Adding LAISS feature names to be used in the get_features method.

src/laiss_resspect_classifier/laiss_feature_extractor.py

@@ -1,13 +1,121 @@
+import itertools
 import numpy as np
 
 from resspect.feature_extractors import LightCurve
 
 class LaissFeatureExtractor(LightCurve):
+    # these features were previously used with ZTF bands [r,g].
+    #! But they don't have to be presumably???
+    feature_names = [
+        'feature_amplitude_magn_*',
+        'feature_anderson_darling_normal_magn_*',
+        'feature_beyond_1_std_magn_*',
+        'feature_beyond_2_std_magn_*',
+        'feature_cusum_magn_*',
+        'feature_inter_percentile_range_2_magn_*',
+        'feature_inter_percentile_range_10_magn_*',
+        'feature_inter_percentile_range_25_magn_*',
+        'feature_kurtosis_magn_*',
+        'feature_linear_fit_slope_magn_*',
+        'feature_linear_fit_slope_sigma_magn_*',
+        'feature_magnitude_percentage_ratio_40_5_magn_*',
+        'feature_magnitude_percentage_ratio_20_5_magn_*',
+        'feature_mean_magn_',
+        'feature_median_absolute_deviation_magn_*',
+        'feature_percent_amplitude_magn_*',
+        'feature_median_buffer_range_percentage_10_magn_*',
+        'feature_median_buffer_range_percentage_20_magn_*',
+        'feature_percent_difference_magnitude_percentile_5_magn_*',
+        'feature_percent_difference_magnitude_percentile_10_magn_*',
+        'feature_skew_magn_*',
+        'feature_standard_deviation_magn_*',
+        'feature_stetson_k_magn_*',
+        'feature_weighted_mean_magn_*',
+        'feature_anderson_darling_normal_flux_*',
+        'feature_cusum_flux_*',
+        'feature_excess_variance_flux_*',
+        'feature_kurtosis_flux_*',
+        'feature_mean_variance_flux_*',
+        'feature_skew_flux_*',
+        'feature_stetson_k_flux_*'
+    ]
+
+    # these are galaxy features are used with the LSST bands [u,g,r,i,z,y]
+    other_feature_names = [
+        '*momentXX',
+        '*momentXY',
+        '*momentYY',
+        '*momentR1',
+        '*momentRH',
+        '*PSFFlux',
+        '*ApFlux',
+        '*KronFlux',
+        '*KronRad',
+        '*ExtNSigma',
+        '*ApMag_*KronMag', #! Confirm with Haille that this is correct
+        '*ApMag_*KronMag', #! Confirm with Haille that this is correct
+        '*ApMag_*KronMag', #! Confirm with Haille that this is correct
+        '*ApMag_*KronMag', #! Confirm with Haille that this is correct
+        '*ApMag_*KronMag', #! Confirm with Haille that this is correct
+    ]
+
+    # these features don't seem to be related to bands
+    yet_more_feature_names = [
+        'i-z',
+        '7DCD',
+        'dist/DLR'
+    ]
+
     def __init__(self, **kwargs):
         super().__init__(**kwargs)
 
+    def get_features(self, filters: list) -> list[str]:
+        """Produce the full list of feature names for the given filters.
+
+        Parameters
+        ----------
+        filters : list[str]
+            List of filters to use in the feature names. i.e. ['r', 'g']
+
+        Returns
+        -------
+        list[str]
+            The complete list of feature names for the given filters, plus the
+            other features that don't depend on the filter.
+        """
+        features = []
+        features.extend(self._get_features_per_filter(self.feature_names, filters))
+
+        LSST_filters = ['u', 'g', 'r', 'i', 'z', 'y']
+        features.extend(self._get_features_per_filter(self.other_feature_names, LSST_filters))
+
+        features.extend(self.yet_more_feature_names)
+
+        return features
+
     def fit(self, band:str = None) -> np.ndarray:
         pass
 
     def fit_all(self) -> np.ndarray:
         pass
+
+    def _get_features_per_filter(self, features: list, filters: list) -> list[str]:
+        """Simple function to get all possible combinations of features and filters.
+        Will replace the '*' in the feature name with the filter name.
+
+        i.e. features = ['example_*'], filters = ['r', 'g']. Returns ['example_r', 'example_g']
+
+        Parameters
+        ----------
+        features : list[str]
+            List of features where each '*' in the name will be replaced with the filter name.
+        filters : list[str]
+            List of filters to replace the '*' in the feature names.
+
+        Returns
+        -------
+        list[str]
+            List of features with the '*' replaced by the filter name.
+        """
+
+        return [pair[0].replace('*', pair[1]) for pair in itertools.product(features, filters)]