def test_create_userdata_w_profile():
randomstr = 'test-' + create_jobid()
s3 = boto3.client('s3')
s3.put_object(Body='haha'.encode('utf-8'),
Bucket='tibanna-output',
Key=randomstr)
input_dict = {
'args': {
'input_files': {
'input_file': {
'bucket_name': 'tibanna-output',
'object_key': randomstr
}
},
'output_S3_bucket': 'somebucket',
'app_name': 'md5',
'cwl_main_filename': 'md5.cwl',
'cwl_directory_url': 'someurl'
},
'config': {
'log_bucket': 'tibanna-output'
},
'jobid': 'myjobid'
}
execution = Execution(input_dict)
profile = {'access_key': 'haha', 'secret_key': 'lala'}
userdata = execution.create_userdata(profile=profile)
print(userdata)
assert userdata
assert '-a haha -s lala' in userdata
# cleanup afterwards
s3.delete_objects(Bucket='tibanna-output',
Delete={'Objects': [{
'Key': randomstr
}]})
def test_launch_args():
"""test creating launch arguments - also test spot_instance"""
jobid = create_jobid()
log_bucket = 'tibanna-output'
input_dict = {
'args': {
'output_S3_bucket': 'somebucket',
'cwl_main_filename': 'md5.cwl',
'cwl_directory_url': 'someurl'
},
'config': {
'log_bucket': log_bucket,
'mem': 1,
'cpu': 1,
'spot_instance': True
},
'jobid': jobid
}
execution = Execution(input_dict)
# userdata is required before launch_args is created
execution.userdata = execution.create_userdata()
launch_args = execution.launch_args
print(launch_args)
assert launch_args
assert 't3.micro' in str(launch_args)
assert 'InstanceMarketOptions' in str(launch_args)
def test_create_run_json_dict():
randomstr = 'test-' + create_jobid()
s3 = boto3.client('s3')
s3.put_object(Body='haha'.encode('utf-8'),
Bucket='tibanna-output',
Key=randomstr)
input_dict = {
'args': {
'input_files': {
'input_file': {
'bucket_name': 'tibanna-output',
'object_key': randomstr
}
},
'output_S3_bucket': 'somebucket',
'app_name': 'md5',
'cwl_main_filename': 'md5.cwl',
'cwl_directory_url': 'someurl'
},
'config': {
'log_bucket': 'tibanna-output'
}
}
execution = Execution(input_dict)
runjson = execution.create_run_json_dict()
assert runjson
# cleanup afterwards
s3.delete_objects(Bucket='tibanna-output',
Delete={'Objects': [{
'Key': randomstr
}]})
def test_ec2_exception_coordinator9():
"""ec2 exceptions with 'other_instance_types' with both instance_type and mem/cpu
specified"""
jobid = create_jobid()
log_bucket = 'tibanna-output'
input_dict = {
'args': {
'output_S3_bucket': 'somebucket',
'cwl_main_filename': 'md5.cwl',
'cwl_directory_url': 'someurl'
},
'config': {
'log_bucket': log_bucket,
'mem': 2,
'cpu': 1,
'behavior_on_capacity_limit': 'other_instance_types'
},
'jobid': jobid
}
execution = Execution(input_dict, dryrun=True)
assert execution.cfg.instance_type == 't3.small'
execution.userdata = execution.create_userdata()
res = execution.ec2_exception_coordinator(fun)()
assert res == 'continue'
assert execution.cfg.instance_type == 't2.small'
def test_upload_run_json():
jobid = create_jobid()
log_bucket = 'tibanna-output'
input_dict = {
'args': {
'output_S3_bucket': 'somebucket',
'cwl_main_filename': 'md5.cwl',
'cwl_directory_url': 'someurl'
},
'config': {
'log_bucket': log_bucket,
'mem': 1,
'cpu': 1
},
'jobid': jobid
}
somejson = {'haha': 'lala'}
execution = Execution(input_dict)
execution.upload_run_json(somejson)
s3 = boto3.client('s3')
res = s3.get_object(Bucket=log_bucket, Key=jobid + '.run.json')
assert res
# clean up afterwards
s3.delete_objects(Bucket=log_bucket,
Delete={'Objects': [{
'Key': jobid + '.run.json'
}]})
def test_check_dependency_failed():
job_statuses = {'running_jobs': ['jid1'], 'failed_jobs': ['jid2']}
with mock.patch('tibanna.job.Jobs.status', return_value=job_statuses):
with pytest.raises(DependencyFailedException) as exec_info:
Execution.check_dependency(job_id=['jid1', 'jid2'])
assert 'failed' in str(exec_info.value)
assert 'jid1' not in str(exec_info.value)
assert 'jid2' in str(exec_info.value)
def test_get_input_size_in_bytes_with_secondary_files():
randomstr, randomstr_1, randomstr_2 = 'test-' + create_jobid(
), 'test-' + create_jobid(), 'test-' + create_jobid()
s3 = boto3.client('s3')
s3.put_object(Body='haha'.encode('utf-8'),
Bucket='tibanna-output',
Key=randomstr)
s3.put_object(Body='fooooooo'.encode('utf-8'),
Bucket='tibanna-output',
Key=randomstr_1)
s3.put_object(Body='pippo'.encode('utf-8'),
Bucket='tibanna-output',
Key=randomstr_2)
input_dict = {
'args': {
'input_files': {
'input_file': {
'bucket_name': 'tibanna-output',
'object_key': randomstr
}
},
'secondary_files': {
'input_file': {
'bucket_name': 'tibanna-output',
'object_key': [randomstr_1, randomstr_2]
}
},
'output_S3_bucket': 'somebucket',
'app_name': 'md5',
'cwl_main_filename': 'md5.cwl',
'cwl_directory_url': 'someurl'
},
'config': {
'log_bucket': 'tibanna-output'
}
}
execution = Execution(input_dict)
execution.input_size_in_bytes = execution.get_input_size_in_bytes()
assert execution.total_input_size_in_gb == 1.5832483768463135E-8
# cleanup afterwards
s3.delete_objects(Bucket='tibanna-output',
Delete={
'Objects': [{
'Key': randomstr
}, {
'Key': randomstr_1
}, {
'Key': randomstr_2
}]
})
def test_execution_mem_cpu():
"""mem and cpu are provided but not app_name or instance_type,
which should be fine.
language is snakemake this time"""
input_dict = {
'args': {
'input_files': {},
'language': 'snakemake',
'output_S3_bucket': 'somebucket',
'snakemake_main_filename': 'Snakefile',
'snakemake_directory_url': 'someurl',
'command': 'snakemake',
'container_image': 'quay.io/snakemake/snakemake'
},
'config': {
'log_bucket': 'tibanna-output',
'mem': 1,
'cpu': 1
}
}
execution = Execution(input_dict)
unicorn_dict = execution.input_dict
assert len(execution.instance_type_list) == 10
assert 'args' in unicorn_dict
assert 'config' in unicorn_dict
assert 'instance_type' in unicorn_dict['config']
assert unicorn_dict['config']['instance_type'] == 't3.micro'
def test_execution_benchmark_app_name_but_w_nonmatching_inputarg(
run_task_awsem_event_wdl_md5_benchmark_error):
cfg = Config(**run_task_awsem_event_wdl_md5_benchmark_error['config'])
assert cfg.use_benchmark == True
with pytest.raises(Exception) as exec_info:
execution = Execution(run_task_awsem_event_wdl_md5_benchmark_error)
assert 'Benchmarking not working' in str(exec_info.value)
def test_launch_and_get_instance_id():
"""test dryrun of ec2 launch"""
jobid = create_jobid()
log_bucket = 'tibanna-output'
input_dict = {
'args': {
'output_S3_bucket': 'somebucket',
'cwl_main_filename': 'md5.cwl',
'cwl_directory_url': 'someurl'
},
'config': {
'log_bucket': log_bucket,
'mem': 1,
'cpu': 1,
'spot_instance': True
},
'jobid': jobid
}
execution = Execution(input_dict, dryrun=True)
# userdata is required before launch_args is created
execution.userdata = execution.create_userdata()
with pytest.raises(Exception) as ex:
execution.launch_and_get_instance_id()
assert 'Request would have succeeded, but DryRun flag is set' in str(
ex.value)
def test_update_config_ebs_size2():
"""ebs_size is given as the 'x' format. The total estimated ebs_size is larger than 10"""
randomstr = 'test-' + create_jobid()
s3 = boto3.client('s3')
s3.put_object(Body='haha'.encode('utf-8'),
Bucket='tibanna-output',
Key=randomstr)
input_dict = {
'args': {
'input_files': {
'input_file': {
'bucket_name': 'tibanna-output',
'object_key': randomstr
}
},
'output_S3_bucket': 'somebucket',
'app_name': 'md5',
'cwl_main_filename': 'md5.cwl',
'cwl_directory_url': 'someurl'
},
'config': {
'log_bucket': 'tibanna-output',
'ebs_size': '5000000000x'
}
}
execution = Execution(input_dict)
execution.input_size_in_bytes = execution.get_input_size_in_bytes()
execution.update_config_ebs_size()
assert execution.cfg.ebs_size == 19
# cleanup afterwards
s3.delete_objects(Bucket='tibanna-output',
Delete={'Objects': [{
'Key': randomstr
}]})
def test_ec2_exception_coordinator6():
"""ec2 exceptions with 'retry_without_spot'"""
jobid = create_jobid()
log_bucket = 'tibanna-output'
input_dict = {
'args': {
'output_S3_bucket': 'somebucket',
'cwl_main_filename': 'md5.cwl',
'cwl_directory_url': 'someurl'
},
'config': {
'log_bucket': log_bucket,
'instance_type': 't2.micro',
'spot_instance': True,
'behavior_on_capacity_limit': 'retry_without_spot'
},
'jobid': jobid
}
execution = Execution(input_dict, dryrun=True)
execution.userdata = execution.create_userdata()
res = execution.ec2_exception_coordinator(fun)()
assert res == 'continue'
assert execution.cfg.spot_instance is False # changed to non-spot
assert execution.cfg.behavior_on_capacity_limit == 'fail' # changed to non-spot
with pytest.raises(EC2InstanceLimitException) as exec_info:
res = execution.ec2_exception_coordinator(fun)() # this time, it fails
assert exec_info
def test_execution_missing_field6():
"""language is shell but container_image is missing"""
input_dict = {
'args': {
'input_files': {},
'language': 'shell',
'output_S3_bucket': 'somebucket',
'command': 'some command'
},
'config': {
'log_bucket': 'tibanna-output',
'mem': 1,
'cpu': 1
}
}
with pytest.raises(MissingFieldInInputJsonException) as ex:
Execution(input_dict)
assert ex
assert 'container_image' in str(ex.value)
def test_execution_missing_field5():
"""language is snakemake but command is missing"""
input_dict = {
'args': {
'input_files': {},
'language': 'snakemake',
'output_S3_bucket': 'somebucket',
'snakemake_main_filename': 'Snakefile',
'snakemake_directory_url': 'someurl',
'container_image': 'quay.io/snakemake/snakemake'
},
'config': {
'log_bucket': 'tibanna-output',
'mem': 1,
'cpu': 1
}
}
with pytest.raises(MissingFieldInInputJsonException) as ex:
Execution(input_dict)
assert ex
assert 'command' in str(ex.value)
def test_execution_benchmark():
randomstr = 'test-' + create_jobid()
s3 = boto3.client('s3')
s3.put_object(Body='haha'.encode('utf-8'),
Bucket='tibanna-output',
Key=randomstr)
input_dict = {
'args': {
'input_files': {
'input_file': {
'bucket_name': 'tibanna-output',
'object_key': randomstr
}
},
'output_S3_bucket': 'somebucket',
'app_name': 'md5',
'cwl_main_filename': 'md5.cwl',
'cwl_directory_url': 'someurl'
},
'config': {
'log_bucket': 'tibanna-output'
}
}
execution = Execution(input_dict)
unicorn_dict = execution.input_dict
print(unicorn_dict)
assert 'args' in unicorn_dict
assert 'config' in unicorn_dict
assert 'instance_type' in unicorn_dict['config']
assert unicorn_dict['config']['instance_type'] == 't3.micro'
assert unicorn_dict['config']['ebs_size'] == 10
# cleanup afterwards
s3.delete_objects(Bucket='tibanna-output',
Delete={'Objects': [{
'Key': randomstr
}]})
def test_ec2_exception_coordinator2():
"""ec2 limit exceptions with 'fail'"""
jobid = create_jobid()
log_bucket = 'tibanna-output'
input_dict = {
'args': {
'output_S3_bucket': 'somebucket',
'cwl_main_filename': 'md5.cwl',
'cwl_directory_url': 'someurl'
},
'config': {
'log_bucket': log_bucket,
'instance_type': 'c5.4xlarge',
'spot_instance': True
},
'jobid': jobid
}
execution = Execution(input_dict, dryrun=True)
execution.userdata = execution.create_userdata()
with pytest.raises(EC2InstanceLimitException) as exec_info:
execution.ec2_exception_coordinator(fun)()
assert exec_info
def test_ec2_exception_coordinator7():
"""ec2 exceptions with 'retry_without_spot' without spot instance"""
jobid = create_jobid()
log_bucket = 'tibanna-output'
input_dict = {
'args': {
'output_S3_bucket': 'somebucket',
'cwl_main_filename': 'md5.cwl',
'cwl_directory_url': 'someurl'
},
'config': {
'log_bucket': log_bucket,
'instance_type': 't2.micro',
'behavior_on_capacity_limit': 'retry_without_spot'
},
'jobid': jobid
}
execution = Execution(input_dict, dryrun=True)
assert execution.cfg.spot_instance is False
execution.userdata = execution.create_userdata()
with pytest.raises(Exception) as exec_info:
execution.ec2_exception_coordinator(fun)()
assert "'retry_without_spot' works only with 'spot_instance'" in str(
exec_info.value)
def test_ec2_exception_coordinator5():
"""ec2 exceptions with 'other_instance_types' but had only one option"""
jobid = create_jobid()
log_bucket = 'tibanna-output'
input_dict = {
'args': {
'output_S3_bucket': 'somebucket',
'cwl_main_filename': 'md5.cwl',
'cwl_directory_url': 'someurl'
},
'config': {
'log_bucket': log_bucket,
'instance_type': 't2.micro',
'spot_instance': True,
'behavior_on_capacity_limit': 'other_instance_types'
},
'jobid': jobid
}
execution = Execution(input_dict, dryrun=True)
assert execution.cfg.instance_type == 't2.micro'
execution.userdata = execution.create_userdata()
with pytest.raises(EC2InstanceLimitException) as exec_info:
execution.ec2_exception_coordinator(fun)()
assert 'No more instance type available' in str(exec_info.value)
def test_execution_benchmark_app_name_but_w_instance_type_ebs(
run_task_awsem_event_wdl_md5_do_not_use_benchmark):
execution = Execution(run_task_awsem_event_wdl_md5_do_not_use_benchmark)
