Create app.py

app.py

@@ -0,0 +1,88 @@
+import streamlit as st
+import requests
+
+API_URL = "https://lokiiparihar-superkart-api.hf.space/predict"  # use the working API
+
+st.set_page_config(page_title="Superkart Sales Prediction", layout="centered")
+
+st.title("Sales Prediction App")
+st.write("This tool predicts Superkart sales. Enter the required information below.")
+
+# Model Choice
+model_choice = st.selectbox(
+    "Select Model",
+    options=["dt", "xgb"],
+    format_func=lambda x: {
+        "dt": "Decision Tree",
+        "xgb": "XGBoost",
+        "rf": "Random Forest",
+        "lr": "Linear Regression",
+    }.get(x, x),
+)
+
+# Inputs (set defaults so the API call has valid values)
+col1, col2 = st.columns(2)
+
+with col1:
+    product_weight = st.number_input("Product Weight", min_value=0.0, value=12.5, step=0.1)
+    sugar = st.selectbox("Sugar Content", [0, 1, 2], index=0)
+    area = st.number_input("Allocated Area", min_value=0.0, value=0.08, step=0.01)
+    product_type = st.number_input("Product Type Code", min_value=0, value=0, step=1)
+
+with col2:
+    mrp = st.number_input("Product MRP", min_value=0.0, value=249.99, step=1.0)
+    store_size = st.selectbox("Store Size Code", [0, 1, 2], index=1)
+    city = st.selectbox("City Type Code", [0, 1, 2], index=0)
+    store_type = st.number_input("Store Type Code", min_value=0, value=1, step=1)
+    store_age = st.number_input("Store Age", min_value=0, value=15, step=1)
+
+# Build payload EXACTLY as your working notebook request expects
+sample = {
+    "Product_Weight": float(product_weight),
+    "Product_Sugar_Content": float(sugar),
+    "Product_Allocated_Area": float(area),
+    "Product_Type": int(product_type),
+    "Product_MRP": float(mrp),
+    "Store_Size": int(store_size),
+    "Store_Location_City_Type": int(city),
+    "Store_Type": int(store_type),
+    "Store_Age": int(store_age),
+    "model": model_choice,
+}
+
+st.subheader("Payload being sent")
+st.json(sample)
+
+if st.button("Predict", type="primary"):
+    try:
+        headers = {"Content-Type": "application/json"}
+        with st.spinner("Calling prediction API..."):
+            response = requests.post(API_URL, json=sample, headers=headers, timeout=30)
+
+        st.write("Status Code:", response.status_code)
+
+        # Show raw response for debugging
+        st.write("Raw Response:")
+        st.code(response.text)
+
+        if response.headers.get("content-type", "").startswith("application/json"):
+            result = response.json()
+            st.write("Parsed JSON:")
+            st.json(result)
+
+            # Try common keys (your API might return a different one)
+            pred_key = next((k for k in ["Prediction", "prediction", "pred", "result", "output"] if k in result), None)
+            if pred_key:
+                st.success(f"Prediction: {result[pred_key]}")
+            else:
+                st.info("Prediction key not found. See JSON above.")
+        else:
+            st.error("API did not return JSON. See raw response above.")
+
+    except requests.exceptions.RequestException as e:
+        st.error(f"Request failed: {e}")
+
+# IMPORTANT:
+# Streamlit apps do NOT use app.run(). Remove any Flask-related code.
+# if __name__ == '__main__':
+#     app.run(debug=True)