def test_header_only(live_server):
ui = PickableMock()
base_url = '{webhost}/api/v1/'.format(webhost=live_server.url())
with pytest.raises(ValueError) as ctx:
run_batch_predictions(base_url=base_url,
base_headers={},
user='******',
pwd='password',
api_token=None,
create_api_token=False,
pid='56dd9570018e213242dfa93c',
lid='56dd9570018e213242dfa93d',
import_id=None,
n_retry=3,
concurrent=1,
resume=False,
n_samples=10,
out_file='out.csv',
keep_cols=None,
delimiter=',',
dataset='tests/fixtures/header_only.csv',
pred_name=None,
timeout=30,
ui=ui,
auto_sample=False,
fast_mode=False,
dry_run=False,
encoding='',
skip_dialect=False)
assert str(ctx.value) == ("Input file 'tests/fixtures/header_only.csv' "
"is empty.")
def test_explicit_delimiter_gzip(live_server):
ui = PickableMock()
base_url = '{webhost}/api/v1/'.format(webhost=live_server.url())
ret = run_batch_predictions(
base_url=base_url,
base_headers={},
user='******',
pwd='password',
api_token=None,
create_api_token=False,
pid='56dd9570018e213242dfa93c',
lid='56dd9570018e213242dfa93d',
import_id=None,
n_retry=3,
concurrent=1,
resume=False,
n_samples=10,
out_file='out.csv',
keep_cols=None,
delimiter=',',
dataset='tests/fixtures/temperatura_predict.csv.gz',
pred_name=None,
timeout=30,
ui=ui,
auto_sample=False,
fast_mode=False,
dry_run=False,
encoding='',
skip_dialect=False)
assert ret is None
def test_bad_newline(live_server):
ui = PickableMock()
base_url = '{webhost}/predApi/v1.0/'.format(webhost=live_server.url())
run_batch_predictions(base_url=base_url,
base_headers={},
user='******',
pwd='password',
api_token=None,
create_api_token=False,
pid='56dd9570018e213242dfa93c',
lid='56dd9570018e213242dfa93d',
import_id=None,
n_retry=3,
concurrent=1,
resume=False,
n_samples=10,
out_file='out.csv',
keep_cols=None,
delimiter=',',
dataset='tests/fixtures/diabetes_bad_newline.csv',
pred_name=None,
timeout=None,
ui=ui,
auto_sample=False,
fast_mode=False,
dry_run=False,
encoding='',
skip_dialect=False)
lines = len(open('out.csv', 'rb').readlines())
assert lines == 5
ui.warning.assert_any_call('Detected empty rows in the CSV file. '
'These rows will be discarded.')
def test_create_and_acquire_api_token(live_server):
ui = PickableMock()
base_url = '{webhost}/api/v1/'.format(webhost=live_server.url())
ret = acquire_api_token(base_url, {}, 'username', 'password', True, ui)
assert ret == 'Som3tok3n'
ui.info.assert_called_with('api-token acquired')
ui.debug.assert_called_with('api-token: Som3tok3n')
def test_acquire_api_token_bad_status(live_server):
ui = PickableMock()
base_url = '{webhost}/api/v1/'.format(webhost=live_server.url())
with pytest.raises(ValueError) as ctx:
acquire_api_token(base_url, {}, 'bad_status', 'passwd', False, ui)
assert not ui.info.called
assert not ui.debug.called
assert str(ctx.value) == 'api_token request returned status code 500'
def test_acquire_api_token_unauthorized(live_server):
ui = PickableMock()
base_url = '{webhost}/api/v1/'.format(webhost=live_server.url())
with pytest.raises(ValueError) as ctx:
acquire_api_token(base_url, {}, 'unknown', 'passwd', False, ui)
assert not ui.info.called
assert not ui.debug.called
assert str(ctx.value) == 'wrong credentials'
def test_acquire_api_token_no_token2(live_server):
ui = PickableMock()
base_url = '{webhost}/api/v1/'.format(webhost=live_server.url())
with pytest.raises(ValueError) as ctx:
acquire_api_token(base_url, {}, 'no_token2', 'passwd', False, ui)
assert not ui.info.called
assert not ui.debug.called
assert str(ctx.value) == ('no api-token registered; '
'please run with --create_api_token flag.')
def test_quotechar_in_keep_cols(live_server):
base_url = '{webhost}/predApi/v1.0/'.format(webhost=live_server.url())
ui = PickableMock()
with tempfile.NamedTemporaryFile(prefix='test_',
suffix='.csv',
delete=False) as fd:
head = open("tests/fixtures/quotes_input_head.csv", "rb").read()
body_1 = open("tests/fixtures/quotes_input_first_part.csv",
"rb").read()
body_2 = open("tests/fixtures/quotes_input_bad_part.csv", "rb").read()
fd.file.write(head)
size = 0
while size < DETECT_SAMPLE_SIZE_SLOW:
fd.file.write(body_1)
size += len(body_1)
fd.file.write(body_2)
fd.close()
ret = run_batch_predictions(base_url=base_url,
base_headers={},
user='******',
pwd='password',
api_token=None,
create_api_token=False,
pid='56dd9570018e213242dfa93c',
lid='56dd9570018e213242dfa93d',
import_id=None,
n_retry=3,
concurrent=1,
resume=False,
n_samples=10,
out_file='out.csv',
keep_cols=["b", "c"],
delimiter=None,
dataset=fd.name,
pred_name=None,
timeout=None,
ui=ui,
auto_sample=True,
fast_mode=False,
dry_run=False,
encoding='',
skip_dialect=False)
assert ret is None
last_line = open("out.csv", "rb").readlines()[-1]
expected_last_line = b'1044,2,"eeeeeeee ""eeeeee"" eeeeeeeeeeee'
assert last_line[:len(expected_last_line)] == expected_last_line
def check_regression_jp(live_server, tmpdir, fast_mode, gzipped):
"""Use utf8 encoded input data.
"""
if fast_mode:
out_fname = 'out_fast.csv'
else:
out_fname = 'out.csv'
out = tmpdir.join(out_fname)
dataset_suffix = '.gz' if gzipped else ''
ui = PickableMock()
base_url = '{webhost}/predApi/v1.0/'.format(webhost=live_server.url())
ret = run_batch_predictions(
base_url=base_url,
base_headers={},
user='******',
pwd='password',
api_token=None,
create_api_token=False,
pid='56dd9570018e213242dfa93c',
lid='56dd9570018e213242dfa93e',
import_id=None,
n_retry=3,
concurrent=1,
resume=False,
n_samples=500,
out_file=str(out),
keep_cols=None,
delimiter=None,
dataset='tests/fixtures/regression_jp.csv' + dataset_suffix,
pred_name='new_name',
timeout=None,
ui=ui,
auto_sample=False,
fast_mode=fast_mode,
dry_run=False,
encoding='',
skip_dialect=False,
compression=True
)
assert ret is None
actual = out.read_text('utf-8')
with open('tests/fixtures/regression_output_jp.csv', 'rU') as f:
assert actual == f.read()
def test_pred_threshold_classification(live_server, tmpdir, func_params):
# train one model in project
out = tmpdir.join('out.csv')
ui = PickableMock()
base_url = '{webhost}/predApi/v1.0/'.format(webhost=live_server.url())
ret = run_batch_predictions(
base_url=base_url,
base_headers={},
user='******',
pwd='password',
api_token=None,
create_api_token=False,
deployment_id=func_params['deployment_id'],
pid=func_params['pid'],
lid=func_params['lid'],
import_id=None,
n_retry=3,
concurrent=1,
resume=False,
n_samples=10,
out_file=str(out),
keep_cols=None,
delimiter=None,
dataset='tests/fixtures/temperatura_predict.csv',
pred_name='healthy',
pred_threshold_name='threshold',
timeout=None,
ui=ui,
auto_sample=False,
fast_mode=False,
dry_run=False,
encoding='',
skip_dialect=False
)
assert ret is None
expected = out.read_text('utf-8')
with open(
'tests/fixtures/temperatura_output_healthy_threshold.csv', 'rU'
) as f:
assert expected == f.read(), expected
def test_prediction_explanations_keepcols(live_server, tmpdir):
# train one model in project
out = tmpdir.join('out.csv')
ui = PickableMock()
base_url = '{webhost}/predApi/v1.0/'.format(webhost=live_server.url())
ret = run_batch_predictions(
base_url=base_url,
base_headers={},
user='******',
pwd='password',
api_token=None,
create_api_token=False,
pid='5afb150782c7dd45fcc03951',
lid='5b2cad28aa1d12847310acf4',
import_id=None,
n_retry=3,
concurrent=1,
resume=False,
n_samples=10,
out_file=str(out),
keep_cols=['medical_specialty', 'number_diagnoses'],
delimiter=None,
dataset='tests/fixtures/10kDiabetes.csv',
pred_name=None,
pred_threshold_name=None,
pred_decision_name=None,
timeout=None,
ui=ui,
auto_sample=False,
fast_mode=False,
dry_run=False,
encoding='',
skip_dialect=False,
max_prediction_explanations=5
)
assert ret is None
actual = out.read_text('utf-8')
file_path = 'tests/fixtures/10kDiabetes_5explanations_keepcols.csv'
with open(file_path, 'rU') as f:
expected = f.read()
assert str(actual) == str(expected), expected
def test_simple_transferable(live_server, tmpdir):
# train one model in project
out = tmpdir.join('out.csv')
ui = PickableMock()
base_url = '{webhost}/predApi/v1.0/'.format(webhost=live_server.url())
ret = run_batch_predictions(
base_url=base_url,
base_headers={},
user='******',
pwd='password',
api_token=None,
create_api_token=False,
import_id='0ec5bcea7f0f45918fa88257bfe42c09',
pid=None,
lid=None,
n_retry=3,
concurrent=1,
resume=False,
n_samples=10,
out_file=str(out),
keep_cols=None,
delimiter=None,
dataset='tests/fixtures/regression_predict.csv',
pred_name=None,
pred_threshold_name=None,
pred_decision_name=None,
timeout=None,
ui=ui,
auto_sample=False,
fast_mode=False,
dry_run=False,
encoding='',
skip_dialect=False
)
assert ret is None
actual = out.read_text('utf-8')
with open('tests/fixtures/regression_output.csv', 'rU') as f:
expected = f.read()
assert str(actual) == str(expected), expected
def test_simple_with_unicode(live_server, tmpdir, func_params, dataset_name):
# train one model in project
out = tmpdir.join('out.csv')
ui = PickableMock()
base_url = '{webhost}/predApi/v1.0/'.format(webhost=live_server.url())
ret = run_batch_predictions(
base_url=base_url,
base_headers={},
user='******',
pwd='password',
api_token=None,
create_api_token=False,
deployment_id=func_params['deployment_id'],
pid=func_params['pid'],
lid=func_params['lid'],
import_id=None,
n_retry=3,
concurrent=1,
resume=False,
n_samples=10,
out_file=str(out),
keep_cols=None,
delimiter=None,
dataset='tests/fixtures/{}'.format(dataset_name),
pred_name=None,
pred_threshold_name=None,
pred_decision_name=None,
timeout=None,
ui=ui,
auto_sample=False,
fast_mode=False,
dry_run=False,
encoding='',
skip_dialect=False)
assert ret is None
actual = out.read_text('utf-8')
with open('tests/fixtures/jpReview_books_reg_out.csv', 'rU') as f:
expected = f.read()
assert str(actual) == str(expected), expected
def test_simple_api_v1(live_server, tmpdir):
# train one model in project
out = tmpdir.join('out.csv')
ui = PickableMock()
base_url = '{webhost}/api/v1/'.format(webhost=live_server.url())
ret = run_batch_predictions(
base_url=base_url,
base_headers={},
user='******',
pwd='password',
api_token=None,
create_api_token=False,
pid='56dd9570018e213242dfa93c',
lid='56dd9570018e213242dfa93f',
import_id=None,
n_retry=3,
concurrent=1,
resume=False,
n_samples=10,
out_file=str(out),
keep_cols=None,
delimiter=None,
dataset='tests/fixtures/temperatura_predict.csv.gz',
pred_name=None,
timeout=None,
ui=ui,
auto_sample=False,
fast_mode=False,
dry_run=False,
encoding='',
skip_dialect=False
)
assert ret is None
actual = out.read_text('utf-8')
with open('tests/fixtures/temperatura_api_v1_output.csv', 'rU') as f:
expected = f.read()
assert str(actual) == str(expected), expected
def test_simple_with_wrong_encoding(live_server, tmpdir, func_params):
out = tmpdir.join('out.csv')
ui = PickableMock()
base_url = '{webhost}/predApi/v1.0/'.format(webhost=live_server.url())
with pytest.raises(UnicodeDecodeError) as execinfo:
run_batch_predictions(base_url=base_url,
base_headers={},
user='******',
pwd='password',
api_token=None,
create_api_token=False,
deployment_id=func_params['deployment_id'],
pid=func_params['pid'],
lid=func_params['lid'],
import_id=None,
n_retry=3,
concurrent=1,
resume=False,
n_samples=10,
out_file=str(out),
keep_cols=None,
delimiter=None,
dataset='tests/fixtures/jpReview_books_reg.csv',
pred_name=None,
pred_threshold_name=None,
pred_decision_name=None,
timeout=None,
ui=ui,
auto_sample=False,
fast_mode=False,
dry_run=False,
encoding='cp932',
skip_dialect=False)
# Fixture dataset encoding 'utf-8' and we trying to decode it with 'cp932'
assert "'cp932' codec can't decode byte" in str(execinfo.value)