Apply suggestions from code review [skip ci]

Co-Authored-By: Chris Markiewicz <[email protected]>

Author: oesteban

niworkflows/interfaces/freesurfer.py

@@ -374,9 +374,9 @@ def inject_skullstripped(subjects_dir, subject_id, skullstripped):
     if not op.exists(bm_auto):
         img = nb.load(t1)
         mask = nb.load(skullstripped)
-        bmask = new_img_like(mask, mask.get_fdata() > 0)
+        bmask = new_img_like(mask, np.asanyarray(mask.dataobj) > 0)
         resampled_mask = resample_to_img(bmask, img, 'nearest')
-        masked_image = new_img_like(img, img.get_fdata() * resampled_mask.get_fdata())
+        masked_image = new_img_like(img, img.get_fdata() * resampled_mask.dataobj)
         masked_image.to_filename(bm_auto)
 
     if not op.exists(bm):

niworkflows/interfaces/images.py

@@ -65,7 +65,7 @@ def _run_interface(self, runtime):
                     nb.Nifti1Image(sqdata, filenii.affine,
                                    filenii.header).to_filename(in_files[0])
 
-            if np.squeeze(nb.load(in_files[0]).get_fdata()).ndim < 4:
+            if filenii.dataobj.ndim < 4:
                 self._results['out_file'] = in_files[0]
                 self._results['out_avg'] = in_files[0]
                 # TODO: generate identity out_mats and zero-filled out_movpar
@@ -265,7 +265,7 @@ def _run_interface(self, runtime):
                 offset = (reoriented.affine[:3, 3] * size_factor - reoriented.affine[:3, 3])
                 target_affine[:3, 3] = reoriented.affine[:3, 3] + offset.astype(int)
 
-            data = nli.resample_img(reoriented, target_affine, target_shape).get_fdata()
+            data = nli.resample_img(reoriented, target_affine, target_shape).dataobj
             conform_xfm = np.linalg.inv(reoriented.affine).dot(target_affine)
             reoriented = reoriented.__class__(data, target_affine, reoriented.header)
 
@@ -534,7 +534,7 @@ def _run_interface(self, runtime):
         out_file = fname_presuffix(self.inputs.in_file, suffix='_hdr',
                                    newpath=runtime.cwd)
 
-        imgnii.__class__(imgnii.get_fdata(), imghdr.get_best_affine(),
+        imgnii.__class__(imgnii.dataobj, imghdr.get_best_affine(),
                          imghdr).to_filename(out_file)
         self._results['out_file'] = out_file
         return runtime
@@ -554,7 +554,7 @@ def extract_wm(in_seg, wm_label=3, newpath=None):
 
     nii = nb.load(in_seg)
     data = np.zeros(nii.shape, dtype=np.uint8)
-    data[nii.get_fdata() == wm_label] = 1
+    data[np.asanyarray(nii.dataobj) == wm_label] = 1
 
     out_file = fname_presuffix(in_seg, suffix='_wm', newpath=newpath)
     new = nb.Nifti1Image(data, nii.affine, nii.header)
@@ -591,7 +591,7 @@ def normalize_xform(img):
             int(qform_code) == xform_code, int(sform_code) == xform_code)):
         return img
 
-    new_img = img.__class__(img.get_fdata(), xform, img.header)
+    new_img = img.__class__(img.dataobj, xform, img.header)
     # Unconditionally set sform/qform
     new_img.set_sform(xform, xform_code)
     new_img.set_qform(xform, xform_code)
@@ -608,7 +608,7 @@ def demean(in_file, in_mask, only_mask=False, newpath=None):
     out_file = fname_presuffix(in_file, suffix='_demeaned',
                                newpath=os.getcwd())
     nii = nb.load(in_file)
-    msk = nb.load(in_mask).get_fdata()
+    msk = np.asanyarray(nb.load(in_mask).dataobj)
     data = nii.get_fdata()
     if only_mask:
         data[msk > 0] -= np.median(data[msk > 0])
@@ -706,10 +706,10 @@ def _run_interface(self, runtime):
         # If mask is a list, each mask is treated as its own ROI/parcel
         # If mask is a 3D, each integer is treated as its own ROI/parcel
         if len(mask_imgs) > 1:
-            masks = [mask_img.get_fdata() >= self.inputs.prob_thres
+            masks = [np.asanyarray(mask_img.dataobj) >= self.inputs.prob_thres
                      for mask_img in mask_imgs]
         else:
-            labelsmap = mask_imgs[0].get_fdata()
+            labelsmap = np.asanyarray(mask_imgs[0].dataobj)
             labels = np.unique(labelsmap)
             labels = labels[labels != 0]
             masks = [labelsmap == l for l in labels]

niworkflows/interfaces/masks.py

@@ -113,7 +113,7 @@ def _run_interface(self, runtime):
 
         mask_nii = compute_epi_mask(padded_nii, exclude_zeros=True)
 
-        mask_data = mask_nii.get_fdata()
+        mask_data = np.asanyarray(mask_nii.dataobj).astype(np.uint8)
         if isdefined(self.inputs.dilation):
             mask_data = nd.morphology.binary_dilation(mask_data).astype(np.uint8)
 
@@ -267,7 +267,7 @@ def _generate_report(self):
             nsegs = len(levels)
             if nsegs == 0:
                 levels = np.unique(np.round(
-                    nb.load(seg_files[0]).get_fdata()).astype(int))
+                    np.asanyarray(nb.load(seg_files[0]).dataobj).astype('int32'))
                 levels = (levels[levels > 0] - 0.5).tolist()
                 nsegs = len(levels)
 

niworkflows/interfaces/mni.py

@@ -418,7 +418,7 @@ def mask(in_file, mask_file, new_name):
     mask_nii = nb.load(mask_file)
     # Set all non-mask voxels in the input file to zero.
     data = in_nii.get_fdata()
-    data[mask_nii.get_fdata() == 0] = 0
+    data[np.asanyarray(mask_nii.dataobj) == 0] = 0
     # Save the new masked image.
     new_nii = nb.Nifti1Image(data, in_nii.affine, in_nii.header)
     new_nii.to_filename(new_name)
@@ -472,7 +472,7 @@ def create_cfm(in_file, lesion_mask=None, global_mask=True, out_path=None):
     in_img = nb.load(in_file)
 
     # If we want a global mask, create one based on the input image.
-    data = np.ones(in_img.shape, dtype=np.uint8) if global_mask else in_img.get_fdata()
+    data = np.ones(in_img.shape, dtype=np.uint8) if global_mask else np.asanyarray(in_img.dataobj)
     if set(np.unique(data)) - {0, 1}:
         raise ValueError("`global_mask` must be true if `in_file` is not a binary mask")
 
@@ -482,7 +482,7 @@ def create_cfm(in_file, lesion_mask=None, global_mask=True, out_path=None):
         lm_img = nb.as_closest_canonical(nb.load(lesion_mask))
 
         # Subtract lesion mask from secondary mask, set negatives to 0
-        data = np.fmax(data - lm_img.get_fdata(), 0)
+        data = np.fmax(data - lm_img.dataobj, 0)
         # Cost function mask will be created from subtraction
     # Otherwise, CFM will be created from global mask
 

niworkflows/interfaces/nilearn.py

@@ -72,13 +72,13 @@ def _run_interface(self, runtime):
         )
 
         if self.inputs.closing:
-            closed = sim.binary_closing(masknii.get_fdata().astype(
+            closed = sim.binary_closing(np.asanyarray(masknii.dataobj).astype(
                 np.uint8), sim.ball(1)).astype(np.uint8)
             masknii = masknii.__class__(closed, masknii.affine,
                                         masknii.header)
 
         if self.inputs.fill_holes:
-            filled = binary_fill_holes(masknii.get_fdata().astype(
+            filled = binary_fill_holes(np.asanyarray(masknii.dataobj).astype(
                 np.uint8), sim.ball(6)).astype(np.uint8)
             masknii = masknii.__class__(filled, masknii.affine,
                                         masknii.header)

niworkflows/interfaces/utils.py

@@ -106,7 +106,7 @@ class CopyHeader(SimpleInterface):
     def _run_interface(self, runtime):
         in_img = nb.load(self.inputs.hdr_file)
         out_img = nb.load(self.inputs.in_file)
-        new_img = out_img.__class__(out_img.get_fdata(), in_img.affine, in_img.header)
+        new_img = out_img.__class__(out_img.dataobj, in_img.affine, in_img.header)
         new_img.set_data_dtype(out_img.get_data_dtype())
 
         out_name = fname_presuffix(self.inputs.in_file,
@@ -215,7 +215,7 @@ def _copyxform(ref_image, out_image, message=None):
     header.set_sform(sform, int(sform_code))
     header['descrip'] = 'xform matrices modified by %s.' % (message or '(unknown)')
 
-    newimg = resampled.__class__(resampled.get_fdata(), orig.affine, header)
+    newimg = resampled.__class__(resampled.dataobj, orig.affine, header)
     newimg.to_filename(out_image)
 
 
@@ -266,7 +266,7 @@ def _gen_reference(fixed_image, moving_image, fov_mask=None, out_file=None,
 
         # Calculate a bounding box for the input mask
         # with an offset of 2 voxels per face
-        bbox = np.argwhere(masknii.get_fdata() > 0)
+        bbox = np.argwhere(np.asanyarray(masknii.dataobj) > 0)
         new_origin = np.clip(bbox.min(0) - 2, a_min=0, a_max=None)
         new_end = np.clip(bbox.max(0) + 2, a_min=0,
                           a_max=res_shape - 1)
@@ -552,7 +552,7 @@ def _run_interface(self, runtime):
             return runtime
 
         im = nb.concat_images([in_files[i] for i in indices])
-        data = im.get_fdata().astype(float).sum(axis=3)
+        data = im.get_fdata().sum(axis=3)
         data = np.clip(data, a_min=0.0, a_max=1.0)
 
         out_file = fname_presuffix(first_fname, suffix='_tpmsum',
@@ -911,7 +911,7 @@ def _tpm2roi(in_tpm, in_mask, mask_erosion_mm=None, erosion_mm=None,
         eroded_mask_file = fname_presuffix(in_mask, suffix='_eroded',
                                            newpath=newpath)
         mask_img = nb.load(in_mask)
-        mask_data = mask_img.get_fdata().astype(np.uint8)
+        mask_data = np.asanyarray(mask_img.dataobj).astype(np.uint8)
         if mask_erosion_mm:
             iter_n = max(int(mask_erosion_mm / max(mask_img.header.get_zooms())), 1)
             mask_data = nd.binary_erosion(mask_data, iterations=iter_n)

niworkflows/tests/test_viz.py

@@ -15,9 +15,9 @@ def test_carpetplot():
         out_file = os.path.join(save_artifacts, 'carpetplot.svg')
     viz.plot_carpet(
         os.path.join(datadir, 'sub-ds205s03_task-functionallocalizer_run-01_bold_volreg.nii.gz'),
-        nb.load(os.path.join(
+        np.asanyarray(nb.load(os.path.join(
             datadir,
-            'sub-ds205s03_task-functionallocalizer_run-01_bold_parc.nii.gz')).get_fdata(),
+            'sub-ds205s03_task-functionallocalizer_run-01_bold_parc.nii.gz')).dataobj),
         output_file=out_file,
         legend=True
     )

niworkflows/viz/plots.py

@@ -37,13 +37,13 @@ def __init__(self, func_file, mask_file=None, data=None, conf_file=None, seg_fil
         func_nii = nb.load(func_file)
         self.tr = tr if tr is not None else func_nii.header.get_zooms()[-1]
 
-        self.mask_data = np.ones_like(func_nii.get_fdata(), dtype='uint8')
+        self.mask_data = nb.fileslice.strided_scalar(func_nii.shape[:3], np.uint8(1))
         if mask_file:
-            self.mask_data = nb.load(mask_file).get_fdata().astype('uint8')
+            self.mask_data = np.asanyarray(nb.load(mask_file).dataobj).astype('uint8')
 
         self.seg_data = None
         if seg_file:
-            self.seg_data = nb.load(seg_file).get_fdata()
+            self.seg_data = np.asanyarray(nb.load(seg_file).dataobj)
 
         if units is None:
             units = {}

niworkflows/viz/utils.py

@@ -200,7 +200,7 @@ def cuts_from_bbox(mask_nii, cuts=3):
     """Finds equi-spaced cuts for presenting images"""
     from nibabel.affines import apply_affine
 
-    mask_data = mask_nii.get_fdata() > 0.0
+    mask_data = np.asanyarray(mask_nii.dataobj) > 0.0
 
     # First, project the number of masked voxels on each axes
     ijk_counts = [
@@ -262,7 +262,7 @@ def _3d_in_file(in_file):
     except AttributeError:
         in_file = in_file
 
-    if in_file.get_fdata().ndim == 3:
+    if len(in_file.shape) == 3:
         return in_file
 
     return nlimage.index_img(in_file, 0)