Update app.py

app.py

@@ -389,36 +389,21 @@ def predict(
     )
     fig_in.update_layout(width=1200, height=400)
 
-    # 8) Bar chart for predicted labels (0 and 1) - vertical stacked, actual counts
-    label_counts_data = []
+    # 8) Bar chart for predicted labels
+    label_counts = {}
     for lbl_col, (pred_val, _) in label_prediction_info.items():
         if lbl_col in df.columns:
-            # Count how many in df have label_col == 0 vs label_col == 1
-            val_counts = df[lbl_col].value_counts(dropna=False)
-            # Only gather data for 0 or 1
-            for val, cnt in val_counts.items():
-                if val in [0, 1]:  # filter out anything else
-                    label_counts_data.append({
-                        "Label Column": lbl_col,
-                        "Label Value": str(val),
-                        "Count": cnt
-                    })
-
-    if label_counts_data:
-        bar_lbl_df = pd.DataFrame(label_counts_data)
+            label_counts[lbl_col] = len(df[df[lbl_col] == pred_val])
+    if label_counts:
+        bar_lbl_df = pd.DataFrame({
+            "Label": list(label_counts.keys()),
+            "Count": list(label_counts.values())
+        })
         fig_lbl = px.bar(
-            bar_lbl_df,
-            x="Label Column",
-            y="Count",
-            color="Label Value",
-            title="Stacked Bar of 0 and 1 (Exact Counts) per Predicted Label Column"
-        )
-        fig_lbl.update_layout(
-            width=1200,
-            height=400,
-            barmode="stack",  # stacked bars
-            barnorm=None      # no normalization => actual counts
+            bar_lbl_df, x="Label", y="Count",
+            title="Number of Patients with the Same Predicted Label"
         )
+        fig_lbl.update_layout(width=1200, height=400)
     else:
         fig_lbl = px.bar(title="No valid predicted labels to display.")
         fig_lbl.update_layout(width=1200, height=400)
@@ -429,7 +414,7 @@ def predict(
         total_count_md,    # 3) Total Patient Count
         nn_md,             # 4) Nearest Neighbors Summary
         fig_in,            # 5) Bar Chart (input features)
-        fig_lbl            # 6) Bar Chart (labels: vertical stacked, counts for 0/1 only)
+        fig_lbl            # 6) Bar Chart (labels)
     )
 
 ######################################
@@ -440,6 +425,7 @@ def combined_plot(feature_list, label_col):
     If user picks 1 feature => distribution plot.
     If user picks 2 features => co-occurrence plot.
     Otherwise => show error or empty plot.
+
     This function also maps numeric codes to text using 'input_mapping'
     and 'predictor.prediction_map' so that the plots display more readable labels.
     """
@@ -452,10 +438,11 @@ def combined_plot(feature_list, label_col):
     # A) Convert numeric codes -> text for each feature in `input_mapping`
     for col, text_to_num_dict in input_mapping.items():
         if col in df_copy.columns:
+            # Reverse mapping: "Yes"->1 becomes 1->"Yes"
             num_to_text = {v: k for k, v in text_to_num_dict.items()}
             df_copy[col] = df_copy[col].map(num_to_text).fillna(df_copy[col])
 
-    # B) Convert label 0/1 to text if label_col is in predictor.prediction_map
+    # B) Convert label 0/1 to text in df_copy if label_col is in predictor.prediction_map
     if label_col in predictor.prediction_map and label_col in df_copy.columns:
         zero_text, one_text = predictor.prediction_map[label_col]
         label_map = {0: zero_text, 1: one_text}
@@ -514,8 +501,8 @@ with gr.Blocks(css=".gradio-container {max-width: 1200px;}") as demo:
             ("YMDERSUD5ANY", "YMDERSUD5ANY: MDE or SUD in past year?"),
             ("YMSUD5YANY",   "YMSUD5YANY: Past-year MDE & substance use disorder"),
             ("YMIUD5YANY",   "YMIUD5YANY: Past-year MDE & illicit drug use disorder"),
-            ("YMIMS5YANY",   "YMIUD5YANY: Past-year MDE + severe impairment + substance use"),
-            ("YMIMI5YANY",   "YIMI5YANY: Past-year MDE w/ severe impairment & illicit drug use")
+            ("YMIMS5YANY",   "YMIMS5YANY: Past-year MDE + severe impairment + substance use"),
+            ("YMIMI5YANY",   "YMIMI5YANY: Past-year MDE w/ severe impairment & illicit drug use")
         ]
         cat1_inputs = []
         for col, label_text in cat1_col_labels:
@@ -593,7 +580,7 @@ with gr.Blocks(css=".gradio-container {max-width: 1200px;}") as demo:
         out_count    = gr.Markdown(label="Total Patient Count")
         out_nn       = gr.Markdown(label="Nearest Neighbors Summary")
         out_bar_input= gr.Plot(label="Input Feature Counts")
-        out_bar_label= gr.Plot(label="Predicted Label Counts (Stacked Bar: 0 & 1)")
+        out_bar_label= gr.Plot(label="Predicted Label Counts")
 
         # Connect the predict button to the predict function
         predict_btn.click(
@@ -613,7 +600,7 @@ with gr.Blocks(css=".gradio-container {max-width: 1200px;}") as demo:
     with gr.Tab("Distribution/Co-occurrence"):
         gr.Markdown("### Select 1 or 2 features + 1 label to see a bar chart.")
 
-        # Show only your input features
+        # Show only your 25 input features
         list_of_features = sorted(input_mapping.keys())
         # Show all label columns from the predictor map
         list_of_labels = sorted(predictor.prediction_map.keys())
@@ -637,4 +624,4 @@ with gr.Blocks(css=".gradio-container {max-width: 1200px;}") as demo:
         )
 
 # Finally, launch the Gradio app
-demo.launch()
+demo.launch()