updated compute fn and test cases

app.py

@@ -7,30 +7,40 @@ from evaluate.utils import infer_gradio_input_types, json_to_string_type, parse_
 from fixed_f1 import FixedF1
 from pathlib import Path
 
-def compute(input: pd.DataFrame):
-
-    metric._compute()
-
 metric = FixedF1()
 
 if isinstance(metric.features, list):
     (feature_names, feature_types) = zip(*metric.features[0].items())
 else:
     (feature_names, feature_types) = zip(*metric.features.items())
-
 gradio_input_types = infer_gradio_input_types(feature_types)
 
 local_path = Path(sys.path[0])
 test_cases = [ {"predictions":[1,2,3,4,5], "references":[1,2,5,4,3]} ] # configure this randomly using randint generator and feature names?
 
+# configure this based on the input type, etc. for launch_gradio_widget
+def compute(input_df: pd.DataFrame, feature_names: tuple[str]):
+
+    predicted = [int(num) for num in input_df[feature_names[0]].to_list()]
+    references = [int(num) for num in input_df[feature_names[1]].to_list()]
+
+    metric.add_batch(predictions=predicted, references=references)
+    
+    outputs = metric._compute()
+
+    f"Your metrics are as follows: \n {outputs}"
+
 space = gr.Interface(
     fn=compute,
-    inputs=gr.Dataframe(
+    inputs=[
+        gr.Dataframe(
         headers=feature_names,
         col_count=len(feature_names),
         row_count=5,
         datatype=json_to_string_type(gradio_input_types),
         ),
+        feature_names
+    ],
     outputs=gr.Textbox(label=metric.name),
     description=(
         metric.info.description + "\nIf this is a text-based metric, make sure to wrap your input in double quotes."
@@ -39,7 +49,14 @@ space = gr.Interface(
     title=f"Metric: {metric.name}",
     article=parse_readme(local_path / "README.md"),
     # TODO: load test cases and use them to populate examples
-    examples=[parse_test_cases(test_cases, feature_names, gradio_input_types)]
+    examples=[
+        [
+            # consider how to generalize this
+            parse_test_cases(test_cases, feature_names, gradio_input_types)[0], 
+            feature_names
+        ]
+    ],
+    cache_examples=False
     )
 
 space.launch()