def test_kNearestDatasets(self):
kND = KNearestDatasets()
kND.fit(pd.DataFrame([self.krvskp, self.labor]),
self.runs.loc[:, [233, 234]])
neighbor = kND.kNearestDatasets(self.anneal, 1)
self.assertEqual([233], neighbor)
neighbor, distance = kND.kNearestDatasets(self.anneal,
1,
return_distance=True)
self.assertEqual([233], neighbor)
np.testing.assert_array_almost_equal([3.8320802803440586], distance)
neighbors = kND.kNearestDatasets(self.anneal, 2)
self.assertEqual([233, 234], neighbors)
neighbors, distance = kND.kNearestDatasets(self.anneal,
2,
return_distance=True)
self.assertEqual([233, 234], neighbors)
np.testing.assert_array_almost_equal(
[3.8320802803440586, 4.367919719655942], distance)
neighbors = kND.kNearestDatasets(self.anneal, -1)
self.assertEqual([233, 234], neighbors)
neighbors, distance = kND.kNearestDatasets(self.anneal,
-1,
return_distance=True)
self.assertEqual([233, 234], neighbors)
np.testing.assert_array_almost_equal(
[3.8320802803440586, 4.367919719655942], distance)
self.assertRaises(ValueError, kND.kNearestDatasets, self.anneal, 0)
self.assertRaises(ValueError, kND.kNearestDatasets, self.anneal, -2)
def test_kNearestDatasets(self):
kND = KNearestDatasets()
kND.fit(pd.DataFrame([self.krvskp, self.labor]),
self.runs.loc[:, [233, 234]])
neighbor = kND.kNearestDatasets(self.anneal, 1)
self.assertEqual([233], neighbor)
neighbor, distance = kND.kNearestDatasets(self.anneal,
1,
return_distance=True)
self.assertEqual([233], neighbor)
assert_array_almost_equal([1.82298937], distance)
neighbors = kND.kNearestDatasets(self.anneal, 2)
self.assertEqual([233, 234], neighbors)
neighbors, distance = kND.kNearestDatasets(self.anneal,
2,
return_distance=True)
self.assertEqual([233, 234], neighbors)
assert_array_almost_equal([1.822989, 2.267919], distance)
neighbors = kND.kNearestDatasets(self.anneal, -1)
self.assertEqual([233, 234], neighbors)
neighbors, distance = kND.kNearestDatasets(self.anneal,
-1,
return_distance=True)
self.assertEqual([233, 234], neighbors)
assert_array_almost_equal([1.822989, 2.267919], distance)
self.assertRaises(ValueError, kND.kNearestDatasets, self.anneal, 0)
self.assertRaises(ValueError, kND.kNearestDatasets, self.anneal, -2)
def test_random_metric(self):
kND = KNearestDatasets(metric=get_random_metric(random_state=1))
kND.fit(pd.DataFrame([self.krvskp, self.labor]),
self.runs.loc[:, [233, 234]])
distances = []
for i in range(20):
neighbor = kND.kBestSuggestions(self.anneal, 1)
distances.append(neighbor[0][1])
self.assertEqual(len(np.unique(distances)), 20)
def test_random_metric(self):
kND = KNearestDatasets(metric=get_random_metric(random_state=1))
kND.fit(pd.DataFrame([self.krvskp, self.labor]),
self.runs.loc[:,[233, 234]])
distances = []
for i in range(20):
neighbor = kND.kBestSuggestions(self.anneal, 1)
distances.append(neighbor[0][1])
self.assertEqual(len(np.unique(distances)), 20)
def test_fit_l1_distance(self):
kND = KNearestDatasets()
kND.fit(pd.DataFrame([self.anneal, self.krvskp, self.labor]), self.runs)
self.assertEqual(kND.best_configuration_per_dataset[232], 0)
self.assertEqual(kND.best_configuration_per_dataset[233], 1)
self.assertEqual(kND.best_configuration_per_dataset[234], 2)
self.assertTrue((kND.metafeatures ==
pd.DataFrame([self.anneal, self.krvskp, self.labor])).all().all())
def test_scale(self):
kND = KNearestDatasets()
metafeatures = pd.DataFrame([self.anneal, self.krvskp])
metafeatures, other = kND._scale(metafeatures, self.labor)
from pandas.util.testing import assert_series_equal
# Series.equal does not work properly with floats...
assert_series_equal(metafeatures.iloc[0],
pd.Series({"number_of_instances": 0.267919719656,
"number_of_classes": 1,
"number_of_features": 1}))
def test_kNearestDatasets(self):
kND = KNearestDatasets()
kND.fit(pd.DataFrame([self.krvskp, self.labor]),
self.runs.loc[:, [233, 234]])
neighbor = kND.kNearestDatasets(self.anneal, 1)
self.assertEqual([233], neighbor)
neighbor, distance = kND.kNearestDatasets(self.anneal, 1,
return_distance=True)
self.assertEqual([233], neighbor)
assert_array_almost_equal([1.82298937], distance)
neighbors = kND.kNearestDatasets(self.anneal, 2)
self.assertEqual([233, 234], neighbors)
neighbors, distance = kND.kNearestDatasets(self.anneal, 2,
return_distance=True)
self.assertEqual([233, 234], neighbors)
assert_array_almost_equal([1.822989, 2.267919], distance)
neighbors = kND.kNearestDatasets(self.anneal, -1)
self.assertEqual([233, 234], neighbors)
neighbors, distance = kND.kNearestDatasets(self.anneal, -1,
return_distance=True)
self.assertEqual([233, 234], neighbors)
assert_array_almost_equal([1.822989, 2.267919], distance)
self.assertRaises(ValueError, kND.kNearestDatasets, self.anneal, 0)
self.assertRaises(ValueError, kND.kNearestDatasets, self.anneal, -2)
def _learn(self, exclude_double_configurations=True):
dataset_metafeatures, all_other_metafeatures = \
self._split_metafeature_array()
# Remove metafeatures which could not be calculated for the target
# dataset
keep = []
for idx in dataset_metafeatures.index:
if np.isfinite(dataset_metafeatures.loc[idx]):
keep.append(idx)
dataset_metafeatures = dataset_metafeatures.loc[keep]
all_other_metafeatures = all_other_metafeatures.loc[:, keep]
# Do mean imputation of all other metafeatures
all_other_metafeatures = all_other_metafeatures.fillna(
all_other_metafeatures.mean())
if self.kND is None:
# In case that we learn our distance function, get_value the parameters for
# the random forest
if self.distance_kwargs:
rf_params = ast.literal_eval(self.distance_kwargs)
else:
rf_params = None
# To keep the distance the same in every iteration, we create a new
# random state
random_state = sklearn.utils.check_random_state(self.seed)
kND = KNearestDatasets(metric=self.distance,
random_state=random_state,
logger=self.logger,
metric_params=rf_params)
runs = dict()
# TODO move this code to the metabase
for task_id in all_other_metafeatures.index:
try:
runs[task_id] = self.meta_base.get_runs(task_id)
except KeyError:
# TODO should I really except this?
self.logger.info("Could not find runs for instance %s" %
task_id)
runs[task_id] = pd.Series([], name=task_id)
runs = pd.DataFrame(runs)
kND.fit(all_other_metafeatures, runs)
self.kND = kND
return self.kND.kBestSuggestions(
dataset_metafeatures,
k=-1,
exclude_double_configurations=exclude_double_configurations,
)
def _learn(self, exclude_double_configurations=True):
dataset_metafeatures, all_other_metafeatures = \
self._split_metafeature_array()
# Remove metafeatures which could not be calculated for the target
# dataset
keep = []
for idx in dataset_metafeatures.index:
if np.isfinite(dataset_metafeatures.loc[idx]):
keep.append(idx)
dataset_metafeatures = dataset_metafeatures.loc[keep]
all_other_metafeatures = all_other_metafeatures.loc[:,keep]
# Do mean imputation of all other metafeatures
all_other_metafeatures = all_other_metafeatures.fillna(
all_other_metafeatures.mean())
if self.kND is None:
# In case that we learn our distance function, get_value the parameters for
# the random forest
if self.distance_kwargs:
rf_params = ast.literal_eval(self.distance_kwargs)
else:
rf_params = None
# To keep the distance the same in every iteration, we create a new
# random state
random_state = sklearn.utils.check_random_state(self.seed)
kND = KNearestDatasets(metric=self.distance,
random_state=random_state,
metric_params=rf_params)
runs = dict()
# TODO move this code to the metabase
for task_id in all_other_metafeatures.index:
try:
runs[task_id] = self.meta_base.get_runs(task_id)
except KeyError:
# TODO should I really except this?
self.logger.warning("Could not find runs for instance %s" % task_id)
runs[task_id] = pd.Series([], name=task_id)
runs = pd.DataFrame(runs)
kND.fit(all_other_metafeatures, runs)
self.kND = kND
return self.kND.kBestSuggestions(dataset_metafeatures, k=-1,
exclude_double_configurations=exclude_double_configurations)
def test_kBestSuggestions(self):
kND = KNearestDatasets()
kND.fit(pd.DataFrame([self.krvskp, self.labor]),
self.runs.loc[:, [233, 234]])
neighbor = kND.kBestSuggestions(self.anneal, 1)
self.assertEqual([(233, 1.8229893712531495, 1)], neighbor)
neighbors = kND.kBestSuggestions(self.anneal, 2)
self.assertEqual([(233, 1.8229893712531495, 1),
(234, 2.2679197196559415, 2)], neighbors)
neighbors = kND.kBestSuggestions(self.anneal, -1)
self.assertEqual([(233, 1.8229893712531495, 1),
(234, 2.2679197196559415, 2)], neighbors)
self.assertRaises(ValueError, kND.kBestSuggestions, self.anneal, 0)
self.assertRaises(ValueError, kND.kBestSuggestions, self.anneal, -2)
def test_kBestSuggestions(self):
kND = KNearestDatasets()
kND.fit(pd.DataFrame([self.krvskp, self.labor]),
self.runs.loc[:,[233, 234]])
neighbor = kND.kBestSuggestions(self.anneal, 1)
self.assertEqual([(233, 1.8229893712531495, 1)],
neighbor)
neighbors = kND.kBestSuggestions(self.anneal, 2)
self.assertEqual([(233, 1.8229893712531495, 1),
(234, 2.2679197196559415, 2)],
neighbors)
neighbors = kND.kBestSuggestions(self.anneal, -1)
self.assertEqual([(233, 1.8229893712531495, 1),
(234, 2.2679197196559415, 2)],
neighbors)
self.assertRaises(ValueError, kND.kBestSuggestions, self.anneal, 0)
self.assertRaises(ValueError, kND.kBestSuggestions, self.anneal, -2)
def test_kBestSuggestions(self):
kND = KNearestDatasets()
kND.fit(pd.DataFrame([self.krvskp, self.labor]),
self.runs.loc[:, [233, 234]])
neighbor = kND.kBestSuggestions(self.anneal, 1)
np.testing.assert_array_almost_equal(
[(233, 3.8320802803440586, 1)],
neighbor,
)
neighbors = kND.kBestSuggestions(self.anneal, 2)
np.testing.assert_array_almost_equal(
[(233, 3.8320802803440586, 1), (234, 4.367919719655942, 2)],
neighbors,
)
neighbors = kND.kBestSuggestions(self.anneal, -1)
np.testing.assert_array_almost_equal(
[(233, 3.8320802803440586, 1), (234, 4.367919719655942, 2)],
neighbors,
)
self.assertRaises(ValueError, kND.kBestSuggestions, self.anneal, 0)
self.assertRaises(ValueError, kND.kBestSuggestions, self.anneal, -2)
