Allow integer or boolean mask to be passed to Masker

CHANGELOG.md

@@ -17,6 +17,7 @@
   - Bug fix in the adjoint of the Diagonal Operator for complex values
   - Update conda build action to v2 for 2.5x quicker builds
   - Add docker image & push to [`ghcr.io/tomographicimaging/cil`](https://github.com/TomographicImaging/CIL/pkgs/container/cil)
+  - Allow Masker to take integer arrays in addition to boolean
 
 * 23.1.0
   - Fix bug in IndicatorBox proximal_conjugate
@@ -132,8 +133,8 @@
   - Recon.FBP allows 'astra' backend
   - Fixed PowerMethod for square/non-square, complex/float matrices with stopping criterion.
   - CofR image_sharpness improved for large datasets
-  - Geometry alignmentment fix for 2D datasets
-  - CGLS update for sapyb to enable complex data, bugfix in use of initial
+  - Geometry alignment fix for 2D datasets
+  - CGLS update for sapyb to enable complex data, bugfix in use of initial 
   - added sapyb and deprecated axpby. All algorithm updated to use sapyb.
   - Allow use of square brackets in file paths to TIFF and Nikon datasets
 
@@ -175,7 +176,7 @@
   - Fixed bug in Zeiss reader geometry direction of rotation
 
 * 21.0.0
-  - Show2D now takes 4D datasets and slice infomation as input
+  - Show2D now takes 4D datasets and slice information as input
   - TIGRE reconstruction package wrapped for cone-beam tomography
   - Datacontainers have get_slice method which returns a dataset with a single slice of the data
   - Datacontainers have reorder method which reorders the data in memory as requested, or for use with 'astra' or 'tigre'

Wrappers/Python/cil/processors/CentreOfRotationCorrector.py

@@ -70,7 +70,7 @@ def xcorrelation(slice_index='centre', projection_index=0, ang_tol=0.1):
 
         >>> processor = CentreOfRotationCorrector.xcorrelation(slice_index=120)
         >>> processor.set_input(data)
-        >>> data_centred = processor.get_ouput()
+        >>> data_centred = processor.get_output()
 
 
         Note

Wrappers/Python/cil/processors/Masker.py

@@ -18,23 +18,52 @@
 # CIL Developers, listed at: https://github.com/TomographicImaging/CIL/blob/master/NOTICE.txt
 
 from cil.framework import DataProcessor, AcquisitionData, ImageData, ImageGeometry, DataContainer
-import warnings
 import numpy
 from scipy import interpolate
 
 class Masker(DataProcessor):
     r'''
-    Processor to fill missing values provided by mask. Please use the desiried method to configure a processor for your needs.
+    Processor to fill missing values provided by mask.
+    Please use the desired method to configure a processor for your needs.
+
+    Parameters
+    ----------
+    mask : DataContainer, ImageData, AcquisitionData, numpy.ndarray
+        A boolean array with the same dimensions as input, where 'False' represents masked values.
+        Alternatively an integer array where 0 represents masked values, and any other value represents unmasked values.
+        Mask can be generated using 'MaskGenerator' processor to identify outliers. 
+    mode : {'value', 'mean', 'median', 'interpolate'}, default='value'
+        The method to fill in missing values 
+    value : float, default=0
+        Substitute all outliers with a specific value if method='value', otherwise discarded.
+    axis : str or int
+        Specify axis as int or from 'dimension_labels' to calculate mean, median or interpolation
+        (depending on mode) along that axis
+    method : {'linear', 'nearest', 'zeros', 'linear', 'quadratic', 'cubic', 'previous', 'next'}, default='linear'
+        Interpolation method to use.
+
+    Returns
+    -------
+    DataContainer or its subclass with masked outliers
+
     '''
 
     @staticmethod
     def value(mask=None, value=0):
         r'''This sets the masked values of the input data to the requested value.
 
-        :param mask: A boolean array with the same dimensions as input, where 'False' represents masked values. Mask can be generated using 'MaskGenerator' processor to identify outliers. 
-        :type mask: DataContainer, ImageData, AcquisitionData, numpy.ndarray
-        :param value: values to be assigned to missing elements
-        :type value: float, default=0
+        Parameters
+        ----------
+        mask : DataContainer, ImageData, AcquisitionData, numpy.ndarray
+            A boolean array with the same dimensions as input, where 'False' represents masked values.
+            Alternatively an integer array where 0 represents masked values, and any other value represents unmasked values.
+            Mask can be generated using 'MaskGenerator' processor to identify outliers. 
+        value : float, default=0
+            Values to be assigned to missing elements
+
+        Returns
+        -------
+        Masker processor
         '''
 
         processor = Masker(mode='value', mask=mask, value=value)
@@ -45,10 +74,18 @@ def value(mask=None, value=0):
     def mean(mask=None, axis=None):
         r'''This sets the masked values of the input data to the mean of the unmasked values across the array or axis.
 
-        :param mask: A boolean array with the same dimensions as input, where 'False' represents masked values. Mask can be generated using 'MaskGenerator' processor to identify outliers.  
-        :type mask: DataContainer, ImageData, AcquisitionData, numpy.ndarray
-        :param axis: specify axis as int or from 'dimension_labels' to calculate mean. 
-        :type axis: str, int
+        Parameters
+        ----------
+        mask : DataContainer, ImageData, AcquisitionData, numpy.ndarray
+            A boolean array with the same dimensions as input, where 'False' represents masked values.
+            Alternatively an integer array where 0 represents masked values, and any other value represents unmasked values.
+            Mask can be generated using 'MaskGenerator' processor to identify outliers. 
+        axis : str, int
+            Specify axis as int or from 'dimension_labels' to calculate mean.
+
+        Returns
+        -------
+        Masker processor
         '''
 
         processor = Masker(mode='mean', mask=mask, axis=axis)
@@ -59,10 +96,18 @@ def mean(mask=None, axis=None):
     def median(mask=None, axis=None):
         r'''This sets the masked values of the input data to the median of the unmasked values across the array or axis.
 
-        :param mask: A boolean array with the same dimensions as input, where 'False' represents masked values. Mask can be generated using 'MaskGenerator' processor to identify outliers.  
-        :type mask: DataContainer, ImageData, AcquisitionData, numpy.ndarray
-        :param axis: specify axis as int or from 'dimension_labels' to calculate median. 
-        :type axis: str, int
+        Parameters
+        ----------
+        mask : DataContainer, ImageData, AcquisitionData, numpy.ndarray
+            A boolean array with the same dimensions as input, where 'False' represents masked values.
+            Alternatively an integer array where 0 represents masked values, and any other value represents unmasked values.
+            Mask can be generated using 'MaskGenerator' processor to identify outliers. 
+        axis : str, int
+            Specify axis as int or from 'dimension_labels' to calculate median.
+
+        Returns
+        -------
+        Masker processor
         '''
 
         processor = Masker(mode='median', mask=mask, axis=axis)
@@ -71,14 +116,22 @@ def median(mask=None, axis=None):
 
     @staticmethod
     def interpolate(mask=None, axis=None, method='linear'):
-        r'''This operates over the specified axis and uses 1D interpolation over remaining flattened array to fill in missing vaues.
+        r'''This operates over the specified axis and uses 1D interpolation over remaining flattened array to fill in missing values.
 
-        :param mask: A boolean array with the same dimensions as input, where 'False' represents masked values. Mask can be generated using 'MaskGenerator' processor to identify outliers.  
-        :type mask: DataContainer, ImageData, AcquisitionData, numpy.ndarray
-        :param axis: specify axis as int or from 'dimension_labels' to loop over and perform 1D interpolation. 
-        :type axis: str, int
-        :param method: One of the following interpoaltion methods: linear, nearest, zeros, linear, quadratic, cubic, previous, next
-        :param method: str, default='linear'
+        Parameters
+        ----------
+        mask : DataContainer, ImageData, AcquisitionData, numpy.ndarray
+            A boolean array with the same dimensions as input, where 'False' represents masked values.
+            Alternatively an integer array where 0 represents masked values, and any other value represents unmasked values.
+            Mask can be generated using 'MaskGenerator' processor to identify outliers. 
+        axis : str, int
+            Specify axis as int or from 'dimension_labels' to loop over and perform 1D interpolation.
+        method : {'linear', 'nearest', 'zeros', 'linear', 'quadratic', 'cubic', 'previous', 'next'}, default='linear'
+            Interpolation method to use.
+
+        Returns
+        -------
+        Masker processor
         '''
 
         processor = Masker(mode='interpolate', mask=mask, axis=axis, method=method)
@@ -91,22 +144,6 @@ def __init__(self,
                  value = 0,
                  axis = None,
                  method = 'linear'):
-
-        r'''Processor to fill missing values provided by mask.
-
-        :param mask: A boolean array with the same dimensions as input, where 'False' represents masked values. Mask can be generated using 'MaskGenerator' processor to identify outliers. 
-        :type mask: DataContainer, ImageData, AcquisitionData, numpy.ndarray
-        :param mode: a method to fill in missing values (value, mean, median, interpolate)
-        :type mode: str, default=value
-        :param value: substitute all outliers with a specific value
-        :type value: float, default=0
-        :param axis: specify axis as int or from 'dimension_labels' to calculate mean or median in respective modes 
-        :type axis: str or int
-        :param method: One of the following interpoaltion methods: linear, nearest, zeros, linear, quadratic, cubic, previous, next
-        :param method: str, default='linear'
-        :return: DataContainer or it's subclass with masked outliers
-        :rtype: DataContainer or it's subclass   
-        '''
 
         kwargs = {'mask': mask,
                   'mode': mode,
@@ -153,11 +190,10 @@ def process(self, out=None):
         except:
             mask_arr = self.mask
 
-        try:
-            mask_invert = ~mask_arr
-        except TypeError:
-            raise TypeError("Mask expected to be a boolean array got {}".format(mask_arr.dtype))
-
+        mask_arr = numpy.array(mask_arr, dtype=bool)
+
+        mask_invert = ~mask_arr
+
         try:
             axis_index = data.dimension_labels.index(self.axis)             
         except:
@@ -199,7 +235,7 @@ def process(self, out=None):
         elif self.mode == 'interpolate':
             if self.method not in ['linear', 'nearest', 'zeros', 'linear', \
                                         'quadratic', 'cubic', 'previous', 'next']:
-                raise TypeError("Wrong interpolation method, one of the follwoing is expected:\n" + 
+                raise TypeError("Wrong interpolation method, one of the following is expected:\n" + 
                                 "linear, nearest, zeros, linear, quadratic, cubic, previous, next")
 
             ndim = data.number_of_dimensions

Wrappers/Python/cil/processors/Normaliser.py

@@ -30,7 +30,7 @@ class Normaliser(Processor):
     Input: AcquisitionData
     Parameter: 2D projection with flat field (or stack)
                2D projection with dark field (or stack)
-    Output: AcquisitionDataSetn
+    Output: AcquisitionDataSet
     '''
 
     def __init__(self, flat_field = None, dark_field = None, tolerance = 1e-5):

Wrappers/Python/test/test_DataProcessor.py

@@ -23,7 +23,6 @@
 from cil.framework import ImageGeometry, VectorGeometry, AcquisitionGeometry
 from cil.framework import ImageData, AcquisitionData
 from cil.utilities import dataexample
-from timeit import default_timer as timer
 
 from cil.framework import AX, CastDataContainer, PixelByPixelDataProcessor
 from cil.recon import FBP
@@ -2521,15 +2520,33 @@ def setUp(self):
         self.mask_manual = DataContainer(mask_manual, dimension_labels=self.data.dimension_labels) 
         self.mask_generated = MaskGenerator.special_values()(self.data)
 
+        # make a copy of mask_manual with 1s and 0s instead of bools:
+        mask_int_manual = mask_manual.astype(numpy.int32)
+        self.mask_int_manual = DataContainer(mask_int_manual, dimension_labels=self.data.dimension_labels)
+
+
     def test_Masker_Manual(self):
         self.Masker_check(self.mask_manual, self.data, self.data_init)
 
     def test_Masker_generated(self):
         self.Masker_check(self.mask_generated, self.data, self.data_init)
 
+    def test_Masker_with_integer_mask(self):
+        self.Masker_check(self.mask_int_manual, self.data, self.data_init)
+
+    def test_Masker_doesnt_modify_input_mask(self):
+        mask = self.mask_manual.copy()
+        self.Masker_check(self.mask_manual, self.data, self.data_init)
+        numpy.testing.assert_array_equal(mask.as_array(), self.mask_manual.as_array())
+
+    def test_Masker_doesnt_modify_input_integer_mask(self):
+        mask = self.mask_int_manual.copy()
+        self.Masker_check(self.mask_int_manual, self.data, self.data_init)
+        numpy.testing.assert_array_equal(mask.as_array(), self.mask_int_manual.as_array())
+
     def Masker_check(self, mask, data, data_init): 
 
-        # test vaue mode
+        # test value mode
         m = Masker.value(mask=mask, value=10)
         m.set_input(data)
         res = m.process()