fix: commit artifacts, remove unused import, extend YrSold range

- Unignore ml/artifacts/*.joblib so Streamlit Cloud can load the model
(stack.joblib is 19MB, within GitHub's 100MB limit)
- Remove unused `import xgboost as xgb` from export_model.py
- Extend YrSold input range to 2030 (default 2024) in Streamlit app

.gitignore

@@ -1,5 +1,5 @@
-# Model artifacts (generated by ml/export_model.py)
-ml/artifacts/*.joblib
+# Model artifacts — committed to repo so Streamlit Cloud can load them
+# ml/artifacts/*.joblib
 
 # Training data
 data/*.csv

ml/artifacts/core_features.joblib

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20ef738072a798ec4c9bef5a7c5778903f2b52f3d0178930fdc1e3323f3214a8
+size 663

ml/artifacts/manual_processor.joblib

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:859115054589f37674e7cfcfe876821a2dfe18b7d6fb6bd6168b483ff729face
+size 313

ml/artifacts/stack.joblib

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4edd6699f54362215a59935d9299e5c8f1e19ba715faf00f1c7f0dc885f8fe89
+size 19565408

ml/export_model.py

@@ -18,7 +18,6 @@ from sklearn.ensemble import GradientBoostingRegressor, StackingRegressor
 from sklearn.linear_model import SGDRegressor
 from sklearn.kernel_ridge import KernelRidge
 from sklearn.metrics import mean_squared_error, mean_absolute_error
-import xgboost as xgb
 from catboost import CatBoostRegressor
 
 from ml.preprocessing import (

streamlit_app.py

@@ -81,7 +81,7 @@ 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)
+    YrSold       = st.number_input("Year sold", 2006, 2030, 2024)
     Neighborhood = st.selectbox("Neighborhood", [
         "NAmes", "CollgCr", "OldTown", "Edwards", "Somerst", "NridgHt",
         "Gilbert", "Sawyer", "NWAmes", "SawyerW", "BrkSide", "Crawfor",