def test_compute_shape(im_seg, expected, params): metrics = process_seg.compute_shape(im_seg, algo_fitting=PARAM.algo_fitting, angle_correction=params['angle_corr'], verbose=VERBOSE) for key in expected.keys(): # fetch obtained_value if 'slice' in params: obtained_value = float(metrics['area'].data[params['slice']]) else: obtained_value = float(np.mean(metrics[key].data)) # fetch expected_value if expected[key] is np.nan: assert math.isnan(obtained_value) break else: expected_value = pytest.approx(expected[key], rel=0.05) assert obtained_value == expected_value
def test_compute_shape_noangle(dummy_segmentation): """Test computation of cross-sectional area from input segmentation.""" # Using hanning because faster metrics = process_seg.compute_shape(dummy_segmentation(shape='ellipse', angle=0, a=50.0, b=30.0), algo_fitting=PARAM.algo_fitting, verbose=VERBOSE) assert np.mean(metrics['area'].data[30:70]) == pytest.approx(47.01, rel=0.05) assert np.mean(metrics['AP_diameter'].data[30:70]) == pytest.approx( 6.0, rel=0.05) assert np.mean(metrics['RL_diameter'].data[30:70]) == pytest.approx( 10.0, rel=0.05) assert np.mean(metrics['ratio_minor_major'].data[30:70]) == pytest.approx( 0.6, rel=0.05) assert np.mean(metrics['eccentricity'].data[30:70]) == pytest.approx( 0.8, rel=0.05) assert np.mean(metrics['orientation'].data[30:70]) == pytest.approx( 0.0, rel=0.05) assert np.mean(metrics['solidity'].data[30:70]) == pytest.approx(1.0, rel=0.05)
def test_compute_shape(im_seg, expected, params): metrics, fit_results = process_seg.compute_shape(im_seg, angle_correction=params['angle_corr'], param_centerline=ParamCenterline(), verbose=VERBOSE) for key in expected.keys(): # fetch obtained_value if 'slice' in params: obtained_value = float(metrics['area'].data[params['slice']]) else: if key == 'length': # when computing length, sums values across slices obtained_value = metrics[key].data.sum() else: # otherwise, average across slices obtained_value = metrics[key].data.mean() # fetch expected_value if expected[key] is np.nan: assert math.isnan(obtained_value) break else: expected_value = pytest.approx(expected[key], rel=0.05) assert obtained_value == expected_value
def main(args=None): parser = get_parser() if args: arguments = parser.parse_args(args) else: arguments = parser.parse_args( args=None if sys.argv[1:] else ['--help']) # Initialization slices = '' group_funcs = (('MEAN', func_wa), ('STD', func_std) ) # functions to perform when aggregating metrics along S-I fname_segmentation = get_absolute_path(arguments.i) fname_vert_levels = '' if arguments.o is not None: file_out = os.path.abspath(arguments.o) else: file_out = '' if arguments.append is not None: append = arguments.append else: append = 0 if arguments.vert is not None: vert_levels = arguments.vert else: vert_levels = '' remove_temp_files = arguments.r if arguments.vertfile is not None: fname_vert_levels = arguments.vertfile if arguments.perlevel is not None: perlevel = arguments.perlevel else: perlevel = None if arguments.z is not None: slices = arguments.z if arguments.perslice is not None: perslice = arguments.perslice else: perslice = None angle_correction = arguments.angle_corr param_centerline = ParamCenterline(algo_fitting=arguments.centerline_algo, smooth=arguments.centerline_smooth, minmax=True) path_qc = arguments.qc qc_dataset = arguments.qc_dataset qc_subject = arguments.qc_subject verbose = int(arguments.v) init_sct(log_level=verbose, update=True) # Update log level # update fields metrics_agg = {} if not file_out: file_out = 'csa.csv' metrics, fit_results = compute_shape(fname_segmentation, angle_correction=angle_correction, param_centerline=param_centerline, verbose=verbose) for key in metrics: if key == 'length': # For computing cord length, slice-wise length needs to be summed across slices metrics_agg[key] = aggregate_per_slice_or_level( metrics[key], slices=parse_num_list(slices), levels=parse_num_list(vert_levels), perslice=perslice, perlevel=perlevel, vert_level=fname_vert_levels, group_funcs=(('SUM', func_sum), )) else: # For other metrics, we compute the average and standard deviation across slices metrics_agg[key] = aggregate_per_slice_or_level( metrics[key], slices=parse_num_list(slices), levels=parse_num_list(vert_levels), perslice=perslice, perlevel=perlevel, vert_level=fname_vert_levels, group_funcs=group_funcs) metrics_agg_merged = merge_dict(metrics_agg) save_as_csv(metrics_agg_merged, file_out, fname_in=fname_segmentation, append=append) # QC report (only show CSA for clarity) if path_qc is not None: generate_qc(fname_segmentation, args=args, path_qc=os.path.abspath(path_qc), dataset=qc_dataset, subject=qc_subject, path_img=_make_figure(metrics_agg_merged, fit_results), process='sct_process_segmentation') display_open(file_out)
def main(args): parser = get_parser() arguments = parser.parse(args) # Initialization slices = '' group_funcs = (('MEAN', func_wa), ('STD', func_std)) # functions to perform when aggregating metrics along S-I fname_segmentation = sct.get_absolute_path(arguments['-i']) fname_vert_levels = '' if '-o' in arguments: file_out = os.path.abspath(arguments['-o']) else: file_out = '' if '-append' in arguments: append = int(arguments['-append']) else: append = 0 if '-vert' in arguments: vert_levels = arguments['-vert'] else: vert_levels = '' if '-r' in arguments: remove_temp_files = arguments['-r'] if '-vertfile' in arguments: fname_vert_levels = arguments['-vertfile'] if '-perlevel' in arguments: perlevel = arguments['-perlevel'] else: perlevel = None if '-z' in arguments: slices = arguments['-z'] if '-perslice' in arguments: perslice = arguments['-perslice'] else: perslice = None if '-angle-corr' in arguments: if arguments['-angle-corr'] == '1': angle_correction = True elif arguments['-angle-corr'] == '0': angle_correction = False param_centerline = ParamCenterline( algo_fitting=arguments['-centerline-algo'], smooth=arguments['-centerline-smooth'], minmax=True) path_qc = arguments.get("-qc", None) qc_dataset = arguments.get("-qc-dataset", None) qc_subject = arguments.get("-qc-subject", None) verbose = int(arguments.get('-v')) sct.init_sct(log_level=verbose, update=True) # Update log level # update fields metrics_agg = {} if not file_out: file_out = 'csa.csv' metrics, fit_results = process_seg.compute_shape(fname_segmentation, angle_correction=angle_correction, param_centerline=param_centerline, verbose=verbose) for key in metrics: metrics_agg[key] = aggregate_per_slice_or_level(metrics[key], slices=parse_num_list(slices), levels=parse_num_list(vert_levels), perslice=perslice, perlevel=perlevel, vert_level=fname_vert_levels, group_funcs=group_funcs) metrics_agg_merged = _merge_dict(metrics_agg) save_as_csv(metrics_agg_merged, file_out, fname_in=fname_segmentation, append=append) # QC report (only show CSA for clarity) if path_qc is not None: generate_qc(fname_segmentation, args=args, path_qc=os.path.abspath(path_qc), dataset=qc_dataset, subject=qc_subject, path_img=_make_figure(metrics_agg_merged, fit_results), process='sct_process_segmentation') sct.display_open(file_out)
def main(args): parser = get_parser() arguments = parser.parse(args) param = Param() # Initialization slices = param.slices group_funcs = (('MEAN', func_wa), ('STD', func_std) ) # functions to perform when aggregating metrics along S-I fname_segmentation = sct.get_absolute_path(arguments['-i']) fname_vert_levels = '' if '-o' in arguments: file_out = os.path.abspath(arguments['-o']) else: file_out = '' if '-append' in arguments: append = int(arguments['-append']) else: append = 0 if '-vert' in arguments: vert_levels = arguments['-vert'] else: vert_levels = '' if '-r' in arguments: remove_temp_files = arguments['-r'] if '-vertfile' in arguments: fname_vert_levels = arguments['-vertfile'] if '-perlevel' in arguments: perlevel = arguments['-perlevel'] else: perlevel = Param().perlevel if '-z' in arguments: slices = arguments['-z'] if '-perslice' in arguments: perslice = arguments['-perslice'] else: perslice = Param().perslice if '-angle-corr' in arguments: if arguments['-angle-corr'] == '1': angle_correction = True elif arguments['-angle-corr'] == '0': angle_correction = False verbose = int(arguments.get('-v')) sct.init_sct(log_level=verbose, update=True) # Update log level # update fields metrics_agg = {} if not file_out: file_out = 'csa.csv' metrics = process_seg.compute_shape(fname_segmentation, algo_fitting='bspline', angle_correction=angle_correction, verbose=verbose) for key in metrics: metrics_agg[key] = aggregate_per_slice_or_level( metrics[key], slices=parse_num_list(slices), levels=parse_num_list(vert_levels), perslice=perslice, perlevel=perlevel, vert_level=fname_vert_levels, group_funcs=group_funcs) metrics_agg_merged = _merge_dict(metrics_agg) save_as_csv(metrics_agg_merged, file_out, fname_in=fname_segmentation, append=append) sct.display_open(file_out)
def main(args): parser = get_parser() arguments = parser.parse(args) param = Param() # Initialization slices = param.slices angle_correction = True use_phys_coord = True group_funcs = (('MEAN', func_wa), ('STD', func_std)) # functions to perform when aggregating metrics along S-I fname_segmentation = sct.get_absolute_path(arguments['-i']) name_process = arguments['-p'] fname_vert_levels = '' if '-o' in arguments: file_out = os.path.abspath(arguments['-o']) else: file_out = '' if '-append' in arguments: append = int(arguments['-append']) else: append = 0 if '-vert' in arguments: vert_levels = arguments['-vert'] else: vert_levels = '' if '-r' in arguments: remove_temp_files = arguments['-r'] if '-vertfile' in arguments: fname_vert_levels = arguments['-vertfile'] if '-perlevel' in arguments: perlevel = arguments['-perlevel'] else: perlevel = Param().perlevel if '-v' in arguments: verbose = int(arguments['-v']) if '-z' in arguments: slices = arguments['-z'] if '-perslice' in arguments: perslice = arguments['-perslice'] else: perslice = Param().perslice if '-a' in arguments: param.algo_fitting = arguments['-a'] if '-no-angle' in arguments: if arguments['-no-angle'] == '1': angle_correction = False elif arguments['-no-angle'] == '0': angle_correction = True if '-use-image-coord' in arguments: if arguments['-use-image-coord'] == '1': use_phys_coord = False if arguments['-use-image-coord'] == '0': use_phys_coord = True # update fields param.verbose = verbose metrics_agg = {} if not file_out: file_out = name_process + '.csv' if name_process == 'centerline': process_seg.extract_centerline(fname_segmentation, verbose=param.verbose, algo_fitting=param.algo_fitting, use_phys_coord=use_phys_coord, file_out=file_out) if name_process == 'csa': metrics = process_seg.compute_csa(fname_segmentation, algo_fitting=param.algo_fitting, type_window=param.type_window, window_length=param.window_length, angle_correction=angle_correction, use_phys_coord=use_phys_coord, remove_temp_files=remove_temp_files, verbose=verbose) for key in metrics: metrics_agg[key] = aggregate_per_slice_or_level(metrics[key], slices=parse_num_list(slices), levels=parse_num_list(vert_levels), perslice=perslice, perlevel=perlevel, vert_level=fname_vert_levels, group_funcs=group_funcs) metrics_agg_merged = merge_dict(metrics_agg) save_as_csv(metrics_agg_merged, file_out, fname_in=fname_segmentation, append=append) sct.printv('\nFile created: '+file_out, verbose=1, type='info') if name_process == 'label-vert': if '-discfile' in arguments: fname_discs = arguments['-discfile'] else: sct.printv('\nERROR: Disc label file is mandatory (flag: -discfile).\n', 1, 'error') process_seg.label_vert(fname_segmentation, fname_discs, verbose=verbose) if name_process == 'shape': fname_discs = None if '-discfile' in arguments: fname_discs = arguments['-discfile'] metrics = process_seg.compute_shape(fname_segmentation, remove_temp_files=remove_temp_files, verbose=verbose) for key in metrics: metrics_agg[key] = aggregate_per_slice_or_level(metrics[key], slices=parse_num_list(slices), levels=parse_num_list(vert_levels), perslice=perslice, perlevel=perlevel, vert_level=fname_vert_levels, group_funcs=group_funcs) metrics_agg_merged = merge_dict(metrics_agg) save_as_csv(metrics_agg_merged, file_out, fname_in=fname_segmentation, append=append) sct.printv('\nFile created: ' + file_out, verbose=1, type='info')
def main(args): parser = get_parser() arguments = parser.parse(args) param = Param() # Initialization slices = param.slices group_funcs = (('MEAN', func_wa), ('STD', func_std)) # functions to perform when aggregating metrics along S-I fname_segmentation = sct.get_absolute_path(arguments['-i']) fname_vert_levels = '' if '-o' in arguments: file_out = os.path.abspath(arguments['-o']) else: file_out = '' if '-append' in arguments: append = int(arguments['-append']) else: append = 0 if '-vert' in arguments: vert_levels = arguments['-vert'] else: vert_levels = '' if '-r' in arguments: remove_temp_files = arguments['-r'] if '-vertfile' in arguments: fname_vert_levels = arguments['-vertfile'] if '-perlevel' in arguments: perlevel = arguments['-perlevel'] else: perlevel = Param().perlevel if '-z' in arguments: slices = arguments['-z'] if '-perslice' in arguments: perslice = arguments['-perslice'] else: perslice = Param().perslice if '-angle-corr' in arguments: if arguments['-angle-corr'] == '1': angle_correction = True elif arguments['-angle-corr'] == '0': angle_correction = False verbose = int(arguments.get('-v')) sct.init_sct(log_level=verbose, update=True) # Update log level # update fields metrics_agg = {} if not file_out: file_out = 'csa.csv' metrics = process_seg.compute_shape(fname_segmentation, algo_fitting='bspline', angle_correction=angle_correction, verbose=verbose) for key in metrics: metrics_agg[key] = aggregate_per_slice_or_level(metrics[key], slices=parse_num_list(slices), levels=parse_num_list(vert_levels), perslice=perslice, perlevel=perlevel, vert_level=fname_vert_levels, group_funcs=group_funcs) metrics_agg_merged = _merge_dict(metrics_agg) save_as_csv(metrics_agg_merged, file_out, fname_in=fname_segmentation, append=append) sct.display_open(file_out)