def create_dataset(db, run_config, aws_config=None):
"""
Create a dataset and add it to the ModelHub database.
db: initialized Database object
run_config: RunConfig object describing the dataset to create
aws_config: optional. AWS credentials for downloading data from S3.
"""
# download data to the local filesystem to extract metadata
train_local, test_local = download_data(run_config.train_path,
run_config.test_path, aws_config)
# create the name of the dataset from the path to the data
name = os.path.basename(train_local)
name = name.replace("_train.csv", "").replace(".csv", "")
# process the data into the form ATM needs and save it to disk
meta = MetaData(run_config.class_column, train_local, test_local)
# enter dataset into database
dataset = db.create_dataset(name=name,
description=run_config.data_description,
train_path=run_config.train_path,
test_path=run_config.test_path,
class_column=run_config.class_column,
n_examples=meta.n_examples,
k_classes=meta.k_classes,
d_features=meta.d_features,
majority=meta.majority,
size_kb=old_div(meta.size, 1000))
return dataset
def create_dataset(self):
"""
Create a dataset and add it to the ModelHub database.
"""
# download data to the local filesystem to extract metadata
train_local, test_local = download_data(self.run_conf.train_path,
self.run_conf.test_path,
self.aws_conf)
# create the name of the dataset from the path to the data
name = os.path.basename(train_local)
name = name.replace("_train.csv", "").replace(".csv", "")
# process the data into the form ATM needs and save it to disk
meta = MetaData(self.run_conf.class_column, train_local, test_local)
# enter dataset into database
dataset = self.db.create_dataset(
name=name,
description=self.run_conf.data_description,
train_path=self.run_conf.train_path,
test_path=self.run_conf.test_path,
class_column=self.run_conf.class_column,
n_examples=meta.n_examples,
k_classes=meta.k_classes,
d_features=meta.d_features,
majority=meta.majority,
size_kb=old_div(meta.size, 1000))
return dataset
def model(dataset):
train_path, _ = download_data(dataset.train_path)
model = Model(method='dt', params=DT_PARAMS,
judgment_metric='roc_auc',
class_column=dataset.class_column)
model.train_test(train_path=train_path)
return model
def model(dataset):
train_path, _ = download_data(dataset.train_path)
model = Model(method='dt', params=DT_PARAMS,
judgment_metric='roc_auc',
class_column=dataset.class_column)
model.train_test(train_path=train_path)
return model
def test_classifier(self, method, params):
"""
Given a set of fully-qualified hyperparameters, create and test a
classifier model.
Returns: Model object and metrics dictionary
"""
model = Model(method=method,
params=params,
judgment_metric=self.datarun.metric,
class_column=self.dataset.class_column,
verbose_metrics=self.verbose_metrics)
train_path, test_path = download_data(self.dataset.train_path,
self.dataset.test_path,
self.aws_config)
metrics = model.train_test(train_path=train_path, test_path=test_path)
target = self.datarun.score_target
def metric_string(model):
if 'cv' in target or 'mu_sigma' in target:
return '%.3f +- %.3f' % (model.cv_judgment_metric,
2 * model.cv_judgment_metric_stdev)
else:
return '%.3f' % model.test_judgment_metric
logger.info('Judgment metric (%s, %s): %s' %
(self.datarun.metric, target[:-len('_judgment_metric')],
metric_string(model)))
old_best = self.db.get_best_classifier(datarun_id=self.datarun.id,
score_target=target)
if old_best is not None:
if getattr(model, target) > getattr(old_best, target):
logger.info(
'New best score! Previous best (classifier %s): %s',
old_best.id, metric_string(old_best))
else:
logger.info('Best so far (classifier %s): %s', old_best.id,
metric_string(old_best))
return model, metrics
def enter_dataset(db, run_config, aws_config=None):
"""
Generate a dataset, and update run_config with the dataset ID.
db: Database object with active connection to ModelHub
run_config: all attributes necessary to initialize a Datarun, including
Dataset info
aws_config: all attributes necessary to connect to an S3 bucket.
Returns: the generated dataset object
"""
print 'downloading data...'
train_path, test_path = download_data(run_config.train_path,
run_config.test_path, aws_config)
print 'creating dataset...'
dataset = create_dataset(db, run_config.label_column, train_path, test_path,
run_config.data_description)
run_config.dataset_id = dataset.id
return dataset
