# emacs: -*- mode: python-mode; py-indent-offset: 4; indent-tabs-mode: nil -*- # vi: set ft=python sts=4 ts=4 sw=4 et: ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ## # # See COPYING file distributed along with the NiBabel package for the # copyright and license terms. # ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ## import bz2 import gzip from io import BytesIO from os.path import join as pjoin import numpy as np import pytest from numpy.testing import assert_array_equal from .. import Nifti1Image, load, minc1 from .._compression import HAVE_ZSTD, zstd from ..externals.netcdf import netcdf_file from ..minc1 import Minc1File, Minc1Image, MincHeader from ..testing import assert_data_similar, data_path from ..tmpdirs import InTemporaryDirectory from . import test_spatialimages as tsi from .test_fileslice import slicer_samples EG_FNAME = pjoin(data_path, 'tiny.mnc') # Example images in format expected for ``test_image_api``, adding ``zooms`` # item. EXAMPLE_IMAGES = [ dict( fname=pjoin(data_path, 'tiny.mnc'), shape=(10, 20, 20), dtype=np.uint8, affine=np.array( [ [0, 0, 2.0, -20], [0, 2.0, 0, -20], [2.0, 0, 0, -10], [0, 0, 0, 1], ] ), zooms=(2.0, 2.0, 2.0), # These values from SPM2 data_summary=dict(min=0.20784314, max=0.74901961, mean=0.60602819), is_proxy=True, ), dict( fname=pjoin(data_path, 'minc1_1_scale.mnc'), shape=(10, 20, 20), dtype=np.uint8, affine=np.array( [ [0, 0, 2.0, -20], [0, 2.0, 0, -20], [2.0, 0, 0, -10], [0, 0, 0, 1], ] ), zooms=(2.0, 2.0, 2.0), # These values from mincstats data_summary=dict(min=0.2082842439, max=0.2094327615, mean=0.2091292083), is_proxy=True, ), dict( fname=pjoin(data_path, 'minc1_4d.mnc'), shape=(2, 10, 20, 20), dtype=np.uint8, affine=np.array( [ [0, 0, 2.0, -20], [0, 2.0, 0, -20], [2.0, 0, 0, -10], [0, 0, 0, 1], ] ), zooms=(1.0, 2.0, 2.0, 2.0), # These values from mincstats data_summary=dict(min=0.2078431373, max=1.498039216, mean=0.9090422837), is_proxy=True, ), dict( fname=pjoin(data_path, 'minc1-no-att.mnc'), shape=(10, 20, 20), dtype=np.uint8, affine=np.array( [ [0, 0, 1.0, 0], [0, 1.0, 0, 0], [1.0, 0, 0, 0], [0, 0, 0, 1], ] ), zooms=(1.0, 1.0, 1.0), # These values from SPM2/mincstats data_summary=dict(min=0.20784314, max=0.74901961, mean=0.6061103), is_proxy=True, ), ] class _TestMincFile: module = minc1 file_class = Minc1File fname = EG_FNAME opener = netcdf_file test_files = EXAMPLE_IMAGES def test_mincfile(self): for tp in self.test_files: mnc_obj = self.opener(tp['fname'], 'r') mnc = self.file_class(mnc_obj) assert mnc.get_data_dtype().type == tp['dtype'] assert mnc.get_data_shape() == tp['shape'] assert mnc.get_zooms() == tp['zooms'] assert_array_equal(mnc.get_affine(), tp['affine']) data = mnc.get_scaled_data() assert data.shape == tp['shape'] # Can't close mmapped NetCDF with live mmap arrays del mnc, data def test_mincfile_slicing(self): # Test slicing and scaling of mincfile data for tp in self.test_files: mnc_obj = self.opener(tp['fname'], 'r') mnc = self.file_class(mnc_obj) data = mnc.get_scaled_data() for slicedef in ( (slice(None),), (1,), (slice(None), 1), (1, slice(None)), (slice(None), 1, 1), (1, slice(None), 1), (1, 1, slice(None)), ): sliced_data = mnc.get_scaled_data(slicedef) assert_array_equal(sliced_data, data[slicedef]) # Can't close mmapped NetCDF with live mmap arrays del mnc, data def test_load(self): # Check highest level load of minc works for tp in self.test_files: img = load(tp['fname']) data = img.get_fdata() assert data.shape == tp['shape'] # min, max, mean values from read in SPM2 / minctools assert_data_similar(data, tp) # check if mnc can be converted to nifti ni_img = Nifti1Image.from_image(img) assert_array_equal(ni_img.affine, tp['affine']) assert_array_equal(ni_img.get_fdata(), data) def test_array_proxy_slicing(self): # Test slicing of array proxy for tp in self.test_files: img = load(tp['fname']) arr = img.get_fdata() prox = img.dataobj assert prox.is_proxy for sliceobj in slicer_samples(img.shape): assert_array_equal(arr[sliceobj], prox[sliceobj]) class TestMinc1File(_TestMincFile): def test_compressed(self): # we can read minc compressed # Not so for MINC2; hence this small sub-class for tp in self.test_files: content = open(tp['fname'], 'rb').read() openers_exts = [(gzip.open, '.gz'), (bz2.BZ2File, '.bz2')] if HAVE_ZSTD: # add .zst to test if installed openers_exts += [(zstd.ZstdFile, '.zst')] with InTemporaryDirectory(): for opener, ext in openers_exts: fname = 'test.mnc' + ext fobj = opener(fname, 'wb') fobj.write(content) fobj.close() img = self.module.load(fname) data = img.get_fdata() assert_data_similar(data, tp) del img # Test the Minc header def test_header_data_io(): bio = BytesIO() hdr = MincHeader() arr = np.arange(24).reshape((2, 3, 4)) with pytest.raises(NotImplementedError): hdr.data_to_fileobj(arr, bio) with pytest.raises(NotImplementedError): hdr.data_from_fileobj(bio) class TestMinc1Image(tsi.TestSpatialImage): image_class = Minc1Image eg_images = (pjoin(data_path, 'tiny.mnc'),) module = minc1 def test_data_to_from_fileobj(self): # Check data_from_fileobj of header raises an error for fpath in self.eg_images: img = self.module.load(fpath) bio = BytesIO() arr = np.arange(24).reshape((2, 3, 4)) with pytest.raises(NotImplementedError): img.header.data_to_fileobj(arr, bio) with pytest.raises(NotImplementedError): img.header.data_from_fileobj(bio)