Update app.py

app.py

@@ -404,7 +404,8 @@ if recipe_submit and uploaded_image:
         predictions = classifyImage(input_image)
         print("Predictions: ", predictions)
         fpredictions = ""
-        predictions_data = []
+        class_names = []
+        confidences = []
     
         # Show the top predictions with percentages
         st.write("Top Predictions:")
@@ -412,26 +413,32 @@ if recipe_submit and uploaded_image:
             fpredictions += f"{class_name}: {confidence:.2f}%,"
             class_name = class_name.replace("_", " ")
             class_name = class_name.title()
-            predictions_data.append({"class_name": class_name, "confidence": confidence})
             st.markdown(f"*{class_name}*: {confidence:.2f}%")
+            class_names.append(class_name)
+            confidences.append(confidence)
         print(fpredictions)
 
         #display as a graph
-        df = pd.DataFrame(predictions_data)
-        bar_chart = alt.Chart(df).mark_bar().encode(
-            x='confidence:Q',  # Quantitative axis for confidence
-            y='class_name:N',  # Nominal axis for class names
-            color=alt.Color('confidence:Q', scale=alt.Scale(domain=[0, 1], range=['gray', 'orange'])),  # Color scale from gray to orange
-            tooltip=['class_name:N', 'confidence:Q']  # Tooltip shows class name and confidence
-        ).properties(
-            width=500,  # Adjust the width of the chart
-            height=300  # Adjust the height of the chart
-        )
-
-        # Display the bar chart in the app
-        st.altair_chart(bar_chart, use_container_width=True)
+        norm = plt.Normalize(min(confidences), max(confidences))
+        cmap = plt.get_cmap('Oranges')
         
+        fig, ax = plt.subplots(figsize=(8, 6))
+        bars = ax.barh(class_names, confidences, color=cmap(norm(confidences)))
+        
+        # Add labels inside the bars, aligned to the right
+        for bar in bars:
+            width = bar.get_width()
+            ax.text(width - 0.02, bar.get_y() + bar.get_height()/2, f'{width*100:.2f}%', 
+                    va='center', ha='right', color='white', fontweight='bold')
+
+
+        ax.set_xlabel('Confidence (%)')
+        ax.set_title('Top Predictions')
 
+        plt.colorbar(plt.cm.ScalarMappable(norm=norm, cmap=cmap), ax=ax, orientation='vertical', label='Confidence')
+        
+        # Display the plot in the Streamlit app
+        st.pyplot(fig)
     # call openai to pick the best classification result based on query
     openAICall = [
         SystemMessage(
@@ -476,8 +483,7 @@ if recipe_submit and uploaded_image:
             RAGresponse = get_response(predictions[0][0])
         print("RAGresponse: ", RAGresponse)
         display_response(RAGresponse)
-else:
-    st.warning("Please upload an image before submitting.", icon=':material/no_meals:')
+
 # elif uploaded_image is not None:
 #     with col1:
 #         # Open the image