def test_ALE_ma_map_reuse(testdata_cbma, tmp_path_factory, caplog): """Test that MA maps are re-used when appropriate.""" from nimare.meta import kernel tmpdir = tmp_path_factory.mktemp("test_ALE_ma_map_reuse") testdata_cbma.update_path(tmpdir) # ALEKernel cannot extract sample_size from a Dataset, # so we need to set it for this kernel and for the later meta-analyses. kern = kernel.ALEKernel(sample_size=20) dset = kern.transform(testdata_cbma, return_type="dataset") # The associated column should be in the new Dataset's images DataFrame cols = dset.images.columns.tolist() assert any(["ALEKernel" in col for col in cols]) # The Dataset without the images will generate them from scratch. # If drop_invalid is False, then there should be an Exception, since two studies in the test # dataset are missing coordinates. meta = ale.ALE(kernel__sample_size=20) with pytest.raises(Exception): meta.fit(testdata_cbma, drop_invalid=False) with caplog.at_level(logging.DEBUG, logger="nimare.meta.cbma.base"): meta.fit(testdata_cbma) assert "Loading pre-generated MA maps" not in caplog.text # The Dataset with the images will re-use them, as evidenced by the logger message. with caplog.at_level(logging.DEBUG, logger="nimare.meta.cbma.base"): meta.fit(dset) assert "Loading pre-generated MA maps" in caplog.text
def test_ALEKernel_smoke(testdata_cbma): """Smoke test for nimare.meta.kernel.ALEKernel.""" # Manually override dataset coordinates file sample sizes # This column would be extracted from metadata and added to coordinates # automatically by the Estimator coordinates = testdata_cbma.coordinates.copy() coordinates["sample_size"] = 20 kern = kernel.ALEKernel() ma_maps = kern.transform(coordinates, testdata_cbma.masker, return_type="image") assert len(ma_maps) == len(testdata_cbma.ids) - 2 ma_maps = kern.transform(coordinates, testdata_cbma.masker, return_type="array") assert ma_maps.shape[0] == len(testdata_cbma.ids) - 2 # Test set_params kern.set_params(fwhm=10, sample_size=None) kern2 = kernel.ALEKernel(fwhm=10) ma_maps1 = kern.transform(coordinates, testdata_cbma.masker, return_type="array") ma_maps2 = kern2.transform(coordinates, testdata_cbma.masker, return_type="array") assert ma_maps1.shape[0] == ma_maps2.shape[0] == len(testdata_cbma.ids) - 2 assert np.array_equal(ma_maps1, ma_maps2)
def test_ALEKernel_fwhm(testdata_cbma): """Peaks of ALE kernel maps should match the foci fed in (assuming focus isn't masked out). Test with explicit FWHM. """ coordinates = testdata_cbma.coordinates.copy() id_ = "pain_03.nidm-1" kern = kernel.ALEKernel(fwhm=10) ma_maps = kern.transform(coordinates, masker=testdata_cbma.masker, return_type="image") ijk = coordinates.loc[coordinates["id"] == id_, ["i", "j", "k"]] ijk = np.squeeze(ijk.values.astype(int)) kern_data = ma_maps[0].get_fdata() max_idx = np.array(np.where(kern_data == np.max(kern_data))).T max_ijk = np.squeeze(max_idx) assert np.array_equal(ijk, max_ijk)
def test_ALEKernel_sample_size(testdata_cbma): """Peaks of ALE kernel maps should match the foci fed in (assuming focus isn't masked out). Test with explicit sample size. """ coordinates = testdata_cbma.coordinates.copy() id_ = "pain_03.nidm-1" kern = kernel.ALEKernel(sample_size=20) ma_maps = kern.transform(coordinates, masker=testdata_cbma.masker, return_type="image") xyz = coordinates.loc[coordinates["id"] == id_, ["x", "y", "z"]] ijk = mm2vox(xyz, testdata_cbma.masker.mask_img.affine) ijk = np.squeeze(ijk.astype(int)) kern_data = ma_maps[0].get_fdata() max_idx = np.array(np.where(kern_data == np.max(kern_data))).T max_ijk = np.squeeze(max_idx) assert np.array_equal(ijk, max_ijk)
def test_ALEKernel_1mm(testdata_cbma): """Peaks of ALE kernel maps should match the foci fed in (assuming focus isn't masked out). Test on 1mm template. """ # Manually override dataset coordinates file sample sizes # This column would be extracted from metadata and added to coordinates # automatically by the Estimator coordinates = testdata_cbma.coordinates.copy() coordinates["sample_size"] = 20 id_ = "pain_03.nidm-1" kern = kernel.ALEKernel() ma_maps = kern.transform(coordinates, testdata_cbma.masker, return_type="image") ijk = coordinates.loc[coordinates["id"] == id_, ["i", "j", "k"]] ijk = ijk.values.astype(int) kern_data = ma_maps[0].get_fdata() max_idx = np.where(kern_data == np.max(kern_data)) max_ijk = np.array(max_idx).T assert np.array_equal(ijk, max_ijk)
def test_ALEKernel_inputdataset_returndataset(testdata_cbma, tmp_path_factory): """Check that all return types produce equivalent results (minus the masking element).""" tmpdir = tmp_path_factory.mktemp("test_ALEKernel_inputdataset_returndataset") testdata_cbma.update_path(tmpdir) kern = kernel.ALEKernel(sample_size=20, memory_limit="1gb") ma_maps = kern.transform(testdata_cbma, return_type="image") ma_arr = kern.transform(testdata_cbma, return_type="array") dset = kern.transform(testdata_cbma, return_type="dataset") ma_maps_from_dset = kern.transform(dset, return_type="image") ma_arr_from_dset = kern.transform(dset, return_type="array") ma_maps_arr = testdata_cbma.masker.transform(ma_maps) ma_maps_from_dset_arr = dset.masker.transform(ma_maps_from_dset) dset_from_dset = kern.transform(dset, return_type="dataset") ids = dset.coordinates["id"].unique() ma_maps_dset = testdata_cbma.masker.transform(dset.get_images(ids=ids, imtype=kern.image_type)) assert isinstance(dset_from_dset, Dataset) assert np.array_equal(ma_arr, ma_maps_arr) assert np.array_equal(ma_arr, ma_maps_dset) assert np.array_equal(ma_arr, ma_maps_from_dset_arr) assert np.array_equal(ma_arr, ma_arr_from_dset)
def test_ALESubtraction_ma_map_reuse(testdata_cbma, tmp_path_factory, caplog): """Test that MA maps are re-used when appropriate.""" from nimare.meta import kernel tmpdir = tmp_path_factory.mktemp("test_ALESubtraction_ma_map_reuse") testdata_cbma.update_path(tmpdir) # ALEKernel cannot extract sample_size from a Dataset, # so we need to set it for this kernel and for the later meta-analyses. kern = kernel.ALEKernel(sample_size=20) dset = kern.transform(testdata_cbma, return_type="dataset") # The Dataset without the images will generate them from scratch. sub_meta = ale.ALESubtraction(n_iters=10, kernel__sample_size=20) with caplog.at_level(logging.DEBUG, logger="nimare.meta.cbma.base"): sub_meta.fit(testdata_cbma, testdata_cbma) assert "Loading pre-generated MA maps" not in caplog.text # The Dataset with the images will re-use them, # as evidenced by the logger message. with caplog.at_level(logging.DEBUG, logger="nimare.meta.cbma.base"): sub_meta.fit(dset, dset) assert "Loading pre-generated MA maps" in caplog.text
# There are three standard kernels that are currently available: {py:class}`~nimare.meta.kernel.MKDAKernel`, {py:class}`~nimare.meta.kernel.KDAKernel`, and {py:class}`~nimare.meta.kernel.ALEKernel`. # Each class may be configured with certain parameters when a new object is initialized. # For example, `MKDAKernel` accepts an `r` parameter, which determines the radius of the spheres that will be created around each peak coordinate. # `ALEKernel` automatically uses the sample size associated with each experiment in the `Dataset` to determine the appropriate full-width-at-half-maximum of its Gaussian distribution, as described in {cite:t}`EICKHOFF20122349`; however, users may provide a constant `sample_size` or `fwhm` parameter when sample size information is not available within the `Dataset` metadata. # # Here we show how these three kernels can be applied to the same `Dataset`. # In[2]: from nimare.meta import kernel mkda_kernel = kernel.MKDAKernel(r=10) mkda_ma_maps = mkda_kernel.transform(sleuth_dset1) kda_kernel = kernel.KDAKernel(r=10) kda_ma_maps = kda_kernel.transform(sleuth_dset1) ale_kernel = kernel.ALEKernel(sample_size=20) ale_ma_maps = ale_kernel.transform(sleuth_dset1) # In[3]: # Here we delete the recent variables for the sake of reducing memory usage del mkda_kernel, kda_kernel, ale_kernel # In[4]: # Generate figure study_idx = 10 # a study with overlapping kernels max_value = np.max(kda_ma_maps[study_idx].get_fdata()) + 1 ma_maps = { "MKDA Kernel": mkda_ma_maps[study_idx],