset_info = dataset.info
+dataset_size = (
+    dataset_info.splits["train"].num_examples
+    if dataset_info.splits.get("train")
+    else "Unknown"
+)
+
 BUFFER_SIZE = 1
 sample_iterator = None
 sample_count = 0
@@ -33,6 +83,7 @@ reshuffle_dataset()  # Initial shuffle
 
 
 def get_next_sample():
+    check_and_refresh_dataset()
     global sample_count
 
     if sample_count >= BUFFER_SIZE:
@@ -60,7 +111,6 @@ with gr.Blocks() as demo:
         """
     This is a preview of the PhotoshopRequest dataset. Each sample represents a Photoshop editing request post.
     Click the 'Sample New Item' button to retrieve a random sample from the dataset.
-    </hr>
     """
     )
 
@@ -72,9 +122,20 @@ with gr.Blocks() as demo:
 
     sample_button = gr.Button("Sample New Item")
 
+    info_md = gr.Markdown()
+
+    def update_info():
+        return f"""
+    <div style="text-align: center;">
+    <hr>
+    Dataset Size: {dataset_size} items<br>
+    Last Refreshed: {last_refresh_time.strftime('%Y-%m-%d %H:%M:%S UTC') if last_refresh_time else 'Unknown'}
+    </div>
+    """
+
     sample_button.click(
         get_next_sample, outputs=[post_info, source_image, edited_image]
-    )
+    ).then(update_info, outputs=[info_md])
 
 if __name__ == "__main__":
     demo.launch()