fix log functionality

app.py

@@ -17,13 +17,7 @@ API_KEY = os.getenv("API_KEY")
 
 df = pd.read_csv('subsectors.csv')
 logs_columns = ['Abstract', 'Model', 'Results']
-
-
-def update_logs(new_log_entry=None):
-    logs_df = pd.read_csv('logs.csv')
-    logs_df = pd.concat([logs_df, new_log_entry], ignore_index=True)
-    logs_df.to_csv('logs.csv', columns=logs_columns)
-    return logs_df
+logs_df = PandasDataFrame(columns=logs_columns)
 
 
 def build_context(row):
@@ -86,18 +80,19 @@ def click_button(model, api_key, abstract):
     response_reasoning = language_model.get_last_message()
 
     dict_pattern = r'\{.*?\}'
-    probabilities_match = re.search(dict_pattern, response_reasoning, re.DOTALL)
+    match = re.search(dict_pattern, response_reasoning, re.DOTALL)
 
-    if probabilities_match and language_model.error is False:
-        probabilities_dict = eval(probabilities_match.group(0))
+    if match and language_model.error is False:
+        match_score_dict = eval(match.group(0))
     else:
-        probabilities_dict = {}
+        match_score_dict = {}
 
-    # Save classification results to logs.csv
-    new_log_entry = pd.DataFrame({'Abstract': [abstract], 'Model': [model], 'Results': [str(probabilities_dict)]})
-    update_logs(new_log_entry=new_log_entry)
+    # Update Logs
+    new_log_entry = pd.DataFrame({'Abstract': [abstract], 'Model': [model], 'Results': [str(match_score_dict)]})
+    global logs_df
+    logs_df = pd.concat([logs_df, new_log_entry], ignore_index=True)
 
-    return probabilities_dict, response_reasoning
+    return match_score_dict, response_reasoning, logs_df
 
 def on_select(evt: gr.SelectData):  # SelectData is a subclass of EventData
     selected = df.iloc[[evt.index[0]]].iloc[0]
@@ -165,13 +160,23 @@ with gr.Blocks(css=css, js=js) as demo:
                     does_include = gr.Textbox(label="Does include", lines=4)
                     does_not_include = gr.Textbox(label="Does not include", lines=3)
     with gr.Tab("Logs"):
-        def load_logs():
-            logs = pd.read_csv('logs.csv')
-            return logs
-        gr.Dataframe(load_logs(), height=1000)
-
-    btn_get_result.click(fn=click_button, inputs=[dropdown_model, api_key, abstract_description], outputs=[label_result, reasoning])
-    df_subsectors.select(fn=on_select, outputs=[subsector_name, s1_definition, s1_keywords, does_include, does_not_include])
+        output_dataframe = gr.Dataframe(
+            value=logs_df,
+            type="pandas",
+            height=500,
+            headers=logs_columns,
+            interactive=False,
+            column_widths=["45%", "10%", "45%"],
+        )
+
+    btn_get_result.click(
+        fn=click_button,
+        inputs=[dropdown_model, api_key, abstract_description],
+        outputs=[label_result, reasoning, output_dataframe])
+    df_subsectors.select(
+        fn=on_select,
+        outputs=[subsector_name, s1_definition, s1_keywords, does_include, does_not_include]
+    )
 
 if __name__ == "__main__":
     # demo.queue()