introduce configurable peak filtering

src/depiction/calibration/spectrum/calibration_method_regress_shift.py

@@ -7,7 +7,9 @@
 from depiction.calibration.models.fit_model import fit_model
 from depiction.calibration.spectrum.reference_peak_distances import ReferencePeakDistances
 from depiction.image import MultiChannelImage
+from depiction.image.smoothing.min_filter import MinFilter
 from depiction.image.smoothing.spatial_smoothing_sparse_aware import SpatialSmoothingSparseAware
+from depiction.spectrum.peak_filtering import FilterNHighestIntensityPartitioned, PeakFilteringType
 
 
 class CalibrationMethodRegressShift(CalibrationMethod):
@@ -37,6 +39,7 @@ def __init__(
         input_smoothing_kernel_size: int = 27,
         input_smoothing_kernel_std: float = 10.0,
         min_points: int = 3,
+        peak_filtering: PeakFilteringType | None = None,
     ) -> None:
         self._ref_mz_arr = ref_mz_arr
         self._max_distance = max_distance
@@ -47,8 +50,15 @@ def __init__(
         self._input_smoothing_activated = input_smoothing_activated
         self._input_smoothing_kernel_size = input_smoothing_kernel_size
         self._input_smoothing_kernel_std = input_smoothing_kernel_std
+        self._peak_filtering = peak_filtering
 
     def extract_spectrum_features(self, peak_mz_arr: NDArray[float], peak_int_arr: NDArray[float]) -> DataArray:
+        if self._peak_filtering:
+            # TODO this is sort of problematic, to be reconsidered how we can avoid passing peak arrays as spectrum
+            #      arrays (it should be part of the API and implemented consistently)
+            peak_mz_arr, peak_int_arr = self._peak_filtering.filter_peaks(
+                peak_mz_arr, peak_int_arr, peak_mz_arr, peak_int_arr
+            )
         distances_mz = ReferencePeakDistances.get_distances_max_peak_in_window(
             peak_mz_arr=peak_mz_arr,
             peak_int_arr=peak_int_arr,

src/depiction/tools/calibrate/calibrate.py

@@ -24,6 +24,7 @@
     CalibrationConstantGlobalShiftConfig,
     CalibrationConfig,
 )
+from depiction.tools.filter_peaks.filter_peaks import get_peak_filter
 
 
 def extract_reference_masses(mass_list: Path) -> NDArray[float]:
@@ -41,6 +42,7 @@ def extract_reference_masses(mass_list: Path) -> NDArray[float]:
 def get_calibration_instance(config: CalibrationConfig, mass_list: Path | None):
     match config.method:
         case CalibrationRegressShiftConfig():
+            peak_filtering = get_peak_filter(config.method.peak_filtering) if config.method.peak_filtering else None
             return CalibrationMethodRegressShift(
                 ref_mz_arr=extract_reference_masses(mass_list),
                 max_distance=config.method.max_distance,
@@ -51,6 +53,7 @@ def get_calibration_instance(config: CalibrationConfig, mass_list: Path | None):
                 input_smoothing_kernel_size=config.method.input_smoothing_kernel_size,
                 input_smoothing_kernel_std=config.method.input_smoothing_kernel_std,
                 min_points=config.method.min_points,
+                peak_filtering=peak_filtering,
             )
         case CalibrationConstantGlobalShiftConfig():
             return CalibrationMethodGlobalConstantShift(

src/depiction/tools/calibrate/config.py

@@ -4,6 +4,8 @@
 
 from pydantic import BaseModel, Field
 
+from depiction.tools.filter_peaks.config import FilterPeaksConfig
+
 
 class CalibrationRegressShiftConfig(BaseModel):
     calibration_method: Literal["RegressShift"] = "RegressShift"
@@ -20,6 +22,8 @@ class CalibrationRegressShiftConfig(BaseModel):
     input_smoothing_kernel_std: float = 10.0
     min_points: int = 3
 
+    peak_filtering: FilterPeaksConfig | None = None
+
 
 class CalibrationChemicalPeptideNoiseConfig(BaseModel):
     calibration_method: Literal["ChemicalPeptideNoise"] = "ChemicalPeptideNoise"