update flexibility

app.py

@@ -18,11 +18,12 @@ def read_data(filepath: str) -> Optional[pd.DataFrame]:
     return df
 
 
-def process(task_name: str,
-            model_name: str,
-            text: str,
-            file=None,
-            ):
+def process(
+        task_name: str,
+        model_name: str,
+        text: str,
+        file=None,
+) -> (None, pd.DataFrame, str):
     # try:
     # load file
     if file:
@@ -56,36 +57,21 @@ task_name_dropdown = gr.components.Dropdown(
     value='Originality',
     choices=['Originality', 'Flexibility']
 )
-
 model_name_dropdown = gr.components.Dropdown(
     label='Model Name',
     value=list_models[0],
     choices=list_models
 )
-
 text_input = gr.components.Textbox(
     value=open('data/example_xlm.csv', 'r').read(),
     lines=10,
-    type='text'
 )
+file_input = gr.components.File(label='Input File', file_types=['.csv', '.xlsx'])
 
 # output
-file_input = gr.components.File(label='Input File',
-                                file_count='single',
-                                file_types=['', '.', '.csv', '.xls', '.xlsx'])
-
-text_output = gr.components.Textbox(
-    label='Output',
-    type='text'
-)
-
-dataframe_output = gr.components.Dataframe(
-    label='DataFrame'
-)
-
-file_output = gr.components.File(label='Output File',
-                                 file_count='single',
-                                 file_types=['', '.', '.csv', '.xls', '.xlsx'])
+text_output = gr.components.Textbox(label='Output')
+dataframe_output = gr.components.Dataframe(label='DataFrame')
+file_output = gr.components.File(label='Output File', file_types=['.csv', '.xlsx'])
 
 app = gr.Interface(
     fn=process,

utils/pipeline.py

@@ -41,15 +41,10 @@ def p1_flexibility(df: pd.DataFrame, model_name: str) -> pd.DataFrame:
 
     def get_flexibility(responses: List[str]) -> float:
         responses_vec = [model(_) for _ in responses]
-        count = 0
         score = 0
-        for i in range(len(responses_vec)):
-            for j in range(len(responses_vec)):
-                if i == j:
-                    continue
-                score += 1 - cos_sim(responses_vec[i], responses_vec[j]).item()
-                count += 1
-        return score / count
+        for i in range(len(responses_vec) - 1):
+            score += 1 - cos_sim(responses_vec[i], responses_vec[i + 1]).item()
+        return score
 
     df_out = df.groupby(by=['id', 'prompt']) \
         .agg({'id': 'first', 'prompt': 'first', 'response': get_flexibility}) \