Merge branch 'master' into kl-multiple-exits

Signed-off-by: Edoardo Pasca <[email protected]>

.github/workflows/build.yml

@@ -122,7 +122,7 @@ jobs:
     - uses: conda-incubator/setup-miniconda@v3
       with:
         python-version: ${{ matrix.python-version }}
-        mamba-version: "*"
+        mamba-version: "1.5"
         channels: conda-forge
     - name: conda build & test
       working-directory: recipe
@@ -243,7 +243,7 @@ jobs:
         registry: ghcr.io
         username: ${{ github.actor }}
         password: ${{ secrets.GITHUB_TOKEN }}
-    - uses: docker/build-push-action@v5
+    - uses: docker/build-push-action@v6
       with:
         cache-from: type=gha
         cache-to: type=gha,mode=max
@@ -254,7 +254,7 @@ jobs:
       run: >
         docker run --rm -v .:/CIL tomographicimaging/cil:test /bin/bash -c
         'python -m unittest discover -v /CIL/Wrappers/Python/test'
-    - uses: docker/build-push-action@v5
+    - uses: docker/build-push-action@v6
       with:
         cache-from: type=gha
         cache-to: type=gha,mode=max

CHANGELOG.md

@@ -12,15 +12,18 @@
     - Use ravel instead of flat in KullbackLeibler numba backend (#1874)
     - Upgrade Python wrapper (#1873, #1875)
     - Updated the documentation for the algorithm base class (#1809)
-    - Add checks on out argument passed to processors to ensure corrrect dtype and size (#1805)
+    - Add checks on out argument passed to processors to ensure correct dtype and size (#1805)
     - Internal refactor: Replaced string-based label checks with enum-based checks for improved type safety and consistency (#1692)
     - Internal refactor: Separate framework into multiple files (#1692)
     - Allow the SIRT algorithm to take `initial=None` (#1906)
+    - Add checks on equality method of `AcquisitionData` and `ImageData` for equality of data type and geometry (#1919)
+    - Add check on equality method of `AcquisitionGeometry` for equality of dimension labels (#1919)
     - Removed multiple exits from numba implementation of KullbackLeibler divergence (#1901)
   - Testing:
     - New unit tests for operators and functions to check for in place errors and the behaviour of `out` (#1805)
     - Updates in SPDHG vs PDHG unit test to reduce test time and adjustments to parameters (#1898)
     - Drop Jenkins in favour of GHA for conda builds (#1914)
+    - New unit tests for `DataContainer`, `AcquisitionData` and `ImageData` to check equality method (`__eq__`) behaves as expected (#1919)
   - Bug fixes:
     - `ImageData` removes dimensions of size 1 from the input array. This fixes an issue where single slice reconstructions from 3D data would fail due to shape mismatches (#1885)
     - Make Binner accept accelerated=False (#1887)

README.md

@@ -27,15 +27,16 @@ conda create --name cil -c conda-forge -c https://software.repos.intel.com/pytho
 To install CIL and the additional packages and plugins needed to run the [CIL demos](https://github.com/TomographicImaging/CIL-Demos) install the environment with:
 
 ```sh
-conda create --name cil -c conda-forge -c https://software.repos.intel.com/python/conda -c ccpi cil=24.1.0 astra-toolbox=*=cuda* tigre ccpi-regulariser tomophantom ipywidgets
+conda create --name cil -c conda-forge -c https://software.repos.intel.com/python/conda -c ccpi cil=24.1.0 astra-toolbox=*=cuda* tigre ccpi-regulariser tomophantom ipykernel ipywidgets
 ```
 
 where:
-- `astra-toolbox` enables CIL support for [ASTRA toolbox](http://www.astra-toolbox.com) CPU projector (2D Parallel beam only) (GPLv3 license)
+- `astra-toolbox=*=py*` enables CIL support for [ASTRA toolbox](http://www.astra-toolbox.com) CPU projector (2D Parallel beam only) (GPLv3 license)
 - `astra-toolbox=*=cuda*` (requires an NVIDIA GPU) enables CIL support for [ASTRA toolbox](http://www.astra-toolbox.com) GPU projectors (GPLv3 license)
 - `tigre` (requires an NVIDIA GPU) enables support for [TIGRE](https://github.com/CERN/TIGRE) toolbox projectors (BSD license)
 - `ccpi-regulariser` is the [CCPi Regularisation Toolkit](https://github.com/TomographicImaging/CCPi-Regularisation-Toolkit)
 - `tomophantom` can generate phantoms to use as test data [Tomophantom](https://github.com/dkazanc/TomoPhantom)
+- `ipykernel`  provides the IPython kernel for Jupyter (allowing jupyter notebooks to be run)
 - `ipywidgets` enables visulisation tools within jupyter noteboooks
 
 ### Dependencies

Wrappers/Python/cil/framework/acquisition_data.py

@@ -73,6 +73,31 @@ def __init__(self,
 
         super(AcquisitionData, self).__init__(array, deep_copy, geometry=geometry,**kwargs)
 
+    def __eq__(self, other):
+        '''
+        Check if two AcquisitionData objects are equal. This is done by checking if the geometry, data and dtype are equal.
+        Also, if the other object is a numpy.ndarray, it will check if the data and dtype are equal.
+        
+        Parameters
+        ----------
+        other: AcquisitionData or numpy.ndarray
+            The object to compare with.
+        
+        Returns
+        -------
+        bool
+            True if the two objects are equal, False otherwise.
+        '''
+
+        if isinstance(other, AcquisitionData):
+            if numpy.array_equal(self.as_array(), other.as_array()) \
+                and self.geometry == other.geometry \
+                and self.dtype == other.dtype:
+                return True 
+        elif numpy.array_equal(self.as_array(), other) and self.dtype==other.dtype:
+            return True
+        else:
+            return False
 
     def get_slice(self,channel=None, angle=None, vertical=None, horizontal=None, force=False):
         '''Returns a new dataset of a single slice in the requested direction.'''

Wrappers/Python/cil/framework/acquisition_geometry.py

@@ -2095,7 +2095,8 @@ def __eq__(self, other):
 
         if isinstance(other, self.__class__) \
             and self.config == other.config \
-            and self.dtype == other.dtype:
+            and self.dtype == other.dtype \
+            and self.dimension_labels == other.dimension_labels:
             return True
         return False
 

Wrappers/Python/cil/framework/image_data.py

@@ -76,7 +76,32 @@ def __init__(self,
 
         super(ImageData, self).__init__(array, deep_copy, geometry=geometry, **kwargs)
 
+    def __eq__(self, other):
+        '''
+        Check if two ImageData objects are equal. This is done by checking if the geometry, data and dtype are equal.
+        Also, if the other object is a numpy.ndarray, it will check if the data and dtype are equal.
+        
+        Parameters
+        ----------
+        other: ImageData or numpy.ndarray
+            The object to compare with.
+        
+        Returns
+        -------
+        bool
+            True if the two objects are equal, False otherwise.
+        '''
 
+        if isinstance(other, ImageData):
+            if numpy.array_equal(self.as_array(), other.as_array()) \
+                and self.geometry == other.geometry \
+                and self.dtype == other.dtype:
+                return True 
+        elif numpy.array_equal(self.as_array(), other) and self.dtype==other.dtype:
+            return True
+        else:
+            return False
+
     def get_slice(self,channel=None, vertical=None, horizontal_x=None, horizontal_y=None, force=False):
         '''
         Returns a new ImageData of a single slice of in the requested direction.

Wrappers/Python/test/test_AcquisitionGeometry.py

@@ -495,10 +495,12 @@ def test_copy(self):
     def test_get_centre_slice(self):
         AG = AcquisitionGeometry.create_Parallel3D(detector_direction_y=[0,1,1])
         AG.set_panel([1000,2000],[1,1])
+        AG.set_angles([0,1,2,3,5])
         AG_cs = AG.get_centre_slice()
 
         AG2 = AcquisitionGeometry.create_Parallel2D()
         AG2.set_panel([1000,1],[1,math.sqrt(0.5)])
+        AG2.set_angles([0,1,2,3,5])
 
         self.assertEqual(AG2, AG_cs)
 

Wrappers/Python/test/test_DataContainer.py

@@ -38,20 +38,6 @@ def aid(x):
 
 
 class TestDataContainer(CCPiTestClass):
-    def create_simple_ImageData(self):
-        N = 64
-        ig = ImageGeometry(voxel_num_x=N, voxel_num_y=N)
-        Phantom = ImageData(geometry=ig)
-
-        x = Phantom.as_array()
-
-        x[int(round(N/4)):int(round(3*N/4)),
-          int(round(N/4)):int(round(3*N/4))] = 0.5
-        x[int(round(N/8)):int(round(7*N/8)),
-          int(round(3*N/8)):int(round(5*N/8))] = 1
-
-        return (ig, Phantom)
-
     def create_DataContainer(self, X,Y,Z, value=1):
         a = value * np.ones((X, Y, Z), dtype='float32')
         #print("a refcount " , sys.getrefcount(a))
@@ -88,6 +74,88 @@ def test_ndim(self):
         self.assertEqual(x_np.ndim, x_cil.ndim)
         self.assertEqual(3, x_cil.ndim)
 
+    def test_DataContainer_equal(self):
+        array = np.linspace(-1, 1, 32, dtype=np.float32).reshape(4, 8)
+        data = DataContainer(array)
+        data1 = data.copy()
+
+        # Check two identical DataContainers are equal
+        self.assertTrue((data == data1).all())
+        # Check it works for comparing DataContainer with numpy array
+        self.assertTrue((data == data1.as_array()).all())
+
+        # # Check two different DataContainers are not equal
+        data1 += 1
+        self.assertFalse((data ==  data1).all())
+
+    def test_AcquisitionData_equal(self):
+        array = np.linspace(-1, 1, 32, dtype=np.float32).reshape(4, 8)
+        geom = AcquisitionGeometry.create_Parallel3D().set_panel((8, 4)).set_channels(1).set_angles([1])
+        data = AcquisitionData(array, geometry=geom)
+
+        data1 = data.copy()
+
+        # Check two identical AcquisitionData are equal
+        self.assertTrue(data == data1)
+        # Check it works for comparing AcquisitionData with numpy array
+        self.assertTrue(data == data1.as_array())
+
+        # # Check two different AcquisitionData are not equal
+        data1 += 1
+        self.assertFalse(data ==  data1)
+
+        # Check the equality of two AcquisitionData with different shapes
+        data_different_shape = data.copy()
+        data_different_shape.array = data_different_shape.array.reshape(8, 4)
+
+        self.assertFalse(data == data_different_shape)
+
+        # Check the equality of two AcquisitionData with different dtypes
+        data_different_dtype = data.copy()
+        data_different_dtype.array = data_different_dtype.array.astype(np.float64)
+        self.assertFalse(data == data_different_dtype)
+
+
+        # Check the equality of two AcquisitionData with different labels
+        data_different_labels = data.copy()
+        print(data_different_labels.geometry.dimension_labels)
+        data_different_labels.geometry.set_labels([AcquisitionDimension("ANGLE"), AcquisitionDimension("CHANNEL") ])
+        self.assertFalse(data == data_different_labels)
+
+    def test_ImageData_equal(self):
+        array = np.linspace(-1, 1, 32, dtype=np.float32).reshape(4, 8)
+        geom = ImageGeometry(voxel_num_x=8, voxel_num_y=4)
+        data = ImageData(array, geometry=geom)
+
+        data1 = data.copy()
+
+        # Check two identical ImageData are equal
+        self.assertTrue(data == data1)
+        # Check it works for comparing ImageData with numpy array
+        self.assertTrue(data == data1.as_array())
+
+        # # Check two different ImageData are not equal
+        data1 += 1
+        self.assertFalse(data ==  data1)
+
+        # Check the equality of two ImageData with different shapes
+        data_different_shape = data.copy()
+        data_different_shape.array = data_different_shape.array.reshape(8, 4)
+
+        self.assertFalse(data == data_different_shape)
+
+        # Check the equality of two ImageData with different dtypes
+        data_different_dtype = data.copy()
+        data_different_dtype.array = data_different_dtype.array.astype(np.float64)
+        self.assertFalse(data == data_different_dtype)
+
+
+        # Check the equality of two ImageData with different labels
+        data_different_labels = data.copy()
+        data_different_labels.geometry.set_labels([ImageDimension("VERTICAL"), ImageDimension("HORIZONTAL_X")])
+        self.assertFalse(data == data_different_labels)
+
+
     def testInlineAlgebra(self):
         X, Y, Z = 8, 16, 32
         a = np.ones((X, Y, Z), dtype='float32')