feat: add Streamlit web app

Interactive UI with sliders and dropdowns for all house features.
Loads model artifacts directly — no API dependency needed.
Deployable for free on Streamlit Community Cloud.

requirements.txt

@@ -1,3 +1,4 @@
+streamlit>=1.35.0
 fastapi>=0.111.0
 uvicorn[standard]>=0.29.0
 sqlalchemy>=2.0.0

streamlit_app.py

@@ -0,0 +1,163 @@
+import os
+import numpy as np
+import pandas as pd
+import joblib
+import streamlit as st
+
+ARTIFACTS_DIR = os.path.join(os.path.dirname(__file__), "ml", "artifacts")
+
+@st.cache_resource
+def load_model():
+    processor = joblib.load(os.path.join(ARTIFACTS_DIR, "manual_processor.joblib"))
+    stack     = joblib.load(os.path.join(ARTIFACTS_DIR, "stack.joblib"))
+    features  = joblib.load(os.path.join(ARTIFACTS_DIR, "core_features.joblib"))
+    return processor, stack, features
+
+
+def predict(processor, stack, core_features, inputs: dict) -> float:
+    df = pd.DataFrame([inputs])
+    X = df[core_features].copy()
+    X_processed = processor.transform(X)
+    log_pred = stack.predict(X_processed)
+    return float(np.expm1(log_pred)[0])
+
+
+# ── Page config ───────────────────────────────────────────────────────────────
+st.set_page_config(
+    page_title="House Price Predictor",
+    page_icon="🏠",
+    layout="wide",
+)
+
+st.title("🏠 House Price Predictor")
+st.caption("Stacking ensemble: GradientBoosting + CatBoost + KernelRidge · Top 16 / 4,000+ on Kaggle")
+st.divider()
+
+# ── Load model ────────────────────────────────────────────────────────────────
+try:
+    processor, stack, core_features = load_model()
+except Exception as e:
+    st.error(f"Failed to load model artifacts: {e}")
+    st.info("Run `python -m ml.export_model` first to generate the artifacts.")
+    st.stop()
+
+# ── Input form ────────────────────────────────────────────────────────────────
+st.subheader("House Features")
+
+col1, col2, col3 = st.columns(3)
+
+with col1:
+    st.markdown("**Size**")
+    GrLivArea    = st.number_input("Above-grade living area (sq ft)", 300, 6000, 1500, step=50)
+    TotalBsmtSF  = st.number_input("Total basement area (sq ft)", 0, 3000, 800, step=50)
+    GarageArea   = st.number_input("Garage area (sq ft)", 0, 1500, 400, step=50)
+    LotArea      = st.number_input("Lot area (sq ft)", 1000, 100000, 8000, step=500)
+    LotFrontage  = st.number_input("Lot frontage (ft)", 0, 200, 65, step=5)
+
+    st.markdown("**Floors & Extras**")
+    floor_1st    = st.number_input("1st floor area (sq ft)", 0, 4000, 856, step=50)
+    floor_2nd    = st.number_input("2nd floor area (sq ft)", 0, 2000, 0, step=50)
+    WoodDeckSF   = st.number_input("Wood deck (sq ft)", 0, 800, 0, step=10)
+    OpenPorchSF  = st.number_input("Open porch (sq ft)", 0, 500, 0, step=10)
+
+with col2:
+    st.markdown("**Quality & Condition**")
+    OverallQual  = st.slider("Overall quality (1–10)", 1, 10, 6)
+    OverallCond  = st.slider("Overall condition (1–9)", 1, 9, 5)
+    KitchenQual  = st.selectbox("Kitchen quality", ["Ex", "Gd", "TA", "Fa", "Po"], index=2)
+    ExterQual    = st.selectbox("Exterior quality", ["Ex", "Gd", "TA", "Fa", "Po"], index=2)
+    BsmtQual     = st.selectbox("Basement quality", ["Ex", "Gd", "TA", "Fa", "Po", None], index=2)
+    HeatingQC    = st.selectbox("Heating quality", ["Ex", "Gd", "TA", "Fa", "Po"], index=2)
+    FireplaceQu  = st.selectbox("Fireplace quality", [None, "Ex", "Gd", "TA", "Fa", "Po"], index=0)
+
+    st.markdown("**Bathrooms & Rooms**")
+    FullBath     = st.number_input("Full baths (above grade)", 0, 4, 2)
+    HalfBath     = st.number_input("Half baths (above grade)", 0, 2, 0)
+    BsmtFullBath = st.number_input("Basement full baths", 0, 2, 0)
+    TotRmsAbvGrd = st.number_input("Total rooms above grade", 2, 14, 7)
+    Fireplaces   = st.number_input("Fireplaces", 0, 4, 0)
+
+with col3:
+    st.markdown("**Year & Location**")
+    YearBuilt    = st.number_input("Year built", 1870, 2024, 2000)
+    YearRemodAdd = st.number_input("Year remodelled", 1950, 2024, 2000)
+    YrSold       = st.number_input("Year sold", 2006, 2010, 2010)
+    Neighborhood = st.selectbox("Neighborhood", [
+        "NAmes", "CollgCr", "OldTown", "Edwards", "Somerst", "NridgHt",
+        "Gilbert", "Sawyer", "NWAmes", "SawyerW", "BrkSide", "Crawfor",
+        "Mitchel", "NoRidge", "Timber", "IDOTRR", "ClearCr", "StoneBr",
+        "SWISU", "MeadowV", "Blmngtn", "BrDale", "Veenker", "NPkVill", "Blueste",
+    ])
+    MSZoning     = st.selectbox("Zoning", ["RL", "RM", "FV", "RH", "C (all)"], index=0)
+    BldgType     = st.selectbox("Building type", ["1Fam", "2fmCon", "Duplex", "TwnhsE", "Twnhs"], index=0)
+    Foundation   = st.selectbox("Foundation", ["PConc", "CBlock", "BrkTil", "Wood", "Slab", "Stone"], index=0)
+
+    st.markdown("**Garage & Utilities**")
+    GarageQual   = st.selectbox("Garage quality", ["TA", "Gd", "Fa", "Ex", "Po"], index=0)
+    GarageCond   = st.selectbox("Garage condition", ["TA", "Gd", "Fa", "Ex", "Po"], index=0)
+    Electrical   = st.selectbox("Electrical system", ["SBrkr", "FuseA", "FuseF", "FuseP", "Mix"], index=0)
+
+# ── Predict ───────────────────────────────────────────────────────────────────
+st.divider()
+
+if st.button("Predict Price", type="primary", use_container_width=True):
+    inputs = {
+        "GrLivArea": float(GrLivArea),
+        "TotalBsmtSF": float(TotalBsmtSF),
+        "LotArea": float(LotArea),
+        "GarageArea": float(GarageArea),
+        "PoolArea": 0.0,
+        "LotFrontage": float(LotFrontage) if LotFrontage else None,
+        "2ndFlrSF": float(floor_2nd),
+        "LowQualFinSF": 0.0,
+        "BsmtUnfSF": 0.0,
+        "1stFlrSF": float(floor_1st),
+        "WoodDeckSF": float(WoodDeckSF),
+        "OpenPorchSF": float(OpenPorchSF),
+        "EnclosedPorch": 0.0,
+        "3SsnPorch": 0.0,
+        "ScreenPorch": 0.0,
+        "FullBath": int(FullBath),
+        "HalfBath": int(HalfBath),
+        "BsmtFullBath": int(BsmtFullBath),
+        "BsmtHalfBath": 0,
+        "TotRmsAbvGrd": int(TotRmsAbvGrd),
+        "Fireplaces": int(Fireplaces),
+        "YearBuilt": int(YearBuilt),
+        "YrSold": int(YrSold),
+        "YearRemodAdd": int(YearRemodAdd),
+        "OverallQual": int(OverallQual),
+        "OverallCond": int(OverallCond),
+        "HeatingQC": HeatingQC,
+        "BsmtQual": BsmtQual,
+        "PoolQC": None,
+        "ExterQual": ExterQual,
+        "KitchenQual": KitchenQual,
+        "Functional": "Typ",
+        "FireplaceQu": FireplaceQu,
+        "BsmtCond": "TA",
+        "ExterCond": "TA",
+        "Neighborhood": Neighborhood,
+        "MSZoning": MSZoning,
+        "MSSubClass": 20,
+        "LandSlope": "Gtl",
+        "Alley": None,
+        "LandContour": "Lvl",
+        "BldgType": BldgType,
+        "Condition1": "Norm",
+        "RoofStyle": "Gable",
+        "Foundation": Foundation,
+        "SaleCondition": "Normal",
+        "Exterior1st": "VinylSd",
+        "Utilities": "AllPub",
+        "Electrical": Electrical,
+        "GarageQual": GarageQual,
+        "GarageCond": GarageCond,
+    }
+
+    with st.spinner("Predicting..."):
+        try:
+            price = predict(processor, stack, core_features, inputs)
+            st.success(f"### Estimated Sale Price: **${price:,.0f}**")
+        except Exception as e:
+            st.error(f"Prediction error: {e}")