def setUp(self):
super(NextJobHandlerTest, self).setUp()
self.exp_id = 'exp_id1'
self.title = 'Testing Classifier storing'
self.category = 'Test'
interaction_id = 'TextInput'
self.algorithm_id = feconf.INTERACTION_CLASSIFIER_MAPPING[
interaction_id]['algorithm_id']
self.training_data = [{
u'answer_group_index': 1,
u'answers': [u'a1', u'a2']
}, {
u'answer_group_index': 2,
u'answers': [u'a2', u'a3']
}]
self.job_id = classifier_models.ClassifierTrainingJobModel.create(
self.algorithm_id, interaction_id, self.exp_id, 1,
datetime.datetime.utcnow(), self.training_data, 'Home',
feconf.TRAINING_JOB_STATUS_NEW, 1)
fs_services.save_classifier_data(self.exp_id, self.job_id, {})
self.expected_response = {
u'job_id': self.job_id,
u'training_data': self.training_data,
u'algorithm_id': self.algorithm_id
}
self.payload = {}
self.payload['vm_id'] = feconf.DEFAULT_VM_ID
secret = feconf.DEFAULT_VM_SHARED_SECRET
self.payload['message'] = json.dumps({})
self.payload['signature'] = classifier.generate_signature(
python_utils.convert_to_bytes(secret), self.payload['message'])
def test_remove_classifier_data(self) -> None:
"""Test that classifier data is removed upon deletion."""
fs_services.save_classifier_data('exp_id', 'job_id',
self.classifier_data_proto)
self.assertTrue(self.fs.isfile('job_id-classifier-data.pb.xz'))
fs_services.delete_classifier_data('exp_id', 'job_id')
self.assertFalse(self.fs.isfile('job_id-classifier-data.pb.xz'))
def test_save_and_get_classifier_data(self) -> None:
"""Test that classifier data is stored and retrieved correctly."""
fs_services.save_classifier_data('exp_id', 'job_id',
self.classifier_data_proto)
filepath = 'job_id-classifier-data.pb.xz'
fs = fs_services.GcsFileSystem(feconf.ENTITY_TYPE_EXPLORATION,
'exp_id')
classifier_data = utils.decompress_from_zlib(fs.get(filepath))
classifier_data_proto = text_classifier_pb2.TextClassifierFrozenModel()
classifier_data_proto.ParseFromString(classifier_data)
self.assertEqual(classifier_data_proto.model_json,
self.classifier_data_proto.model_json)
def _create_classifier_training_job(self, algorithm_id, interaction_id,
exp_id, exp_version,
next_scheduled_check_time,
training_data, state_name, status,
classifier_data, data_schema_version):
"""Creates a new classifier training job model and stores
classfier data in a file.
"""
job_id = classifier_models.ClassifierTrainingJobModel.create(
algorithm_id, interaction_id, exp_id, exp_version,
next_scheduled_check_time, training_data, state_name, status,
data_schema_version)
fs_services.save_classifier_data(exp_id, job_id, classifier_data)
return job_id
def _create_classifier_training_job(
self, algorithm_id, interaction_id, exp_id, exp_version,
next_scheduled_check_time, training_data, state_name, status,
classifier_data, algorithm_version):
"""Creates a new classifier training job model and stores
classfier data in a file.
"""
job_id = classifier_models.ClassifierTrainingJobModel.create(
algorithm_id, interaction_id, exp_id, exp_version,
next_scheduled_check_time, training_data, state_name, status,
algorithm_version)
classifier_data_proto = text_classifier_pb2.TextClassifierFrozenModel()
classifier_data_proto.model_json = json.dumps(classifier_data)
fs_services.save_classifier_data(exp_id, job_id, classifier_data_proto)
return job_id
def setUp(self):
super(ClassifierTrainingJobModelValidatorTests, self).setUp()
self.signup(self.OWNER_EMAIL, self.OWNER_USERNAME)
self.owner_id = self.get_user_id_from_email(self.OWNER_EMAIL)
explorations = [
exp_domain.Exploration.create_default_exploration(
'%s' % i,
title='title %d' % i,
category='category%d' % i,
) for i in python_utils.RANGE(2)
]
for exp in explorations:
exp.add_states(['StateTest%s' % exp.id])
exp_services.save_new_exploration(self.owner_id, exp)
next_scheduled_check_time = datetime.datetime.utcnow()
classifier_data_proto = text_classifier_pb2.TextClassifierFrozenModel()
classifier_data_proto.model_json = json.dumps(
{'classifier_data': 'data'})
id0 = classifier_models.ClassifierTrainingJobModel.create(
'TextClassifier', 'TextInput', '0', 1, next_scheduled_check_time,
[{
'answer_group_index': 1,
'answers': ['a1', 'a2']
}], 'StateTest0', feconf.TRAINING_JOB_STATUS_NEW, 1)
fs_services.save_classifier_data('TextClassifier', id0,
classifier_data_proto)
self.model_instance_0 = (
classifier_models.ClassifierTrainingJobModel.get_by_id(id0))
id1 = classifier_models.ClassifierTrainingJobModel.create(
'TextClassifier', 'TextInput', '1', 1, next_scheduled_check_time,
[{
'answer_group_index': 1,
'answers': ['a1', 'a2']
}], 'StateTest1', feconf.TRAINING_JOB_STATUS_NEW, 1)
fs_services.save_classifier_data('TextClassifier', id1,
classifier_data_proto)
self.model_instance_1 = (
classifier_models.ClassifierTrainingJobModel.get_by_id(id1))
self.job_class = (prod_validation_jobs_one_off.
ClassifierTrainingJobModelAuditOneOffJob)
def store_classifier_data(job_id, classifier_data_proto):
"""Checks for the existence of the model and then updates it.
Args:
job_id: str. ID of the ClassifierTrainingJob domain object.
classifier_data_proto: FrozenModel. The frozen model protobuf object
containing result of training job that needs to be stored.
Raises:
Exception. The ClassifierTrainingJobModel corresponding to the job_id
of the ClassifierTrainingJob does not exist.
"""
classifier_training_job_model = (
classifier_models.ClassifierTrainingJobModel.get(job_id, strict=False))
if not classifier_training_job_model:
raise Exception(
'The ClassifierTrainingJobModel corresponding to the job_id of the '
'ClassifierTrainingJob does not exist.')
classifier_training_job = get_classifier_training_job_from_model(
classifier_training_job_model)
classifier_training_job.validate()
fs_services.save_classifier_data(classifier_training_job_model.exp_id,
job_id, classifier_data_proto)
def store_classifier_data(job_id, classifier_data):
"""Checks for the existence of the model and then updates it.
Args:
job_id: str. ID of the ClassifierTrainingJob domain object.
classifier_data: dict. The classification model which needs to be stored
in the job.
Raises:
Exception. The ClassifierTrainingJobModel corresponding to the job_id
of the ClassifierTrainingJob does not exist.
"""
classifier_training_job_model = (
classifier_models.ClassifierTrainingJobModel.get(job_id, strict=False))
if not classifier_training_job_model:
raise Exception(
'The ClassifierTrainingJobModel corresponding to the job_id of the '
'ClassifierTrainingJob does not exist.')
classifier_training_job = get_classifier_training_job_from_model(
classifier_training_job_model)
classifier_training_job.update_classifier_data(classifier_data)
classifier_training_job.validate()
fs_services.save_classifier_data(classifier_training_job_model.exp_id,
job_id, classifier_data)
def test_save_and_get_classifier_data(self):
"""Test that classifier data is stored and retrieved correctly."""
fs_services.save_classifier_data('exp_id', 'job_id',
self.classifier_data)
classifier_data = fs_services.read_classifier_data('exp_id', 'job_id')
self.assertEqual(classifier_data, self.classifier_data)