def test_multiple_algorithms_engine():
engine = Engine({
'datasource': TestEmptyDataSource,
'algorithms': {
'algo1': {
'class': TestMultiAlgorithm,
'params': {
'p': 'A'
},
},
'algo2': {
'class': TestMultiAlgorithm,
'params': {
'p': 'B'
},
},
},
'serving': TestIdentityServing,
})
models = {
'algo1': object(),
'algo2': object(),
}
nt.assert_equals(engine.predict(models, None), {
'algo1': 'A',
'algo2': 'B',
})
def test_setup():
make_test_data()
engine = Engine({
'datasource': {
'class': TestDataSource,
'params': {
'csv': test_data_file,
},
},
'preparator': TestPreparator,
'algorithm': {
'class': TestAlgorithm,
'params': {
'model.csv': '~/.tidml/tests/model2.csv', # custom
},
},
})
engine.train()
def main():
engine = Engine({'config': 'examples/dase/house_prices/config.yaml'})
engine.train()
models = engine.load_models()
engine.predict(models,
[0.07, 0.99, 0.0, 0.51, 0.69, 0.77, 0.77, 0.75, 0.44])
def create():
return Engine({
"datasource": MyDataSource,
"algorithm": {
"class": MyAlgorithm,
"params": {
"model.pickle": "~/.tidml/hello_world/model.pkl"
}
}
})
def test_insane_datasource():
engine = Engine({
'datasource': {
'class': TestDataSource,
'params': {
'csv': test_data_file,
'insane': True,
},
},
})
nt.assert_raises_regexp(RuntimeError, 'training_data insane!',
engine.train)
def test_insane_algorithm():
engine = Engine({
'datasource': {
'class': TestDataSource,
'params': {
'csv': test_data_file,
},
},
'preparator': TestPreparator,
'algorithm': {
'class': TestAlgorithm,
'params': {
'model.csv': '~/.tidml/tests/model2.csv', # custom
'insane': True,
},
},
})
nt.assert_raises_regexp(RuntimeError, 'model insane!', engine.train)
def test_simple_engine():
engine = Engine({
'datasource': {
'class': TestDataSource,
'params': {
'csv': test_data_file,
},
},
'algorithm': {
'class': TestSimpleAlgorithm,
'params': {
'model.pickle': '~/.tidml/tests/model.pkl', # default built-in
},
},
})
engine.train()
models = engine.load_models()
prediction = engine.predict(models, 3)
nt.assert_equals(prediction, 6)
def create():
return Engine({'config': 'examples/dase/gnb_classifier/config.yaml'})
def create():
return Engine({'config': 'examples/dase/regression/config.yaml'})