def testZeroMeanNormalize(self):
graphs = []
for i in range(11):
nodeFeatures = np.random.randint(low=0, high=100, size=(i + 1, 13))
g = S2VGraph('binaryId',
g=nx.Graph(),
label='label',
node_tags=[],
node_features=nodeFeatures)
graphs.append(g)
normalizeFeatures(graphs, operation='zero_mean')
for (i, g) in enumerate(graphs):
log.debug(f'Feature shape after norm: {g.node_features.shape}')
self.assertEqual(g.node_features.shape[0], i + 1, 'First dim size')
self.assertEqual(g.node_features.shape[1], 13, 'Second dim size')
def testMinMaxNormalize(self):
graphs = []
for i in range(11):
nodeFeatures = np.random.randint(low=0, high=100, size=(i + 1, 13))
g = S2VGraph('binaryId',
g=nx.Graph(),
label='label',
node_tags=[],
node_features=nodeFeatures)
graphs.append(g)
normalizeFeatures(graphs, operation='min_max')
for (i, g) in enumerate(graphs):
log.debug(f'Feature shape after norm: {g.node_features.shape}')
self.assertEqual(g.node_features.shape[0], i + 1, 'First dim size')
self.assertEqual(g.node_features.shape[1], 13, 'Second dim size')
for x in np.nditer(g.node_features):
self.assertLessEqual(x, 1.0, 'Not <= 1')
self.assertGreaterEqual(x, 0.0, 'Not >= 0')
histFile = open(
'%sGpu%sRun%s.csv' % (cmd_args.data, cmd_args.gpu_id, runId), 'w')
histFile.write("# %s\n" % str(gHP))
df.to_csv(histFile, index_label='Epoch', float_format='%.6f')
histFile.close()
if __name__ == '__main__':
log.setLevel("INFO")
random.seed(cmd_args.seed)
np.random.seed(cmd_args.seed)
torch.manual_seed(cmd_args.seed)
startTime = time.process_time()
graphs = loadGraphsMayCache(cmd_args.train_dir)
normalizeFeatures(graphs, isTestSet=False, operation=cmd_args.norm_op)
trainGraphs = filterOutNoEdgeGraphs(trainGraphs)
dataReadyTime = time.process_time() - startTime
log.info('Dataset ready takes %.2fs' % dataReadyTime)
for (id, hp) in enumerate(HyperParameterIterator(cmd_args.hp_path)):
for (key, val) in hp.items():
gHP[key] = val
numNodesList = sorted([g.num_nodes for g in graphs])
idx = int(math.ceil(hp['poolingRatio'] * len(graphs))) - 1
gHP['poolingK'] = numNodesList[idx]
kFoldGraphs = kFoldSplit(gHP['cvFold'], graphs)
crossValidate(kFoldGraphs, id)
classifier = classifier.cuda()
classifier.eval()
startTime = time.process_time()
testPredProb = predictDataset(testGraphs, classifier)
log.info(f'Net test time = {time.process_time() - startTime} seconds')
exportPredictions(testGraphs, testPredProb)
if __name__ == '__main__':
log.setLevel("INFO")
random.seed(cmd_args.seed)
np.random.seed(cmd_args.seed)
torch.manual_seed(cmd_args.seed)
if cmd_args.data == 'YANACFG':
log.warning(f'No testset for YANACFG data')
else:
startTime = time.process_time()
testGraphs = loadGraphsMayCache(cmd_args.test_dir, isTestSet=True)
normalizeFeatures(testGraphs,
isTestSet=True,
operation=cmd_args.norm_op)
dataReadyTime = time.process_time() - startTime
log.info('Testset ready takes %.2fs' % dataReadyTime)
trainGraphs = loadGraphsMayCache(cmd_args.train_dir, isTestSet=False)
trainGraphs = filterOutNoEdgeGraphs(trainGraphs)
decideHyperparameters(trainGraphs)
testWithModel(testGraphs)