QuaPy/quapy/tests/test_datasets.py

44 lines
1.6 KiB
Python

import pytest
from quapy.data.datasets import REVIEWS_SENTIMENT_DATASETS, TWITTER_SENTIMENT_DATASETS_TEST, \
TWITTER_SENTIMENT_DATASETS_TRAIN, UCI_DATASETS, fetch_reviews, fetch_twitter, fetch_UCIDataset
@pytest.mark.parametrize('dataset_name', REVIEWS_SENTIMENT_DATASETS)
def test_fetch_reviews(dataset_name):
dataset = fetch_reviews(dataset_name)
print(f'Dataset {dataset_name}')
print('Training set stats')
dataset.training.stats()
print('Test set stats')
dataset.test.stats()
@pytest.mark.parametrize('dataset_name', TWITTER_SENTIMENT_DATASETS_TEST + TWITTER_SENTIMENT_DATASETS_TRAIN)
def test_fetch_twitter(dataset_name):
try:
dataset = fetch_twitter(dataset_name)
except ValueError as ve:
if dataset_name == 'semeval' and ve.args[0].startswith(
'dataset "semeval" can only be used for model selection.'):
dataset = fetch_twitter(dataset_name, for_model_selection=True)
print(f'Dataset {dataset_name}')
print('Training set stats')
dataset.training.stats()
print('Test set stats')
@pytest.mark.parametrize('dataset_name', UCI_DATASETS)
def test_fetch_UCIDataset(dataset_name):
try:
dataset = fetch_UCIDataset(dataset_name)
except FileNotFoundError as fnfe:
if dataset_name == 'pageblocks.5' and fnfe.args[0].find(
'If this is the first time you attempt to load this dataset') > 0:
print('The pageblocks.5 dataset requires some hand processing to be usable, skipping this test.')
return
print(f'Dataset {dataset_name}')
print('Training set stats')
dataset.training.stats()
print('Test set stats')