Merge branch #davanstrien/collection_dataset_viewer' into 'librarian-bots/Alpaca-Style-Datasets-explorer'

app.py

@@ -52,13 +52,13 @@ def navigate_dataset(dataset_ids, index, direction):
 
 
 def get_display_name(collection_id):
-    # Strip out the extra ID part of the name
-    parts = collection_id.split("/")
-    if len(parts) == 2:
-        owner, name = parts
-        name = re.sub(r"-[a-f0-9]{32}$", "", name)
-        return f"{owner}/{name}"
-    return collection_id  # Return original if not in expected format
+    # Pattern to match username/repo-name with an optional ID of 16 or more hexadecimal characters
+    pattern = r"^(.+?)-([a-f0-9]{16,})$"
+    if match := re.match(pattern, collection_id):
+        return match[1]
+    else:
+        # If no match, return the original
+        return collection_id
 
 
 try:
@@ -73,14 +73,8 @@ try:
 
         gr.Markdown("""
         This app allows you to browse and view datasets from a specific Hugging Face collection. 
-        Use the 'Previous' and 'Next' buttons to navigate through the datasets in the collection.
-
-        **Note**: This space is currently set up to display datasets from a specific collection. 
-        If you'd like to use it for a different collection:
-        1. Duplicate this space
-        2. In your duplicated space, set the `COLLECTION_SLUG_OR_URL` environment variable to your desired collection ID or URL
-        3. Your new space will then display datasets from your chosen collection!
-        """)
+        Use the 'Previous' and 'Next' buttons to navigate through the datasets in the collection. 
+        See below for how to set up this app for a different collection.""")
 
         index_state = gr.State(value=0)
 
@@ -90,7 +84,12 @@ try:
 
         dataset_info = gr.Markdown(f"Dataset 1 of {len(dataset_ids)}: {dataset_ids[0]}")
         iframe_output = gr.HTML()
-
+        gr.Markdown("""**Note**: This space is currently set up to display datasets from a specific collection. 
+        If you'd like to use it for a different collection:
+        1. Duplicate this space
+        2. In your duplicated space, set the `COLLECTION_SLUG_OR_URL` environment variable to your desired collection ID or URL
+        3. Your new space will then display datasets from your chosen collection!
+        """)
         left_btn.click(
             navigate_dataset,
             inputs=[gr.State(dataset_ids), index_state, gr.Number(-1, visible=False)],