src/data/extractors/base.py

"""
Abstract base class for feature extractors.
Dependency Inversion Principle: Depend on abstractions for feature extraction.
"""
from abc import ABC, abstractmethod
from typing import Dict, Any
import geopandas as gpd


class BaseFeatureExtractor(ABC):
    """
    Abstract base class for watershed/basin feature extractors.
    Each extractor extracts specific category of features (topological, soil, etc.)
    """

    def __init__(self, **kwargs):
        """
        Initialize extractor.

        Args:
            **kwargs: Extractor-specific parameters
        """
        self.params = kwargs

    @abstractmethod
    def extract(self, basin_gdf: gpd.GeoDataFrame) -> Dict[str, Any]:
        """
        Extract features from basin geometry.

        Args:
            basin_gdf: GeoDataFrame with basin boundaries

        Returns:
            Dictionary of extracted features
        """
        pass

    def get_feature_names(self) -> list:
        """
        Get list of feature names this extractor produces.

        Returns:
            List of feature names
        """
        return []

    def validate_inputs(self, basin_gdf: gpd.GeoDataFrame) -> bool:
        """
        Validate input geodataframe.

        Args:
            basin_gdf: GeoDataFrame to validate

        Returns:
            True if valid
        """
        if basin_gdf is None or len(basin_gdf) == 0:
            return False
        if basin_gdf.geometry.is_empty.any():
            return False
        return True