Refactor non-dashboard modules

.github/copilot-instructions.md

@@ -42,7 +42,7 @@ For project goals and setup, see [README.md](../README.md).
 
 - Follow the numbered script sequence: `00grids.sh` → `01darts.sh` → `02alphaearth.sh` → `03era5.sh` → `04arcticdem.sh` → `05train.sh`
 - Each pipeline stage should produce **reproducible intermediate outputs**
-- Use `src/entropice/paths.py` for consistent path management
+- Use `src/entropice/utils/paths.py` for consistent path management
 - Environment variable `FAST_DATA_DIR` controls data directory location (default: `./data`)
 
 ### Storage Hierarchy
@@ -64,16 +64,37 @@ DATA_DIR/
 
 ### Core Modules (`src/entropice/`)
 
-- **`grids.py`**: H3/HEALPix spatial grid generation with watermask
+The codebase is organized into four main packages:
+
+- **`entropice.ingest`**: Data ingestion from external sources
+- **`entropice.spatial`**: Spatial operations and grid management
+- **`entropice.ml`**: Machine learning workflows
+- **`entropice.utils`**: Common utilities
+
+#### Data Ingestion (`src/entropice/ingest/`)
+
 - **`darts.py`**: RTS label extraction from DARTS v2 dataset
 - **`era5.py`**: Climate data processing from ERA5 (Arctic-aligned years: Oct 1 - Sep 30)
 - **`arcticdem.py`**: Terrain analysis from 32m Arctic elevation data
 - **`alphaearth.py`**: Satellite image embeddings via Google Earth Engine
+
+#### Spatial Operations (`src/entropice/spatial/`)
+
+- **`grids.py`**: H3/HEALPix spatial grid generation with watermask
 - **`aggregators.py`**: Raster-to-vector spatial aggregation engine
+- **`watermask.py`**: Ocean masking utilities
+- **`xvec.py`**: Extended vector operations for xarray
+
+#### Machine Learning (`src/entropice/ml/`)
+
 - **`dataset.py`**: Multi-source data integration and feature engineering
 - **`training.py`**: Model training with eSPA, XGBoost, Random Forest, KNN
 - **`inference.py`**: Batch prediction pipeline for trained models
+
+#### Utilities (`src/entropice/utils/`)
+
 - **`paths.py`**: Centralized path management
+- **`codecs.py`**: Custom codecs for data serialization
 
 ### Dashboard (`src/entropice/dashboard/`)
 
@@ -110,10 +131,10 @@ pixi run pytest
 
 ### Common Tasks
 
-- **Generate grids**: Use `grids.py` CLI
-- **Process labels**: Use `darts.py` CLI
-- **Train models**: Use `training.py` CLI with TOML config
-- **Run inference**: Use `inference.py` CLI
+- **Generate grids**: Use `spatial/grids.py` CLI
+- **Process labels**: Use `ingest/darts.py` CLI
+- **Train models**: Use `ml/training.py` CLI with TOML config
+- **Run inference**: Use `ml/inference.py` CLI
 - **View results**: `pixi run dashboard`
 
 ## Key Design Patterns
@@ -162,10 +183,10 @@ Training features:
 
 To extend Entropice:
 
-- **New data source**: Follow patterns in `era5.py` or `arcticdem.py`
-- **Custom aggregations**: Add to `_Aggregations` dataclass in `aggregators.py`
-- **Alternative labels**: Implement extractor following `darts.py` pattern
-- **New models**: Add scikit-learn compatible estimators to `training.py`
+- **New data source**: Follow patterns in `ingest/era5.py` or `ingest/arcticdem.py`
+- **Custom aggregations**: Add to `_Aggregations` dataclass in `spatial/aggregators.py`
+- **Alternative labels**: Implement extractor following `ingest/darts.py` pattern
+- **New models**: Add scikit-learn compatible estimators to `ml/training.py`
 - **Dashboard pages**: Add Streamlit pages to `dashboard/` module
 
 ## Important Notes
@@ -175,13 +196,13 @@ To extend Entropice:
 - Handle **antimeridian crossing** in polar regions
 - Use **batch processing** for GPU memory management
 - Notebooks are for exploration only - **keep production code in `src/`**
-- Always use **absolute paths** or paths from `paths.py`
+- Always use **absolute paths** or paths from `utils/paths.py`
 
 ## Common Issues
 
 - **Memory**: Use batch processing and Dask chunking for large datasets
 - **GPU OOM**: Reduce batch size in inference or training
-- **Antimeridian**: Use proper handling in `aggregators.py` for polar grids
+- **Antimeridian**: Use proper handling in `spatial/aggregators.py` for polar grids
 - **Temporal alignment**: ERA5 uses Arctic-aligned years (Oct-Sep)
 - **CRS**: Compute in EPSG:3413, visualize in EPSG:4326