torchio.data.io._write_nibabel() - Code Metrics - Inspection of "Refactor 2D support (#270)" - fepegar/torchio - Measure and Improve Code Quality continuously with Scrutinizer

Passed

Push — master ( 53ab14...c2608f )

by Fernando

created 2020-08-17 20:28 UTC

torchio.data.io._write_nibabel() B

↳ Parent: torchio.data.io

Complexity

Conditions

Size

Total Lines	25
Code Lines	19

Duplication

Lines	0
Ratio	0 %

Importance

Changes

Metric	Value
cc	6
eloc	19
nop	5
dl	0
loc	25
rs	8.5166
c	0
b	0
f	0

import warnings
from pathlib import Path
from typing import Tuple
import torch
import numpy as np
import nibabel as nib
import SimpleITK as sitk
from .. import TypePath, TypeData
from ..utils import nib_to_sitk, sitk_to_nib


FLIPXY = np.diag([-1, -1, 1, 1])


def read_image(
        path: TypePath,
        itk_first: bool = False,
        ) -> Tuple[torch.Tensor, np.ndarray]:
    if itk_first:
        try:
            result = _read_sitk(path)
        except RuntimeError:  # try with NiBabel
            result = _read_nibabel(path)
    else:
        try:
            result = _read_nibabel(path)
        except nib.loadsave.ImageFileError:  # try with ITK
            result = _read_sitk(path)
    return result


def _read_nibabel(path: TypePath) -> Tuple[torch.Tensor, np.ndarray]:
    img = nib.load(str(path), mmap=False)
    data = img.get_fdata(dtype=np.float32)
    tensor = torch.from_numpy(data)
    affine = img.affine
    return tensor, affine


def _read_sitk(
        path: TypePath,
        transpose_2d: bool = True,
        ) -> Tuple[torch.Tensor, np.ndarray]:
    if Path(path).is_dir():  # assume DICOM
        image = _read_dicom(path)
    else:
        image = sitk.ReadImage(str(path))
    if image.GetDimension() == 2 and transpose_2d:
        image = sitk.PermuteAxes(image, (1, 0))
    data, affine = sitk_to_nib(image, keepdim=True)
    if data.dtype != np.float32:
        data = data.astype(np.float32)
    tensor = torch.from_numpy(data)
    return tensor, affine


def _read_dicom(directory: TypePath):
    directory = Path(directory)
    if not directory.is_dir():  # unreachable if called from _read_sitk
        raise FileNotFoundError(f'Directory "{directory}" not found')
    reader = sitk.ImageSeriesReader()
    dicom_names = reader.GetGDCMSeriesFileNames(str(directory))
    if not dicom_names:
        message = (
            f'The directory "{directory}"'
            ' does not seem to contain DICOM files'
        )
        raise FileNotFoundError(message)
    reader.SetFileNames(dicom_names)
    image = reader.Execute()
    return image


def write_image(
        tensor: torch.Tensor,
        affine: TypeData,
        path: TypePath,
        itk_first: bool = False,
        squeeze: bool = True,
        channels_last: bool = True,
        ) -> None:
    args = tensor, affine, path
    kwargs = dict(squeeze=squeeze, channels_last=channels_last)
    if itk_first:
        try:
            _write_sitk(*args, squeeze=squeeze)
        except RuntimeError:  # try with NiBabel
            _write_nibabel(*args, squeeze=squeeze, channels_last=channels_last)
    else:
        try:
            _write_nibabel(*args, squeeze=squeeze, channels_last=channels_last)
        except nib.loadsave.ImageFileError:  # try with ITK
            _write_sitk(*args, squeeze=squeeze)


def _write_nibabel(
        tensor: TypeData,
        affine: TypeData,
        path: TypePath,
        squeeze: bool = True,
        channels_last: bool = True,
        ) -> None:
    """
    Expects a path with an extension that can be used by nibabel.save
    to write a NIfTI-1 image, such as '.nii.gz' or '.img'
    """
    assert tensor.ndim == 4
    if channels_last:
        tensor = tensor.permute(1, 2, 3, 0)
    tensor = tensor.squeeze() if squeeze else tensor
    suffix = Path(str(path).replace('.gz', '')).suffix
    if '.nii' in suffix:
        img = nib.Nifti1Image(np.asarray(tensor), affine)
    elif '.hdr' in suffix or '.img' in suffix:
        img = nib.Nifti1Pair(np.asarray(tensor), affine)
    else:
        raise nib.loadsave.ImageFileError
    img.header['qform_code'] = 1
    img.header['sform_code'] = 0
    img.to_filename(str(path))


def _write_sitk(
        tensor: torch.Tensor,
        affine: TypeData,
        path: TypePath,
        squeeze: bool = True,
        use_compression: bool = True,
        transpose_2d: bool = True,
        ) -> None:
    assert tensor.ndim == 4
    path = Path(path)
    if path.suffix in ('.png', '.jpg', '.jpeg'):
        warnings.warn(f'Casting to uint 8 before saving to {path}')
        tensor = tensor.numpy().astype(np.uint8)
    image = nib_to_sitk(tensor, affine, squeeze=squeeze)
    if image.GetDimension() == 2 and transpose_2d:
        image = sitk.PermuteAxes(image, (1, 0))
    sitk.WriteImage(image, str(path), use_compression)


def read_matrix(path: TypePath):
    """Read an affine transform and convert to tensor."""
    path = Path(path)
    suffix = path.suffix
    if suffix in ('.tfm', '.h5'):  # ITK
        tensor = _read_itk_matrix(path)
    elif suffix in ('.txt', '.trsf'):  # NiftyReg, blockmatching
        tensor = _read_niftyreg_matrix(path)
    return tensor



def write_matrix(matrix: torch.Tensor, path: TypePath):
    """Write an affine transform."""
    path = Path(path)
    suffix = path.suffix
    if suffix in ('.tfm', '.h5'):  # ITK
        _write_itk_matrix(matrix, path)
    elif suffix in ('.txt', '.trsf'):  # NiftyReg, blockmatching
        _write_niftyreg_matrix(matrix, path)


def _to_itk_convention(matrix):
    """RAS to LPS"""
    matrix = np.dot(FLIPXY, matrix)
    matrix = np.dot(matrix, FLIPXY)
    matrix = np.linalg.inv(matrix)
    return matrix


def _from_itk_convention(matrix):
    """LPS to RAS"""
    matrix = np.dot(matrix, FLIPXY)
    matrix = np.dot(FLIPXY, matrix)
    matrix = np.linalg.inv(matrix)
    return matrix


def _read_itk_matrix(path):
    """Read an affine transform in ITK's .tfm format"""
    transform = sitk.ReadTransform(str(path))
    parameters = transform.GetParameters()
    rotation_parameters = parameters[:9]
    rotation_matrix = np.array(rotation_parameters).reshape(3, 3)
    translation_parameters = parameters[9:]
    translation_vector = np.array(translation_parameters).reshape(3, 1)
    matrix = np.hstack([rotation_matrix, translation_vector])
    homogeneous_matrix_lps = np.vstack([matrix, [0, 0, 0, 1]])
    homogeneous_matrix_ras = _from_itk_convention(homogeneous_matrix_lps)
    return torch.from_numpy(homogeneous_matrix_ras)


def _write_itk_matrix(matrix, tfm_path):
    """The tfm file contains the matrix from floating to reference."""
    transform = _matrix_to_itk_transform(matrix)
    transform.WriteTransform(str(tfm_path))


def _matrix_to_itk_transform(matrix, dimensions=3):
    matrix = _to_itk_convention(matrix)
    rotation = matrix[:dimensions, :dimensions].ravel().tolist()
    translation = matrix[:dimensions, 3].tolist()
    transform = sitk.AffineTransform(rotation, translation)
    return transform


def _read_niftyreg_matrix(trsf_path):
    """Read a NiftyReg matrix and return it as a NumPy array"""
    matrix = np.loadtxt(trsf_path)
    matrix = np.linalg.inv(matrix)
    return torch.from_numpy(matrix)


def _write_niftyreg_matrix(matrix, txt_path):
    """Write an affine transform in NiftyReg's .txt format (ref -> flo)"""
    matrix = np.linalg.inv(matrix)
    np.savetxt(txt_path, matrix, fmt='%.8f')


1			import warnings
2			from pathlib import Path
3			from typing import Tuple
4			import torch
5			import numpy as np
6			import nibabel as nib
7			import SimpleITK as sitk
8			from .. import TypePath, TypeData
9			from ..utils import nib_to_sitk, sitk_to_nib
10
11
12			FLIPXY = np.diag([-1, -1, 1, 1])
13
14
15			def read_image(
16			path: TypePath,
17			itk_first: bool = False,
18			) -> Tuple[torch.Tensor, np.ndarray]:
19			if itk_first:
20			try:
21			result = _read_sitk(path)
22			except RuntimeError: # try with NiBabel
23			result = _read_nibabel(path)
24			else:
25			try:
26			result = _read_nibabel(path)
27			except nib.loadsave.ImageFileError: # try with ITK
28			result = _read_sitk(path)
29			return result
30
31
32			def _read_nibabel(path: TypePath) -> Tuple[torch.Tensor, np.ndarray]:
33			img = nib.load(str(path), mmap=False)
34			data = img.get_fdata(dtype=np.float32)
35			tensor = torch.from_numpy(data)
36			affine = img.affine
37			return tensor, affine
38
39
40			def _read_sitk(
41			path: TypePath,
42			transpose_2d: bool = True,
43			) -> Tuple[torch.Tensor, np.ndarray]:
44			if Path(path).is_dir(): # assume DICOM
45			image = _read_dicom(path)
46			else:
47			image = sitk.ReadImage(str(path))
48			if image.GetDimension() == 2 and transpose_2d:
49			image = sitk.PermuteAxes(image, (1, 0))
50			data, affine = sitk_to_nib(image, keepdim=True)
51			if data.dtype != np.float32:
52			data = data.astype(np.float32)
53			tensor = torch.from_numpy(data)
54			return tensor, affine
55
56
57			def _read_dicom(directory: TypePath):
58			directory = Path(directory)
59			if not directory.is_dir(): # unreachable if called from _read_sitk
60			raise FileNotFoundError(f'Directory "{directory}" not found')
61			reader = sitk.ImageSeriesReader()
62			dicom_names = reader.GetGDCMSeriesFileNames(str(directory))
63			if not dicom_names:
64			message = (
65			f'The directory "{directory}"'
66			' does not seem to contain DICOM files'
67			)
68			raise FileNotFoundError(message)
69			reader.SetFileNames(dicom_names)
70			image = reader.Execute()
71			return image
72
73
74			def write_image(
75			tensor: torch.Tensor,
76			affine: TypeData,
77			path: TypePath,
78			itk_first: bool = False,
79			squeeze: bool = True,
80			channels_last: bool = True,
81			) -> None:
82			args = tensor, affine, path
83			kwargs = dict(squeeze=squeeze, channels_last=channels_last)
84			if itk_first:
85			try:
86			_write_sitk(*args, squeeze=squeeze)
87			except RuntimeError: # try with NiBabel
88			_write_nibabel(*args, squeeze=squeeze, channels_last=channels_last)
89			else:
90			try:
91			_write_nibabel(*args, squeeze=squeeze, channels_last=channels_last)
92			except nib.loadsave.ImageFileError: # try with ITK
93			_write_sitk(*args, squeeze=squeeze)
94
95
96			def _write_nibabel(
97			tensor: TypeData,
98			affine: TypeData,
99			path: TypePath,
100			squeeze: bool = True,
101			channels_last: bool = True,
102			) -> None:
103			"""
104			Expects a path with an extension that can be used by nibabel.save
105			to write a NIfTI-1 image, such as '.nii.gz' or '.img'
106			"""
107			assert tensor.ndim == 4
108			if channels_last:
109			tensor = tensor.permute(1, 2, 3, 0)
110			tensor = tensor.squeeze() if squeeze else tensor
111			suffix = Path(str(path).replace('.gz', '')).suffix
112			if '.nii' in suffix:
113			img = nib.Nifti1Image(np.asarray(tensor), affine)
114			elif '.hdr' in suffix or '.img' in suffix:
115			img = nib.Nifti1Pair(np.asarray(tensor), affine)
116			else:
117			raise nib.loadsave.ImageFileError
118			img.header['qform_code'] = 1
119			img.header['sform_code'] = 0
120			img.to_filename(str(path))
121
122
123			def _write_sitk(
124			tensor: torch.Tensor,
125			affine: TypeData,
126			path: TypePath,
127			squeeze: bool = True,
128			use_compression: bool = True,
129			transpose_2d: bool = True,
130			) -> None:
131			assert tensor.ndim == 4
132			path = Path(path)
133			if path.suffix in ('.png', '.jpg', '.jpeg'):
134			warnings.warn(f'Casting to uint 8 before saving to {path}')
135			tensor = tensor.numpy().astype(np.uint8)
136			image = nib_to_sitk(tensor, affine, squeeze=squeeze)
137			if image.GetDimension() == 2 and transpose_2d:
138			image = sitk.PermuteAxes(image, (1, 0))
139			sitk.WriteImage(image, str(path), use_compression)
140
141
142			def read_matrix(path: TypePath):
143			"""Read an affine transform and convert to tensor."""
144			path = Path(path)
145			suffix = path.suffix
146			if suffix in ('.tfm', '.h5'): # ITK
147			tensor = _read_itk_matrix(path)
148			elif suffix in ('.txt', '.trsf'): # NiftyReg, blockmatching
149			tensor = _read_niftyreg_matrix(path)
150			return tensor
			0 ignored issues – show introduced 2020-06-30 09:27 UTC by Report Bug Copy Issue Report The variable `tensor` does not seem to be defined for all execution paths. Loading history...
151
152
153			def write_matrix(matrix: torch.Tensor, path: TypePath):
154			"""Write an affine transform."""
155			path = Path(path)
156			suffix = path.suffix
157			if suffix in ('.tfm', '.h5'): # ITK
158			_write_itk_matrix(matrix, path)
159			elif suffix in ('.txt', '.trsf'): # NiftyReg, blockmatching
160			_write_niftyreg_matrix(matrix, path)
161
162
163			def _to_itk_convention(matrix):
164			"""RAS to LPS"""
165			matrix = np.dot(FLIPXY, matrix)
166			matrix = np.dot(matrix, FLIPXY)
167			matrix = np.linalg.inv(matrix)
168			return matrix
169
170
171			def _from_itk_convention(matrix):
172			"""LPS to RAS"""
173			matrix = np.dot(matrix, FLIPXY)
174			matrix = np.dot(FLIPXY, matrix)
175			matrix = np.linalg.inv(matrix)
176			return matrix
177
178
179			def _read_itk_matrix(path):
180			"""Read an affine transform in ITK's .tfm format"""
181			transform = sitk.ReadTransform(str(path))
182			parameters = transform.GetParameters()
183			rotation_parameters = parameters[:9]
184			rotation_matrix = np.array(rotation_parameters).reshape(3, 3)
185			translation_parameters = parameters[9:]
186			translation_vector = np.array(translation_parameters).reshape(3, 1)
187			matrix = np.hstack([rotation_matrix, translation_vector])
188			homogeneous_matrix_lps = np.vstack([matrix, [0, 0, 0, 1]])
189			homogeneous_matrix_ras = _from_itk_convention(homogeneous_matrix_lps)
190			return torch.from_numpy(homogeneous_matrix_ras)
191
192
193			def _write_itk_matrix(matrix, tfm_path):
194			"""The tfm file contains the matrix from floating to reference."""
195			transform = _matrix_to_itk_transform(matrix)
196			transform.WriteTransform(str(tfm_path))
197
198
199			def _matrix_to_itk_transform(matrix, dimensions=3):
200			matrix = _to_itk_convention(matrix)
201			rotation = matrix[:dimensions, :dimensions].ravel().tolist()
202			translation = matrix[:dimensions, 3].tolist()
203			transform = sitk.AffineTransform(rotation, translation)
204			return transform
205
206
207			def _read_niftyreg_matrix(trsf_path):
208			"""Read a NiftyReg matrix and return it as a NumPy array"""
209			matrix = np.loadtxt(trsf_path)
210			matrix = np.linalg.inv(matrix)
211			return torch.from_numpy(matrix)
212
213
214			def _write_niftyreg_matrix(matrix, txt_path):
215			"""Write an affine transform in NiftyReg's .txt format (ref -> flo)"""
216			matrix = np.linalg.inv(matrix)
217			np.savetxt(txt_path, matrix, fmt='%.8f')
218

fepegar / torchio

Push — master ( 53ab14...c2608f )

torchio.data.io._write_nibabel() B

Complexity

Size

Duplication

Importance

Duplication Side-by-Side

Filter issues like