Update app.py

app.py

@@ -5,51 +5,59 @@ from amazon_apparel_recommender import price_quality_recommendations
 # Load data
 df = pd.read_csv("assets/cleaned_metadata.csv")
 
-# Define dropdown options
-brands = sorted(df['brand'].dropna().unique().tolist())
-colors = sorted(df['color'].dropna().unique().tolist())
-product_types = sorted(df['product_type_name'].dropna().unique().tolist())
-
-# Function to get ASINs based on selected filters
-def get_filtered_asins(brand, color, product_type):
-    filtered = df[
+# Precompute dropdown options
+brands = sorted(df['brand'].dropna().unique())
+colors = sorted(df['color'].dropna().unique())
+product_types = sorted(df['product_type_name'].dropna().unique())
+
+# Filter ASINs based on dropdown selections
+def filter_products(brand, color, product_type):
+    filtered_df = df[
         (df['brand'] == brand) &
         (df['color'] == color) &
         (df['product_type_name'] == product_type)
-    ]
-    options = filtered[['asin', 'title']].drop_duplicates()
-    asin_options = [f"{row['asin']} — {row['title'][:60]}" for _, row in options.iterrows()]
-    return asin_options
-
-# Function to extract ASIN and recommend products
-def recommend_from_asin(selected):
-    if not selected:
-        return "Please select a product."
-    asin = selected.split(" — ")[0]
+    ][['asin', 'title']].drop_duplicates()
+
+    if filtered_df.empty:
+        return ["No matches found"]
+
+    return [f"{row['asin']} — {row['title'][:80]}" for _, row in filtered_df.iterrows()]
+
+# Recommend products based on selected ASIN
+def recommend_products(asin_combo):
+    if not asin_combo or asin_combo.startswith("No matches"):
+        return pd.DataFrame(columns=["Title", "Brand", "Price", "Rating"])
+
+    asin = asin_combo.split(" — ")[0]
     results = price_quality_recommendations(asin, df)
     if results.empty:
-        return "No recommendations found for this product."
-    return results[['title', 'brand', 'price', 'review_score']].head(5)
+        return pd.DataFrame(columns=["Title", "Brand", "Price", "Rating"])
+
+    return results[['title', 'brand', 'price', 'review_score']].head(5).rename(
+        columns={"title": "Title", "brand": "Brand", "price": "Price", "review_score": "Rating"}
+    )
 
-# Build Gradio UI
-with gr.Blocks() as demo:
+# Build Gradio interface
+with gr.Blocks(title="Amazon Apparel Recommender") as app:
     gr.Markdown("Amazon Apparel Recommender")
-    gr.Markdown("Select product filters to view similar item suggestions.")
+    gr.Markdown("Select brand, color, and product type to view similar item recommendations.")
 
     with gr.Row():
-        brand_input = gr.Dropdown(choices=brands, label="Brand", interactive=True)
-        color_input = gr.Dropdown(choices=colors, label="Color", interactive=True)
-        type_input = gr.Dropdown(choices=product_types, label="Product Type", interactive=True)
+        brand_input = gr.Dropdown(label="Brand", choices=brands, interactive=True)
+        color_input = gr.Dropdown(label="Color", choices=colors, interactive=True)
+        type_input = gr.Dropdown(label="Product Type", choices=product_types, interactive=True)
 
-    asin_dropdown = gr.Dropdown(label="Select Matching Product (ASIN — Title)", interactive=True)
+    asin_input = gr.Dropdown(label="Matching Product (ASIN — Title)", interactive=True)
 
     with gr.Row():
-        filter_btn = gr.Button(" Find Matching Products")
-        recommend_btn = gr.Button(" Get Recommendations")
+        filter_btn = gr.Button("Find Matching Products")
+        recommend_btn = gr.Button("Get Recommendations")
 
-    output = gr.Dataframe(headers=["Title", "Brand", "Price", "Rating"])
+    output_table = gr.Dataframe(headers=["Title", "Brand", "Price", "Rating"], interactive=False)
 
-    filter_btn.click(fn=get_filtered_asins, inputs=[brand_input, color_input, type_input], outputs=asin_dropdown)
-    recommend_btn.click(fn=recommend_from_asin, inputs=asin_dropdown, outputs=output)
+    # Logic binding
+    filter_btn.click(fn=filter_products, inputs=[brand_input, color_input, type_input], outputs=asin_input)
+    recommend_btn.click(fn=recommend_products, inputs=asin_input, outputs=output_table)
 
-demo.launch()
+# Launch app
+app.launch()