def test_instantiate_SpatialSVM():
from os.path import dirname, join, abspath
from clinica.pipelines.machine_learning_spatial_svm.spatial_svm_pipeline import SpatialSVM
root = dirname(abspath(join(abspath(__file__), pardir)))
root = join(root, 'data', 'SpatialSVM')
parameters = {'group_label': 'ADNIbl', 'orig_input_data': 't1-volume'}
pipeline = SpatialSVM(caps_directory=join(root, 'in', 'caps'),
tsv_file=join(root, 'in', 'subjects.tsv'),
parameters=parameters)
pipeline.build()
def run_command(self, args):
"""
"""
from tempfile import mkdtemp
from clinica.utils.stream import cprint
from clinica.pipelines.machine_learning_spatial_svm.spatial_svm_pipeline import SpatialSVM
pipeline = SpatialSVM(
caps_directory=self.absolute_path(args.caps_directory),
tsv_file=self.absolute_path(args.subjects_sessions_tsv))
pipeline.parameters = {
'group_id': args.group_id,
'fwhm': args.full_width_half_maximum,
'image_type': args.image_type,
'pet_type': args.pet_tracer,
'no_pvc': args.no_pvc
}
if args.working_directory is None:
args.working_directory = mkdtemp()
pipeline.base_dir = self.absolute_path(args.working_directory)
if args.n_procs:
pipeline.run(plugin='MultiProc',
plugin_args={'n_procs': args.n_procs})
else:
pipeline.run()
cprint(
"The " + self._name +
" pipeline has completed. You can now delete the working directory ("
+ args.working_directory + ").")
def test_instantiate_SpatialSVM(cmdopt):
from clinica.pipelines.machine_learning_spatial_svm.spatial_svm_pipeline import (
SpatialSVM, )
input_dir = Path(cmdopt["input"])
root = input_dir / "SpatialSVM"
parameters = {"group_label": "ADNIbl", "orig_input_data": "t1-volume"}
pipeline = SpatialSVM(
caps_directory=fspath(root / "in" / "caps"),
tsv_file=fspath(root / "in" / "subjects.tsv"),
parameters=parameters,
)
pipeline.build()
def test_instantiate_SpatialSVM():
from os.path import abspath, dirname, join
from clinica.pipelines.machine_learning_spatial_svm.spatial_svm_pipeline import (
SpatialSVM, )
root = dirname(abspath(join(abspath(__file__), pardir)))
root = join(root, "data", "SpatialSVM")
parameters = {"group_label": "ADNIbl", "orig_input_data": "t1-volume"}
pipeline = SpatialSVM(
caps_directory=join(root, "in", "caps"),
tsv_file=join(root, "in", "subjects.tsv"),
parameters=parameters,
)
pipeline.build()
def test_instantiate_SpatialSVM():
from clinica.pipelines.machine_learning_spatial_svm.spatial_svm_pipeline import SpatialSVM
from os.path import dirname, join, abspath
root = join(dirname(abspath(__file__)), 'data', 'SpatialSVM')
pipeline = SpatialSVM(caps_directory=join(root, 'in', 'caps'),
tsv_file=join(root, 'in', 'subjects.tsv'))
pipeline.parameters['group_id'] = 'ADNIbl'
pipeline.parameters['fwhm'] = 4
pipeline.parameters['image_type'] = 't1'
pipeline.parameters['pet_type'] = 'fdg'
pipeline.parameters['no_pvc'] = 'True'
pipeline.build()
def test_run_SpatialSVM(cmdopt):
from os.path import dirname, join, abspath
import shutil
import numpy as np
import nibabel as nib
from clinica.pipelines.machine_learning_spatial_svm.spatial_svm_pipeline import SpatialSVM
working_dir = cmdopt
root = dirname(abspath(join(abspath(__file__), pardir)))
root = join(root, 'data', 'SpatialSVM')
# Remove potential residual of previous UT
clean_folder(join(root, 'out', 'caps'), recreate=False)
clean_folder(join(working_dir, 'SpatialSVM'), recreate=False)
# Copy necessary data from in to out
shutil.copytree(join(root, 'in', 'caps'), join(root, 'out', 'caps'))
parameters = {'group_label': 'ADNIbl', 'orig_input_data': 't1-volume'}
# Instantiate pipeline and run()
pipeline = SpatialSVM(caps_directory=join(root, 'out', 'caps'),
tsv_file=join(root, 'in', 'subjects.tsv'),
base_dir=join(working_dir, 'SpatialSVM'),
parameters=parameters)
pipeline.build()
pipeline.run(plugin='MultiProc',
plugin_args={'n_procs': 4},
bypass_check=True)
# Check output vs ref
subjects = ['sub-ADNI011S0023', 'sub-ADNI013S0325']
out_data_REG_NIFTI = [
nib.load(
join(
root, 'out', 'caps', 'subjects', sub, 'ses-M00',
'machine_learning', 'input_spatial_svm', 'group-ADNIbl', sub +
'_ses-M00_T1w_segm-graymatter_space-Ixi549Space_modulated-on_spatialregularization.nii.gz'
)).get_data() for sub in subjects
]
ref_data_REG_NIFTI = [
nib.load(
join(
root, 'ref', sub +
'_ses-M00_T1w_segm-graymatter_space-Ixi549Space_modulated-on_spatialregularization.nii.gz'
)).get_data() for sub in subjects
]
for i in range(len(out_data_REG_NIFTI)):
assert np.allclose(out_data_REG_NIFTI[i],
ref_data_REG_NIFTI[i],
rtol=1e-3,
equal_nan=True)
# Remove data in out folder
clean_folder(join(root, 'out', 'caps'), recreate=True)
clean_folder(join(working_dir, 'SpatialSVM'), recreate=False)
def run_SpatialSVM(input_dir: Path, output_dir: Path, ref_dir: Path,
working_dir: Path) -> None:
import shutil
from os import fspath
import nibabel as nib
import numpy as np
from clinica.pipelines.machine_learning_spatial_svm.spatial_svm_pipeline import (
SpatialSVM, )
caps_dir = output_dir / "caps"
tsv = input_dir / "subjects.tsv"
# Copy necessary data from in to out
shutil.copytree(input_dir / "caps", caps_dir, copy_function=shutil.copy)
parameters = {"group_label": "ADNIbl", "orig_input_data": "t1-volume"}
# Instantiate pipeline and run()
pipeline = SpatialSVM(
caps_directory=fspath(caps_dir),
tsv_file=fspath(tsv),
base_dir=fspath(working_dir),
parameters=parameters,
)
pipeline.build()
pipeline.run(plugin="MultiProc",
plugin_args={"n_procs": 4},
bypass_check=True)
# Check output vs ref
subjects = ["sub-ADNI011S0023", "sub-ADNI013S0325"]
out_data_REG_NIFTI = [
nib.load(
fspath(
caps_dir / "subjects" / sub / "ses-M00" / "machine_learning" /
"input_spatial_svm" / "group-ADNIbl" /
(sub +
"_ses-M00_T1w_segm-graymatter_space-Ixi549Space_modulated-on_spatialregularization.nii.gz"
))).get_fdata(dtype="float32") for sub in subjects
]
ref_data_REG_NIFTI = [
nib.load(
fspath(ref_dir / (
sub +
"_ses-M00_T1w_segm-graymatter_space-Ixi549Space_modulated-on_spatialregularization.nii.gz"
))).get_fdata(dtype="float32") for sub in subjects
]
for i in range(len(out_data_REG_NIFTI)):
assert np.allclose(out_data_REG_NIFTI[i],
ref_data_REG_NIFTI[i],
rtol=1e-3,
equal_nan=True)
def test_run_SpatialSVM(cmdopt):
import shutil
from os.path import abspath, dirname, join
import nibabel as nib
import numpy as np
from clinica.pipelines.machine_learning_spatial_svm.spatial_svm_pipeline import (
SpatialSVM, )
working_dir = cmdopt
root = dirname(abspath(join(abspath(__file__), pardir, pardir)))
root = join(root, "data", "SpatialSVM")
# Remove potential residual of previous UT
clean_folder(join(root, "out", "caps"), recreate=False)
clean_folder(join(working_dir, "SpatialSVM"), recreate=False)
# Copy necessary data from in to out
shutil.copytree(join(root, "in", "caps"), join(root, "out", "caps"))
parameters = {"group_label": "ADNIbl", "orig_input_data": "t1-volume"}
# Instantiate pipeline and run()
pipeline = SpatialSVM(
caps_directory=join(root, "out", "caps"),
tsv_file=join(root, "in", "subjects.tsv"),
base_dir=join(working_dir, "SpatialSVM"),
parameters=parameters,
)
pipeline.build()
pipeline.run(plugin="MultiProc",
plugin_args={"n_procs": 4},
bypass_check=True)
# Check output vs ref
subjects = ["sub-ADNI011S0023", "sub-ADNI013S0325"]
out_data_REG_NIFTI = [
nib.load(
join(
root,
"out",
"caps",
"subjects",
sub,
"ses-M00",
"machine_learning",
"input_spatial_svm",
"group-ADNIbl",
sub +
"_ses-M00_T1w_segm-graymatter_space-Ixi549Space_modulated-on_spatialregularization.nii.gz",
)).get_data() for sub in subjects
]
ref_data_REG_NIFTI = [
nib.load(
join(
root,
"ref",
sub +
"_ses-M00_T1w_segm-graymatter_space-Ixi549Space_modulated-on_spatialregularization.nii.gz",
)).get_data() for sub in subjects
]
for i in range(len(out_data_REG_NIFTI)):
assert np.allclose(out_data_REG_NIFTI[i],
ref_data_REG_NIFTI[i],
rtol=1e-3,
equal_nan=True)
# Remove data in out folder
clean_folder(join(root, "out", "caps"), recreate=True)
clean_folder(join(working_dir, "SpatialSVM"), recreate=False)
