Update app.py

app.py

@@ -17,6 +17,16 @@ async def get_splits(dataset_name: str) -> Dict[str, List[Dict]]:
             ]
         }
 
+async def get_valid_datasets() -> List[str]:
+    URL = f"https://huggingface.co/api/datasets"
+    async with httpx.AsyncClient() as session:
+        response = await session.get(URL)
+        try:
+            datasets = [dataset["id"] for dataset in response.json() if "/" in dataset["id"]]
+        except (KeyError, json.JSONDecodeError):
+            datasets = []  # Set a default value if the response is not in the expected format
+        return datasets
+
 async def get_first_rows(dataset: str, config: str, split: str) -> Dict[str, Dict[str, List[Dict]]]:
     URL = f"https://huggingface.co/datasets/{dataset}/resolve/main/dataset_info.json"
     async with httpx.AsyncClient() as session:
@@ -30,25 +40,6 @@ async def get_first_rows(dataset: str, config: str, split: str) -> Dict[str, Dic
         }
         return first_rows
 
-# Guido von Roissum: https://www.youtube.com/watch?v=-DVyjdw4t9I
-async def update_URL(dataset: str, config: str, split: str) -> str:
-    URL = f"https://huggingface.co/datasets/{dataset}/tree/main/{config}/{split}"
-    return URL
-
-    
-
-
-async def get_valid_datasets() -> List[str]:
-    URL = f"https://huggingface.co/api/datasets"
-    async with httpx.AsyncClient() as session:
-        response = await session.get(URL)
-        try:
-            datasets = [dataset["id"] for dataset in response.json()]
-        except (KeyError, json.JSONDecodeError):
-            datasets = []  # Set a default value if the response is not in the expected format
-        return datasets
-
-
 def get_df_from_rows(api_output):
     dfFromSort = pd.DataFrame([row["row"] for row in api_output["rows"]])
     try:
@@ -58,26 +49,32 @@ def get_df_from_rows(api_output):
     return dfFromSort
 
 async def update_configs(dataset_name: str):
-    splits = await get_splits(dataset_name)
-    all_configs = sorted(set([s["config"] for s in splits["splits"]]))
-    return (gr.Dropdown.update(choices=all_configs, value=all_configs[0]),
-            splits)
+    try:
+        splits = await get_splits(dataset_name)
+        all_configs = sorted(set([s["config"] for s in splits["splits"]]))
+        return (gr.Dropdown.update(choices=all_configs, value=all_configs[0] if all_configs else None),
+                splits)
+    except json.JSONDecodeError:
+        return (gr.Dropdown.update(choices=[], value=None), {"splits": []})
 
 async def update_splits(config_name: str, state: gr.State):
     splits_for_config = sorted(set([s["split"] for s in state["splits"] if s["config"] == config_name]))
-    dataset_name = state["splits"][0]["dataset"]
-    dataset = await update_dataset(splits_for_config[0], config_name, dataset_name)
-    return (gr.Dropdown.update(choices=splits_for_config, value=splits_for_config[0]), dataset)
+    dataset_name = state["splits"][0]["dataset"] if state["splits"] else None
+    dataset = await update_dataset(splits_for_config[0], config_name, dataset_name) if splits_for_config and dataset_name else gr.DataFrame.update()
+    return (gr.Dropdown.update(choices=splits_for_config, value=splits_for_config[0] if splits_for_config else None), dataset)
 
 async def update_dataset(split_name: str, config_name: str, dataset_name: str):
     rows = await get_first_rows(dataset_name, config_name, split_name)
     df = get_df_from_rows(rows)
     return df
 
+async def update_URL(dataset: str, config: str, split: str) -> str:
+    URL = f"https://huggingface.co/datasets/{dataset}/tree/main/{config}/{split}"
+    return URL
    
 async def openurl(URL: str) -> str:
     html = f"<a href={URL} target=_blank>{URL}</a>"
-    return (html)
+    return html
 
 with gr.Blocks() as demo:
     gr.Markdown("<h1><center>🥫Datasetter📊 Datasets Analyzer and Transformer</center></h1>")
@@ -86,16 +83,13 @@ with gr.Blocks() as demo:
     splits_data = gr.State()
     
     with gr.Row():
-        dataset_name = gr.Dropdown(label="Dataset", interactive=True)
+        dataset_name = gr.Dropdown(label="Dataset", interactive=True, allow_custom_value=True)
         config = gr.Dropdown(label="Subset", interactive=True)
         split = gr.Dropdown(label="Split", interactive=True)
     
     with gr.Row():
-        #filterleft = gr.Textbox(label="First Column Filter",placeholder="Filter Column 1")
         URLcenter = gr.Textbox(label="Dataset URL", placeholder="URL")
         btn = gr.Button("Use Dataset")
-        #URLoutput = gr.Textbox(label="Output",placeholder="URL Output")
-        #URLoutput = gr.HTML(label="Output",placeholder="URL Output")
         URLoutput = gr.HTML(label="Output")
 
     with gr.Row():