def test_Normalize(self):
"""Load unnormalized feature space, normalize,
then compare to stored normalized feature space."""
from wndcharm.utils import compare
result_fs = FeatureSpace.NewFromFitFile(
self.test_fit_path).Normalize(inplace=True)
target_fs = FeatureSpace.NewFromFitFile(self.test_normalized_fit_path)
from numpy.testing import assert_allclose
assert_allclose(result_fs.data_matrix,
target_fs.data_matrix,
rtol=1e-05)
def test_LDATransform( self ):
"""LDA transform"""
tempdir = mkdtemp()
import zipfile
zf = zipfile.ZipFile( pychrm_test_dir + sep + 'lymphoma_iicbu2008_subset_EOSIN_ONLY_t5x6_v3.2features.fit.zip', mode='r')
zf.extractall( tempdir )
fitfile_path = tempdir + sep + 'lymphoma_iicbu2008_subset_eosin_t5x6_v3.2features.fit'
try:
kwargs = {}
kwargs['pathname'] = fitfile_path
kwargs['quiet'] = True
# sampling opts: -l -t5x6 implies 5 columns and 6 rows ... I know it's weird.
kwargs['long'] = True
kwargs['tile_num_rows'] = 6
kwargs['tile_num_cols'] = 5
# against self:
fs = FeatureSpace.NewFromFitFile( **kwargs )
self_transformed = fs.LDATransform( reference_features=None, inplace=False )
fit_on_fit_LDA_result = FeatureSpaceClassification.NewWND5(
self_transformed, self_transformed, feature_weights=None )
# against other:
train, test = fs.Split()
train.LDATransform( reference_features=None, inplace=True )
test.LDATransform( reference_features=train, inplace=True )
split_LDA_result = FeatureSpaceClassification.NewWND5(
train, test, feature_weights=None )
finally:
rmtree( tempdir )
def test_DiscreteTrainTestSplitWithTiling( self ):
"""Uses a curated subset of the IICBU 2008 Lymphoma dataset, preprocessed as follows:
auto-deconvolved, eosin channel only, tiled 5x6, 3 classes, 10 imgs per class,
300 samples per class.
"""
# Inflate the zipped test fit into a temp file
import zipfile
zipped_file_path = pychrm_test_dir + sep + 'lymphoma_iicbu2008_subset_EOSIN_ONLY_t5x6_v3.2features.fit.zip'
zf = zipfile.ZipFile( zipped_file_path, mode='r' )
tempdir = mkdtemp()
zf.extractall( tempdir )
try:
fitfilepath = tempdir + sep + zf.namelist()[0]
#fs = FeatureSet.NewFromFitFile( fitfilepath )
fs = FeatureSpace.NewFromFitFile( fitfilepath, tile_num_rows=5, tile_num_cols=6 )
from numpy.random import RandomState
prng = RandomState(42)
train, test = fs.Split( random_state=prng, quiet=True )
train.Normalize( inplace=True, quiet=True )
fw = FisherFeatureWeights.NewFromFeatureSpace( train ).Threshold()
train.FeatureReduce( fw, inplace=True )
test.FeatureReduce( fw, inplace=True ).Normalize( train, inplace=True, quiet=True )
finally:
rmtree( tempdir )
def test_NewFromFeatureSet(self):
"""Fisher score calculation"""
fs = FeatureSpace.NewFromFitFile(
self.test_fit_path).Normalize(inplace=True)
fw = FisherFeatureWeights.NewFromFeatureSpace(fs)
# test weights generated from test-l.fit:
# wndchrm classify -l -f1.0 -vtest_fit-l.weights test-l.fit test-l.fit
target_fw = FisherFeatureWeights.NewFromFile(
self.test_feat_weight_path)
for target_val, res_val in zip(target_fw.values, fw.values):
self.assertAlmostEqual(target_val, res_val, delta=self.epsilon)
# test slice operator
orig_len = len(fw)
sliced = fw[:10]
self.assertEqual(len(sliced), 10)
self.assertEqual(len(sliced.feature_names), 10)
self.assertEqual(len(sliced.values), 10)
for i in xrange(10):
self.assertEqual(sliced.feature_names[i], fw.feature_names[i])
self.assertEqual(sliced.values[i], fw.values[i])
sliced = fw[50:100:2]
for i, j in zip(range(len(sliced)), range(50, 100, 2)):
self.assertEqual(sliced.feature_names[i], fw.feature_names[j])
self.assertEqual(sliced.values[i], fw.values[j])
def test_ContinuousFitOnFit(self):
from wndcharm.ArtificialFeatureSpace import CreateArtificialFeatureSpace_Discrete
fs_discrete = CreateArtificialFeatureSpace_Discrete(
n_samples=1000,
n_classes=10,
num_features_per_signal_type=30,
noise_gradient=5,
initial_noise_sigma=10,
n_samples_per_group=1,
interpolatable=True)
tempdir = mkdtemp()
path_to_fit = tempdir + sep + 'Artificial.fit'
try:
fs_discrete.ToFitFile(path_to_fit)
fs_continuous = FeatureSpace.NewFromFitFile(path_to_fit,
discrete=False)
fs_continuous.Normalize(quiet=True)
fw_reduced = PearsonFeatureWeights.NewFromFeatureSpace(
fs_continuous).Threshold()
fs_reduced = fs_continuous.FeatureReduce(fw_reduced)
batch_result = FeatureSpaceRegression.NewMultivariateLinear(
fs_reduced, fw_reduced, quiet=True)
finally:
rmtree(tempdir)
def test_SamplesUnion(self):
from wndcharm.ArtificialFeatureSpace import CreateArtificialFeatureSpace_Discrete
n_classes = 2
fs1 = CreateArtificialFeatureSpace_Discrete(
n_samples=20,
n_classes=n_classes,
num_features_per_signal_type=30,
noise_gradient=5,
initial_noise_sigma=10,
n_samples_per_group=1,
interpolatable=True)
fitfile_path = wndchrm_test_dir + sep + 'test-l.fit'
fs2 = FeatureSpace.NewFromFitFile(fitfile_path)
self.assertRaises(ValueError, fs1.SamplesUnion, other_fs=fs2)
fs3 = CreateArtificialFeatureSpace_Discrete(
n_samples=20,
n_classes=n_classes,
num_features_per_signal_type=30,
noise_gradient=5,
initial_noise_sigma=10,
n_samples_per_group=1,
interpolatable=True)
joined_fs = fs1 + fs3
self.assertEqual(n_classes, joined_fs.num_classes)
def test_FitOnFitClassification(self):
fitfile_path = wndchrm_test_dir + sep + 'test-l.fit'
#fs = FeatureSet.NewFromFitFile( fitfile_path )
fs = FeatureSpace.NewFromFitFile(fitfile_path)
fs.Normalize(inplace=True, quiet=True)
fw = FisherFeatureWeights.NewFromFeatureSpace(fs).Threshold(438)
fw.Print(50)
fs.FeatureReduce(fw, inplace=True)
pychrm_split = FeatureSpaceClassification.NewWND5(fs,
fs,
fw,
quiet=False)
from wndcharm.FeatureSpacePredictionExperiment import FeatureSpaceClassificationExperiment
html_path = pychrm_test_dir + sep + 'test-l_training_error_result.html'
html_exp = FeatureSpaceClassificationExperiment.NewFromHTMLReport(
html_path, quiet=False)
# single split in this html
html_split = html_exp.individual_results[0]
for i, (html_result, pychrm_result) in enumerate( zip( html_split.individual_results,\
pychrm_split.individual_results ) ):
try:
self.assertEqual(html_result, pychrm_result)
except:
outstr = "Error in comparison # {0}:\n".format(i)
outstr += "HTML result:\n{0}\n Python API res:\n{1}".format(
html_result, pychrm_result)
raise
def test_Normalize( self ):
"""Load unnormalized feature space, normalize,
then compare to stored normalized feature space."""
raw_fs = FeatureSpace.NewFromFitFile( self.test_fit_path )
result_fs = raw_fs.Normalize( inplace=False )
target_fs = FeatureSpace.NewFromFitFile( self.test_normalized_fit_path )
from numpy.testing import assert_allclose
assert_allclose( result_fs.data_matrix, target_fs.data_matrix, rtol=1e-05 )
# Create reference Z-score feature space
from scipy.stats.mstats import zscore
from wndcharm.utils import ReplaceNonReal
ReplaceNonReal( raw_fs.data_matrix )
oldsettings = np.seterr( all='ignore' )
target_fs = zscore( raw_fs.data_matrix )
target_fs[ np.isnan( target_fs) ] = 0
np.seterr( **oldsettings )
result_fs = raw_fs.Normalize( inplace=False, zscore=True )
assert_allclose( result_fs.data_matrix, target_fs )
def test_WND5_all_features(self):
epsilon = 0.00001
# Define paths to original files
test_sig_path = join(test_dir, 't1_s01_c05_ij-l_precalculated.sig')
test_fit_path = join(test_dir, 'test-l.fit')
test_feat_wght_path = join(test_dir, 'test_fit-l.weights')
test_tif_path = join(test_dir, 't1_s01_c05_ij.tif')
# Here are the correct values that Python API needs to return:
# wndchrm classify -l -f0.75 test-l.fit t1_s01_c05_ij.tif
# t1_s01_c05_ij.tif 1.6e-27 0.083 0.917 * 4cell 3.835
# wndchrm classify -l test-l.fit t1_s01_c05_ij.tif
# t1_s01_c05_ij.tif 3.19e-27 0.076 0.924 * 4cell 3.848
# wndchrm classify -l -f0.05 test-l.fit t1_s01_c05_ij.tif
# t1_s01_c05_ij.tif 1.06e-26 0.066 0.934 * 4cell 3.869
correct_marg_probs = {}
correct_marg_probs[2189] = [0.083, 0.917]
correct_marg_probs[438] = [0.076, 0.924]
correct_marg_probs[146] = [0.066, 0.934]
# Load the original files once and only once for all this class's tests
feature_set = FeatureSpace.NewFromFitFile(test_fit_path)
fs1 = feature_set.feature_names
feature_set.Normalize()
fs2 = feature_set.feature_names
self.assertSequenceEqual(fs1, fs2)
test_sample = FeatureVector(source_filepath=test_tif_path, long=True)
test_sample.LoadSigFile(test_sig_path)
self.assertSequenceEqual(feature_set.feature_names,
test_sample.feature_names)
test_sample.Normalize(feature_set)
all_weights = FisherFeatureWeights.NewFromFile(test_feat_wght_path)
def Check(num_feats):
weights = all_weights.Threshold(num_feats)
feat_set = feature_set.FeatureReduce(weights)
sample = test_sample.FeatureReduce(weights)
result = SingleSampleClassification.NewWND5(
feat_set, weights, sample)
result_marg_probs = [ round( val, 3 ) \
for val in result.marginal_probabilities ]
for target_val, res_val in zip(correct_marg_probs[num_feats],
result_marg_probs):
self.assertAlmostEqual(target_val, res_val, delta=epsilon)
for num_feats in correct_marg_probs:
Check(num_feats)
def test_NewFromFeatureSet(self):
"""Fisher score calculation"""
feature_set = FeatureSpace.NewFromFitFile(self.test_fit_path)
feature_set.Normalize(inplace=True)
result_weights = FisherFeatureWeights.NewFromFeatureSpace(feature_set)
# test weights generated from test-l.fit:
# wndchrm classify -l -f1.0 -vtest_fit-l.weights test-l.fit test-l.fit
target_weights = FisherFeatureWeights.NewFromFile(
self.test_feat_weight_path)
for target_val, res_val in zip(target_weights.values,
result_weights.values):
self.assertAlmostEqual(target_val, res_val, delta=self.epsilon)
def test_DiscreteTrainTestSplitNoTiling( self ):
"""Uses binucleate test set"""
fitfilepath = wndchrm_test_dir + sep + 'test-l.fit'
fs = FeatureSpace.NewFromFitFile( fitfilepath )
from numpy.random import RandomState
prng = RandomState(42)
full_train, full_test = fs.Split( random_state=prng, quiet=True )
full_train.Normalize( quiet=True )
reduced_fw = FisherFeatureWeights.NewFromFeatureSpace( full_train ).Threshold()
reduced_train = full_train.FeatureReduce( reduced_fw )
reduced_test = full_test.FeatureReduce( reduced_fw )
reduced_test.Normalize( reduced_train, quiet=True )
batch_result = FeatureSpaceClassification.NewWND5( reduced_train,
reduced_test, reduced_fw, quiet=True )
parser.add_argument( 'output_filepath', help='Results are written to this file, otherwise to STDOUT',
nargs='?')
parser.add_argument( '-D', help='Write the training set to a pickle file', metavar='<optional path>',
default='unset', nargs='?')
args = parser.parse_args()
num_splits = args.n
num_bins = args.b
input_filename = args.classifier_file_path[0]
outpath = args.output_filepath
dump_pickle = args.D
if input_filename.endswith( ".fit" ):
full_set = FeatureSpace.NewFromFitFile( input_filename )
elif input_filename.endswith( ".fit.pickled" ):
full_set = FeatureSpace.NewFromPickleFile( input_filename )
elif input_filename.endswith( ".fof" ):
full_set = FeatureSpace.NewFromFileOfFiles( input_filename )
else:
raise Exception( 'The classifier must either end in .fit, .fit.pickled, or .fof' )
if not dump_pickle == 'unset':
if dump_pickle:
# user used -D to specify a name for their training set pickle
full_set.PickleMe( dump_pickle )
else:
# user used -D as a flag, use default pickle name pattern
full_set.PickleMe()
def test_NewFromFileOfFiles( self ):
"""Pulls in the lymphoma eosin histology 5x6 tiled featureset via .sig files."""
# Types of files containing features:
# FIT: contains an entire FeatureSpace definition including features.
# FOF: "File Of Files" containing a FeatureSpace structure definition only,
# listing paths to files of pre-calculated features (.sig) or the
# tiff images themselves so features can be calculated
# SIG: A text file containing pre-calculated features for a single sample.
# Test dataset: subset of the IICBU2008 lymphoma dataset. 2 channels (H+E),
# 3 classes ('CLL', 'FL', 'MCL'), 10 images per class per channel,
# 5x6 tiling grid = 30 samples per image resulting in
# 2 x 3 x 10 X 30 = 1800 total samples available
# Files containing features included in this test suite:
# 1. lymphoma_iicbu2008_subset_EOSIN_ONLY_t5x6_v3.2features.fit.zip:
# A zip archive containing a single FIT file with features pre-calculated.
# 2. lymphoma_iicbu2008_subset_HE_t5x6_v3.2features_SIGFILES.zip:
# Contains 1800 SIG files, plus 4 FOF files (items 2-5 below):
# "lymphoma_iicbu2008_subset_EOSIN_ONLY_images.fof.tsv"
# "lymphoma_iicbu2008_subset_EOSIN_ONLY_sigfiles_t5x6-l.fof.tsv"
# "lymphoma_iicbu2008_subset_2CHAN_HE_images.fof.tsv"
# "lymphoma_iicbu2008_subset_2CHAN_HE_sigfiles_t5x6-l.fof.tsv"
# List of possible feature sources:
# 1. Single channel FIT (Eosin only)
# 2. Single channel FOF (Eosin only) referencing to 30 tiffs (requires global sampling options -t5x6 -l to grab sigs)
# 3. Single channel FOF (Eosin only) referencing 900 sig files
# 4. Double channel FOF (Eosin+Haemotoxylin) referencing 60 tiffs (requires global sampling options -t5x6 -l to grab sigs)
# 5. Double channel FOF (Eosin+Haemotoxylin) referencing 1800 sig files.
#=============================================
# BEGIN CODE TO CREATE TESTDATA ZIP PACKAGE
#import zipfile
#import zlib
#path = '/Users/chris/src/wnd-charm/tests/pywndcharm_tests/TESTDATA_lymphoma_iicbu2008_subset_HE_t5x6_v3.2features_SIGFILES.zip'
#zf = zipfile.ZipFile( path, mode='w' )
#import os
#classes = 'CLL', 'FL', 'MCL',
#channels = 'haemotoxylin', 'eosin'
#from collections import defaultdict
#sig_tracker = defaultdict(int)
#samplegroupid_tracker = {}
#samplegroup_counter = 0
#
#eosin_tif_fof = [] # 30 lines
#eosin_sig_fof = [] # 900 lines
#double_tif_fof = [] # 30 lines, 2 feature set columns
#double_sig_fof = [] # 900 lines, 2 feature set columns
#
#for _channel in channels:
# zf.write( './' + _channel, compress_type=zipfile.ZIP_DEFLATED )
# for _class in classes:
# zf.write( './' + _channel + '/' + _class, compress_type=zipfile.ZIP_DEFLATED )
# for root, dirs, files in os.walk( _channel + '/' + _class ):
# for _file in files:
# if _file.endswith( '.tif' ):
# # Strip off the _H.tif or _E.tif
# samplename = _file[:-6]
# eosinpath = './eosin/' + _class + '/' + samplename + '_E.tif'
# haemopath = './haemotoxylin/' + _class + '/' + samplename + '_H.tif'
# if _channel == 'eosin':
# eosin_tif_fof.append( eosinpath + '\t' + _class )
# double_tif_fof.append( samplename + '\t' + _class + '\t' + eosinpath + '\t{\tchannel\t=\teosin\t}\t' + haemopath + '\t{\tchannel\t=\thaemotoxylin\t}')
# elif _file.endswith( '.sig' ):
# zf.write( './' + _channel + '/' + _class + '/' + _file, compress_type=zipfile.ZIP_DEFLATED )
# if _channel == 'eosin':
# # Strip off the _H-t5x6_0_0-l.sig
# samplename = _file[:-17] + '.tif'
# eosinpath = './eosin/' + _class + '/' + _file
# haemopath = './haemotoxylin/' + _class + '/' + _file.replace( '_E-t5x6_', '_H-t5x6_' )
# # count samples from 0:
# samplesequenceid = str( sig_tracker[ samplename ] )
# sig_tracker[ samplename ] += 1
# if samplename not in samplegroupid_tracker:
# samplegroupid_tracker[ samplename ] = samplegroup_counter
# samplegroup_counter += 1
# samplegroupid = str( samplegroupid_tracker[ samplename ] )
# eosin_sig_fof.append( eosinpath + '\t' + _class )
# double_sig_fof.append( samplename + '\t' + _class + '\t' + eosinpath + '\t{\tchannel\t=\teosin\t;\tsamplegroupid\t=\t' + samplegroupid + '\t;\tsamplesequenceid\t=\t' + samplesequenceid + '\t}\t' + haemopath + '\t{\tchannel\t=\thaemotoxylin\t;\tsamplegroupid\t=\t' + samplegroupid + '\t;\tsamplesequenceid\t=\t' + samplesequenceid + '\t}\t')
#
#fof_dir = '/Users/chris/src/wnd-charm/tests/pywndcharm_tests/'
#with open( 'lymphoma_iicbu2008_subset_EOSIN_ONLY_images.fof.tsv', 'w') as out:
# for _ in eosin_tif_fof:
# out.write( _ + '\n')
#with open( 'lymphoma_iicbu2008_subset_EOSIN_ONLY_sigfiles_t5x6-l.fof.tsv', 'w') as out:
# for _ in eosin_sig_fof:
# out.write( _ + '\n')
#with open( 'lymphoma_iicbu2008_subset_2CHAN_HE_images.fof.tsv', 'w') as out:
# for _ in double_tif_fof:
# out.write( _ + '\n')
#with open( 'lymphoma_iicbu2008_subset_2CHAN_HE_sigfiles_t5x6-l.fof.tsv', 'w') as out:
# for _ in double_sig_fof:
# out.write( _ + '\n')
#zf.write( './' + 'lymphoma_iicbu2008_subset_EOSIN_ONLY_images.fof.tsv', compress_type=zipfile.ZIP_DEFLATED )
#zf.write( './' + 'lymphoma_iicbu2008_subset_EOSIN_ONLY_sigfiles_t5x6-l.fof.tsv', compress_type=zipfile.ZIP_DEFLATED )
#zf.write( './' + 'lymphoma_iicbu2008_subset_2CHAN_HE_images.fof.tsv', compress_type=zipfile.ZIP_DEFLATED )
#zf.write( './' + 'lymphoma_iicbu2008_subset_2CHAN_HE_sigfiles_t5x6-l.fof.tsv', compress_type=zipfile.ZIP_DEFLATED )
#zf.printdir()
#zf.close()
# END CODE TO CREATE TESTDATA ZIP PACKAGE
#=============================================
# Inflate the zipped test fit into a temp file
import zipfile
zipped_file_path = pychrm_test_dir + sep + 'lymphoma_iicbu2008_subset_HE_t5x6_v3.2features_SIGFILES.zip'
zf1 = zipfile.ZipFile( zipped_file_path, mode='r' )
tempdir = mkdtemp()
zf1.extractall( tempdir )
# for comparison:
zf2 = zipfile.ZipFile( pychrm_test_dir + sep + 'lymphoma_iicbu2008_subset_EOSIN_ONLY_t5x6_v3.2features.fit.zip', mode='r')
zf2.extractall( tempdir )
try:
kwargs = {}
kwargs['pathname'] = tempdir + sep + 'lymphoma_iicbu2008_subset_EOSIN_ONLY_sigfiles_t5x6-l.fof.tsv'
kwargs['quiet'] = True
# sampling opts: -l -t5x6 implies 5 columns and 6 rows ... I know it's weird.
kwargs['long'] = True
kwargs['tile_num_rows'] = 6
kwargs['tile_num_cols'] = 5
fs_fof = FeatureSpace.NewFromFileOfFiles( **kwargs )
kwargs['pathname'] = tempdir + sep + 'lymphoma_iicbu2008_subset_eosin_t5x6_v3.2features.fit'
fs_fit = FeatureSpace.NewFromFitFile( **kwargs )
# Fit file has less significant figures than Signature files, and it's not
# consistent how many there are. Seems like fit file just lops off numbers
# at the end. Example: (signatures on top, fit on bottom)
#
# Example:
# - 17.232246, # sig
# ? --
#
# + 17.2322, # fit
# - -63.549056, # sig
# ? ^^^
#
# + -63.5491, # fit
# ? ^
#
# - 223.786977, # sig
# ? ---
#
# + 223.787, # fit
# More of the same:
#(Pdb) fs_fof.data_matrix[0,-5:]
#array([ 0.935442, 14.005003, -43.562076, 127.394914, 0.628772])
#(Pdb) fs_fit.data_matrix[0,-5:]
#array([ 0.935442, 14.005 , -43.5621 , 127.395 , 0.628772])
# default is rtol=1e-07, atol=0
#np.testing.assert_allclose( actual=fs_fit.data_matrix, desired=fs_fof.data_matrix,
# rtol=1e-03, atol=0 )
#np.testing.assert_array_almost_equal_nulp( fs_fit.data_matrix, fs_fof.data_matrix )
for row_num, (fit_row, fof_row) in enumerate( zip( fs_fit.data_matrix, fs_fof.data_matrix )):
retval = compare( fit_row, fof_row )
if retval == False:
print "error in sample row", row_num
print "FIT: ", fs_fit._contiguous_sample_names[row_num], "FOF", fs_fof._contiguous_sample_names[row_num]
self.assertTrue( retval )
# Test sorting; scramble the FOF then load and check:
sorted_fof = tempdir + sep + \
'lymphoma_iicbu2008_subset_EOSIN_ONLY_sigfiles_t5x6-l.fof.tsv'
with open( sorted_fof) as fof:
lines = fof.readlines()
from random import shuffle
shuffle(lines)
unsorted_fof = tempdir + sep + \
'lymphoma_iicbu2008_subset_EOSIN_ONLY_sigfiles_t5x6-l_UNSORTED.fof.tsv'
with open( unsorted_fof, 'w' ) as fof:
for line in lines:
fof.write( line )
kwargs = {}
kwargs['pathname'] = unsorted_fof
kwargs['quiet'] = True
# sampling opts: -l -t5x6 implies 5 columns and 6 rows ... I know it's weird.
kwargs['long'] = True
kwargs['tile_num_rows'] = 6
kwargs['tile_num_cols'] = 5
fs_fof = FeatureSpace.NewFromFileOfFiles( **kwargs )
# Check again
for row_num, (fit_row, fof_row) in enumerate( zip( fs_fit.data_matrix, fs_fof.data_matrix )):
retval = compare( fit_row, fof_row )
if retval == False:
print "error in sample row", row_num
print "FIT: ", fs_fit._contiguous_sample_names[row_num], "FOF", fs_fof._contiguous_sample_names[row_num]
self.assertTrue( retval )
# TESTING TAKE TILES:
self.assertRaises( ValueError, fs_fof.TakeTiles, tuple() )
self.assertRaises( ValueError, fs_fof.TakeTiles, (45, 46, 47,) )
self.assertRaises( TypeError, fs_fof.TakeTiles, 'crap' )
# take middle 4
wanted_tiles = ( 14, 15, 20, 21 )
took = fs_fof.TakeTiles( wanted_tiles, inplace=False )
num_sample_groups = len( set( fs_fof._contiguous_sample_group_ids ) )
self.assertEqual( took.num_samples_per_group, len( wanted_tiles ) )
self.assertEqual( took.num_samples, len( wanted_tiles ) * num_sample_groups )
# mid4 = 'lymphoma_iicbu2008_subset_EOSIN_ONLY_sigfiles_MIDDLE_4_TILES_t5x6-l.fof.tsv'
# # fake out wndcharm by putting empty tiffs in the temp dir
# # we don't need them, the sigs are in there already.
# with open( mid4) as fof:
# lines = fof.readlines()
# names, classes, paths, opts = zip( *[ _.split('\t') for _ in lines ] )
# for _path in paths:
# with open( tempdir + sep + _path, 'w' ):
# pass
# took_via_fof = FeatureSpace.NewFromFileOfFiles( mid4, num_samples_per_group=4 )
#
# for row_num, (fit_row, fof_row) in enumerate( zip( took.data_matrix, took_via_fof.data_matrix )):
# retval = compare( fit_row, fof_row )
# if retval == False:
# print "error in sample row", row_num
# print "FIT: ", took._contiguous_sample_names[row_num], "FOF", took_via_fof._contiguous_sample_names[row_num]
# self.assertTrue( retval )
finally:
rmtree( tempdir )
def test_FitOnFit(self):
"""Uses a curated subset of the IICBU 2008 Lymphoma dataset, preprocessed as follows:
auto-deconvolved, eosin channel only, tiled 5x6, 3 classes, 10 imgs per class,
300 samples per class.
"""
# Inflate the zipped test fit into a temp file
import zipfile
zipped_file_path = pychrm_test_dir + sep + 'lymphoma_iicbu2008_subset_EOSIN_ONLY_t5x6_v3.2features.fit.zip'
zf = zipfile.ZipFile(zipped_file_path, mode='r')
tempdir = mkdtemp()
zf.extractall(tempdir)
try:
fitfilepath = tempdir + sep + zf.namelist()[0]
# Do fit on fit WITHOUT tiling and compare with fit on fit results
# generated with wndchrm 1.60
fs = FeatureSpace.NewFromFitFile(fitfilepath).Normalize(
inplace=True, quiet=True)
#fs = FeatureSpace.NewFromFitFile( wndchrm_test_dir + sep + 'test-l.fit' )
#fs.ToFitFile( 'temp.fit' )
fw = FisherFeatureWeights.NewFromFeatureSpace(fs).Threshold()
fs.FeatureReduce(fw, inplace=True)
# #fw.Print()
# #fs.Print(verbose=True)
pychrm_res = FeatureSpaceClassification.NewWND5(fs, fs, fw)
pychrm_res.Print()
#
# import cProfile as pr
# #import profile as pr
# import tempfile
# import pstats
# prof = tempfile.NamedTemporaryFile()
# cmd = 'no_tile_pychrm_result = DiscreteBatchClassificationResult.New( reduced_fs, reduced_fs, fw )'
# pr.runctx( cmd, globals(), locals(), prof.name)
# p = pstats.Stats(prof.name)
# p.sort_stats('time').print_stats(20)
# prof.close()
self.maxDiff = None
html_path = pychrm_test_dir + sep + 'lymphoma_iicbu2008_subset_eosin_t5x6_v3.2feats_REFERENCE_RESULTS_900_samples_TRAINING_ERROR.html'
wres = FeatureSpaceClassificationExperiment.NewFromHTMLReport(
html_path)
wres.Print()
wc_batch_result = wres.individual_results[
0] # only 1 split in fit-on-fit
# This takes WAY too long:
#self.assertSequenceEqual( wc_batch_result.individual_results, pychrm_res.individual_results )
wc_result = np.empty((3 * len(wc_batch_result.individual_results)))
for i, single_result in enumerate(
wc_batch_result.individual_results):
wc_result[i * 3:(i + 1) *
3] = single_result.marginal_probabilities
pc_result = np.empty((3 * len(pychrm_res.individual_results)))
for i, single_result in enumerate(pychrm_res.individual_results):
# HTML report only has 3 decimal places
pc_result[ i*3 : (i+1)*3 ] = \
[ float( "{0:0.3f}".format( val ) ) for val in single_result.marginal_probabilities ]
from numpy.testing import assert_allclose
assert_allclose(actual=pc_result, desired=wc_result, atol=0.003)
#wc_batch_result.Print()
#pres.Print()
# ==========================================================
# Now do the same with tiling, reusing fs from before:
num_samples_per_group = 30
n_groups = fs.num_samples / num_samples_per_group
new_sg_ids = [
i for i in xrange(n_groups)
for j in xrange(num_samples_per_group)
]
fs.Update( tile_num_rows=5, tile_num_cols=6, num_samples_per_group=30,\
_contiguous_sample_group_ids=new_sg_ids )._RebuildViews()
with_tile_pychrm_result = FeatureSpaceClassification.NewWND5(
fs, fs, fw)
html_path = pychrm_test_dir + sep + 'lymphoma_iicbu2008_subset_eosin_t5x6_v3.2feats_REFERENCE_RESULTS_30_samples_tiled_TRAINING_ERROR.html'
with_tile_wndchrm_result = \
FeatureSpaceClassificationExperiment.NewFromHTMLReport( html_path ).individual_results[0]
#self.assertSequenceEqual( with_tile_pychrm_result.averaged_results, with_tile_wndchrm_result.individual_results )
wc_result = np.empty(
(3 * len(with_tile_wndchrm_result.individual_results)))
for i, single_result in enumerate(
with_tile_wndchrm_result.individual_results):
wc_result[i * 3:(i + 1) *
3] = single_result.marginal_probabilities
pc_result = np.empty(
(3 * len(with_tile_pychrm_result.averaged_results)))
for i, single_result in enumerate(
with_tile_pychrm_result.averaged_results):
# HTML report only has 3 decimal places
pc_result[ i*3 : (i+1)*3 ] = \
[ float( "{0:0.3f}".format( val ) ) for val in single_result.marginal_probabilities ]
assert_allclose(actual=pc_result, desired=wc_result, atol=0.003)
finally:
rmtree(tempdir)
def test_TileOptions(self):
fs = FeatureSpace.NewFromFitFile(wndchrm_test_dir + sep + 'test-l.fit',
tile_options)