How to use test_empty method in yandex-tank

Best Python code snippet using yandex-tank

metrics.py

Source:metrics.py Github

copy

Full Screen

1# Copyright 2020 Division of Medical Image Computing, German Cancer Research Center (DKFZ), Heidelberg, Germany2#3# Licensed under the Apache License, Version 2.0 (the "License");4# you may not use this file except in compliance with the License.5# You may obtain a copy of the License at6#7# http://www.apache.org/licenses/LICENSE-2.08#9# Unless required by applicable law or agreed to in writing, software10# distributed under the License is distributed on an "AS IS" BASIS,11# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.12# See the License for the specific language governing permissions and13# limitations under the License.14import numpy as np15from medpy import metric16def assert_shape(test, reference):17 assert test.shape == reference.shape, "Shape mismatch: {} and {}".format(18 test.shape, reference.shape)19class ConfusionMatrix:20 def __init__(self, test=None, reference=None):21 self.tp = None22 self.fp = None23 self.tn = None24 self.fn = None25 self.size = None26 self.reference_empty = None27 self.reference_full = None28 self.test_empty = None29 self.test_full = None30 self.set_reference(reference)31 self.set_test(test)32 def set_test(self, test):33 self.test = test34 self.reset()35 def set_reference(self, reference):36 self.reference = reference37 self.reset()38 def reset(self):39 self.tp = None40 self.fp = None41 self.tn = None42 self.fn = None43 self.size = None44 self.test_empty = None45 self.test_full = None46 self.reference_empty = None47 self.reference_full = None48 def compute(self):49 if self.test is None or self.reference is None:50 raise ValueError("'test' and 'reference' must both be set to compute confusion matrix.")51 assert_shape(self.test, self.reference)52 self.tp = int(((self.test != 0) * (self.reference != 0)).sum())53 self.fp = int(((self.test != 0) * (self.reference == 0)).sum())54 self.tn = int(((self.test == 0) * (self.reference == 0)).sum())55 self.fn = int(((self.test == 0) * (self.reference != 0)).sum())56 self.size = int(np.prod(self.reference.shape, dtype=np.int64))57 self.test_empty = not np.any(self.test)58 self.test_full = np.all(self.test)59 self.reference_empty = not np.any(self.reference)60 self.reference_full = np.all(self.reference)61 def get_matrix(self):62 for entry in (self.tp, self.fp, self.tn, self.fn):63 if entry is None:64 self.compute()65 break66 return self.tp, self.fp, self.tn, self.fn67 def get_size(self):68 if self.size is None:69 self.compute()70 return self.size71 def get_existence(self):72 for case in (self.test_empty, self.test_full, self.reference_empty, self.reference_full):73 if case is None:74 self.compute()75 break76 return self.test_empty, self.test_full, self.reference_empty, self.reference_full77def dice(test=None, reference=None, confusion_matrix=None, nan_for_nonexisting=True, **kwargs):78 """2TP / (2TP + FP + FN)"""79 if confusion_matrix is None:80 confusion_matrix = ConfusionMatrix(test, reference)81 tp, fp, tn, fn = confusion_matrix.get_matrix()82 test_empty, test_full, reference_empty, reference_full = confusion_matrix.get_existence()83 if test_empty and reference_empty:84 if nan_for_nonexisting:85 return float("NaN")86 else:87 return 0.88 return float(2. * tp / (2 * tp + fp + fn))89def jaccard(test=None, reference=None, confusion_matrix=None, nan_for_nonexisting=True, **kwargs):90 """TP / (TP + FP + FN)"""91 if confusion_matrix is None:92 confusion_matrix = ConfusionMatrix(test, reference)93 tp, fp, tn, fn = confusion_matrix.get_matrix()94 test_empty, test_full, reference_empty, reference_full = confusion_matrix.get_existence()95 if test_empty and reference_empty:96 if nan_for_nonexisting:97 return float("NaN")98 else:99 return 0.100 return float(tp / (tp + fp + fn))101def precision(test=None, reference=None, confusion_matrix=None, nan_for_nonexisting=True, **kwargs):102 """TP / (TP + FP)"""103 if confusion_matrix is None:104 confusion_matrix = ConfusionMatrix(test, reference)105 tp, fp, tn, fn = confusion_matrix.get_matrix()106 test_empty, test_full, reference_empty, reference_full = confusion_matrix.get_existence()107 if test_empty:108 if nan_for_nonexisting:109 return float("NaN")110 else:111 return 0.112 return float(tp / (tp + fp))113def sensitivity(test=None, reference=None, confusion_matrix=None, nan_for_nonexisting=True, **kwargs):114 """TP / (TP + FN)"""115 if confusion_matrix is None:116 confusion_matrix = ConfusionMatrix(test, reference)117 tp, fp, tn, fn = confusion_matrix.get_matrix()118 test_empty, test_full, reference_empty, reference_full = confusion_matrix.get_existence()119 if reference_empty:120 if nan_for_nonexisting:121 return float("NaN")122 else:123 return 0.124 return float(tp / (tp + fn))125def recall(test=None, reference=None, confusion_matrix=None, nan_for_nonexisting=True, **kwargs):126 """TP / (TP + FN)"""127 return sensitivity(test, reference, confusion_matrix, nan_for_nonexisting, **kwargs)128def specificity(test=None, reference=None, confusion_matrix=None, nan_for_nonexisting=True, **kwargs):129 """TN / (TN + FP)"""130 if confusion_matrix is None:131 confusion_matrix = ConfusionMatrix(test, reference)132 tp, fp, tn, fn = confusion_matrix.get_matrix()133 test_empty, test_full, reference_empty, reference_full = confusion_matrix.get_existence()134 if reference_full:135 if nan_for_nonexisting:136 return float("NaN")137 else:138 return 0.139 return float(tn / (tn + fp))140def accuracy(test=None, reference=None, confusion_matrix=None, **kwargs):141 """(TP + TN) / (TP + FP + FN + TN)"""142 if confusion_matrix is None:143 confusion_matrix = ConfusionMatrix(test, reference)144 tp, fp, tn, fn = confusion_matrix.get_matrix()145 return float((tp + tn) / (tp + fp + tn + fn))146def fscore(test=None, reference=None, confusion_matrix=None, nan_for_nonexisting=True, beta=1., **kwargs):147 """(1 + b^2) * TP / ((1 + b^2) * TP + b^2 * FN + FP)"""148 precision_ = precision(test, reference, confusion_matrix, nan_for_nonexisting)149 recall_ = recall(test, reference, confusion_matrix, nan_for_nonexisting)150 return (1 + beta*beta) * precision_ * recall_ /\151 ((beta*beta * precision_) + recall_)152def false_positive_rate(test=None, reference=None, confusion_matrix=None, nan_for_nonexisting=True, **kwargs):153 """FP / (FP + TN)"""154 return 1 - specificity(test, reference, confusion_matrix, nan_for_nonexisting)155def false_omission_rate(test=None, reference=None, confusion_matrix=None, nan_for_nonexisting=True, **kwargs):156 """FN / (TN + FN)"""157 if confusion_matrix is None:158 confusion_matrix = ConfusionMatrix(test, reference)159 tp, fp, tn, fn = confusion_matrix.get_matrix()160 test_empty, test_full, reference_empty, reference_full = confusion_matrix.get_existence()161 if test_full:162 if nan_for_nonexisting:163 return float("NaN")164 else:165 return 0.166 return float(fn / (fn + tn))167def false_negative_rate(test=None, reference=None, confusion_matrix=None, nan_for_nonexisting=True, **kwargs):168 """FN / (TP + FN)"""169 return 1 - sensitivity(test, reference, confusion_matrix, nan_for_nonexisting)170def true_negative_rate(test=None, reference=None, confusion_matrix=None, nan_for_nonexisting=True, **kwargs):171 """TN / (TN + FP)"""172 return specificity(test, reference, confusion_matrix, nan_for_nonexisting)173def false_discovery_rate(test=None, reference=None, confusion_matrix=None, nan_for_nonexisting=True, **kwargs):174 """FP / (TP + FP)"""175 return 1 - precision(test, reference, confusion_matrix, nan_for_nonexisting)176def negative_predictive_value(test=None, reference=None, confusion_matrix=None, nan_for_nonexisting=True, **kwargs):177 """TN / (TN + FN)"""178 return 1 - false_omission_rate(test, reference, confusion_matrix, nan_for_nonexisting)179def total_positives_test(test=None, reference=None, confusion_matrix=None, **kwargs):180 """TP + FP"""181 if confusion_matrix is None:182 confusion_matrix = ConfusionMatrix(test, reference)183 tp, fp, tn, fn = confusion_matrix.get_matrix()184 return tp + fp185def total_negatives_test(test=None, reference=None, confusion_matrix=None, **kwargs):186 """TN + FN"""187 if confusion_matrix is None:188 confusion_matrix = ConfusionMatrix(test, reference)189 tp, fp, tn, fn = confusion_matrix.get_matrix()190 return tn + fn191def total_positives_reference(test=None, reference=None, confusion_matrix=None, **kwargs):192 """TP + FN"""193 if confusion_matrix is None:194 confusion_matrix = ConfusionMatrix(test, reference)195 tp, fp, tn, fn = confusion_matrix.get_matrix()196 return tp + fn197def total_negatives_reference(test=None, reference=None, confusion_matrix=None, **kwargs):198 """TN + FP"""199 if confusion_matrix is None:200 confusion_matrix = ConfusionMatrix(test, reference)201 tp, fp, tn, fn = confusion_matrix.get_matrix()202 return tn + fp203def hausdorff_distance(test=None, reference=None, confusion_matrix=None, nan_for_nonexisting=True, voxel_spacing=None, connectivity=1, **kwargs):204 if confusion_matrix is None:205 confusion_matrix = ConfusionMatrix(test, reference)206 test_empty, test_full, reference_empty, reference_full = confusion_matrix.get_existence()207 if test_empty or test_full or reference_empty or reference_full:208 if nan_for_nonexisting:209 return float("NaN")210 else:211 return 0212 test, reference = confusion_matrix.test, confusion_matrix.reference213 return metric.hd(test, reference, voxel_spacing, connectivity)214def hausdorff_distance_95(test=None, reference=None, confusion_matrix=None, nan_for_nonexisting=True, voxel_spacing=None, connectivity=1, **kwargs):215 if confusion_matrix is None:216 confusion_matrix = ConfusionMatrix(test, reference)217 test_empty, test_full, reference_empty, reference_full = confusion_matrix.get_existence()218 if test_empty or test_full or reference_empty or reference_full:219 if nan_for_nonexisting:220 return float("NaN")221 else:222 return 0223 test, reference = confusion_matrix.test, confusion_matrix.reference224 return metric.hd95(test, reference, voxel_spacing, connectivity)225def avg_surface_distance(test=None, reference=None, confusion_matrix=None, nan_for_nonexisting=True, voxel_spacing=None, connectivity=1, **kwargs):226 if confusion_matrix is None:227 confusion_matrix = ConfusionMatrix(test, reference)228 test_empty, test_full, reference_empty, reference_full = confusion_matrix.get_existence()229 if test_empty or test_full or reference_empty or reference_full:230 if nan_for_nonexisting:231 return float("NaN")232 else:233 return 0234 test, reference = confusion_matrix.test, confusion_matrix.reference235 return metric.asd(test, reference, voxel_spacing, connectivity)236def avg_surface_distance_symmetric(test=None, reference=None, confusion_matrix=None, nan_for_nonexisting=True, voxel_spacing=None, connectivity=1, **kwargs):237 if confusion_matrix is None:238 confusion_matrix = ConfusionMatrix(test, reference)239 test_empty, test_full, reference_empty, reference_full = confusion_matrix.get_existence()240 if test_empty or test_full or reference_empty or reference_full:241 if nan_for_nonexisting:242 return float("NaN")243 else:244 return 0245 test, reference = confusion_matrix.test, confusion_matrix.reference246 return metric.assd(test, reference, voxel_spacing, connectivity)247ALL_METRICS = {248 "False Positive Rate": false_positive_rate,249 "Dice": dice,250 "Jaccard": jaccard,251 "Hausdorff Distance": hausdorff_distance,252 "Hausdorff Distance 95": hausdorff_distance_95,253 "Precision": precision,254 "Recall": recall,255 "Avg. Symmetric Surface Distance": avg_surface_distance_symmetric,256 "Avg. Surface Distance": avg_surface_distance,257 "Accuracy": accuracy,258 "False Omission Rate": false_omission_rate,259 "Negative Predictive Value": negative_predictive_value,260 "False Negative Rate": false_negative_rate,261 "True Negative Rate": true_negative_rate,262 "False Discovery Rate": false_discovery_rate,263 "Total Positives Test": total_positives_test,264 "Total Negatives Test": total_negatives_test,265 "Total Positives Reference": total_positives_reference,266 "total Negatives Reference": total_negatives_reference...

Full Screen

Full Screen

Automation Testing Tutorials

Learn to execute automation testing from scratch with LambdaTest Learning Hub. Right from setting up the prerequisites to run your first automation test, to following best practices and diving deeper into advanced test scenarios. LambdaTest Learning Hubs compile a list of step-by-step guides to help you be proficient with different test automation frameworks i.e. Selenium, Cypress, TestNG etc.

LambdaTest Learning Hubs:

YouTube

You could also refer to video tutorials over LambdaTest YouTube channel to get step by step demonstration from industry experts.

Run yandex-tank automation tests on LambdaTest cloud grid

Perform automation testing on 3000+ real desktop and mobile devices online.

Try LambdaTest Now !!

Get 100 minutes of automation test minutes FREE!!

Next-Gen App & Browser Testing Cloud

Was this article helpful?

Helpful

NotHelpful