def test_forward_dense_array_mapper():
mask = np.ones((3, 2), dtype='bool')
map_ = mask_mapper(mask)
# test shape reports
assert_equal(map_.forward1(mask).shape, (6, ))
# test 1sample mapping
assert_array_equal(map_.forward1(np.arange(6).reshape(3, 2)),
[0, 1, 2, 3, 4, 5])
# test 4sample mapping
foursample = map_.forward(np.arange(24).reshape(4, 3, 2))
assert_array_equal(foursample,
[[0, 1, 2, 3, 4, 5], [6, 7, 8, 9, 10, 11],
[12, 13, 14, 15, 16, 17], [18, 19, 20, 21, 22, 23]])
# check incomplete masks
mask[1, 1] = 0
map_ = mask_mapper(mask)
assert_equal(map_.forward1(mask).shape, (5, ))
assert_array_equal(map_.forward1(np.arange(6).reshape(3, 2)),
[0, 1, 2, 4, 5])
# check that it doesn't accept wrong dataspace
assert_raises(ValueError, map_.forward, np.arange(4).reshape(2, 2))
# check fail if neither mask nor shape
assert_raises(ValueError, mask_mapper)
# check that a full mask is automatically created when providing shape
m = mask_mapper(shape=(2, 3, 4))
mp = m.forward1(np.arange(24).reshape(2, 3, 4))
assert_array_equal(mp, np.arange(24))
def test_mapped_classifier(self):
samples = np.array([[0, 0, -1], [1, 0, 1], [-1, -1, 1], [-1, 0, 1],
[1, -1, 1]])
for mask, res in (([1, 1, 0], [1, 1, 1, -1,
-1]), ([1, 0, 1], [-1, 1, -1, -1, 1]),
([0, 1, 1], [-1, 1, -1, 1, -1])):
clf = MappedClassifier(clf=self.clf_sign,
mapper=mask_mapper(
np.array(mask, dtype=bool)))
self.assertEqual(clf.predict(samples), res)
def test_mapped_classifier(self):
samples = np.array([ [ 0, 0, -1], [ 1, 0, 1],
[-1, -1, 1], [-1, 0, 1],
[ 1, -1, 1] ])
for mask, res in (([1, 1, 0], [ 1, 1, 1, -1, -1]),
([1, 0, 1], [-1, 1, -1, -1, 1]),
([0, 1, 1], [-1, 1, -1, 1, -1])):
clf = MappedClassifier(clf=self.clf_sign,
mapper=mask_mapper(np.array(mask,
dtype=bool)))
self.assertEqual(clf.predict(samples), res)
def test_reverse_dense_array_mapper():
mask = np.ones((3, 2), dtype='bool')
mask[1, 1] = 0
map_ = mask_mapper(mask)
rmapped = map_.reverse1(np.arange(1, 6))
assert_equal(rmapped.shape, (3, 2))
assert_equal(rmapped[1, 1], 0)
assert_equal(rmapped[2, 1], 5)
# check that it doesn't accept wrong dataspace
assert_raises(ValueError, map_.forward, np.arange(6))
rmapped2 = map_.reverse(np.arange(1, 11).reshape(2, 5))
assert_equal(rmapped2.shape, (2, 3, 2))
assert_equal(rmapped2[0, 1, 1], 0)
assert_equal(rmapped2[1, 1, 1], 0)
assert_equal(rmapped2[0, 2, 1], 5)
assert_equal(rmapped2[1, 2, 1], 10)
def test_selects():
mask = np.ones((3, 2), dtype='bool')
mask[1, 1] = 0
mask0 = mask.copy()
data = np.arange(6).reshape(mask.shape)
map_ = mask_mapper(mask)
# check if any exception is thrown if we get
# out of the outIds
#assert_raises(IndexError, map_.select_out, [0,1,2,6])
# remove 1,2
map_.append(StaticFeatureSelection([0, 3, 4]))
assert_array_equal(map_.forward1(data), [0, 4, 5])
# remove 1 more
map_.append(StaticFeatureSelection([0, 2]))
assert_array_equal(map_.forward1(data), [0, 5])
# check if original mask wasn't perturbed
assert_array_equal(mask, mask0)
# check if original mask wasn't perturbed
assert_array_equal(mask, mask0)
def from_wizard(cls,
samples,
targets=None,
chunks=None,
mask=None,
mapper=None,
flatten=None,
space=None):
"""Convenience method to create dataset.
Datasets can be created from N-dimensional samples. Data arrays with
more than two dimensions are going to be flattened, while preserving
the first axis (separating the samples) and concatenating all other as
the second axis. Optionally, it is possible to specify targets and
chunk attributes for all samples, and masking of the input data (only
selecting elements corresponding to non-zero mask elements
Parameters
----------
samples : ndarray
N-dimensional samples array. The first axis separates individual
samples.
targets : scalar or ndarray, optional
Labels for all samples. If a scalar is provided its values is assigned
as label to all samples.
chunks : scalar or ndarray, optional
Chunks definition for all samples. If a scalar is provided its values
is assigned as chunk of all samples.
mask : ndarray, optional
The shape of the array has to correspond to the shape of a single
sample (shape(samples)[1:] == shape(mask)). Its non-zero elements
are used to mask the input data.
mapper : Mapper instance, optional
A trained mapper instance that is used to forward-map
possibly already flattened (see flatten) and masked samples
upon construction of the dataset. The mapper must have a
simple feature space (samples x features) as output. Use a
`ChainMapper` to achieve that, if necessary.
flatten : None or bool, optional
If None (default) and no mapper provided, data would get flattened.
Bool value would instruct explicitly either to flatten before
possibly passing into the mapper if no mask is given.
space : str, optional
If provided it is assigned to the mapper instance that performs the
initial flattening of the data.
Returns
-------
instance : Dataset
"""
# for all non-ndarray samples you need to go with the constructor
samples = np.asanyarray(samples)
# compile the necessary samples attributes collection
sa_items = {}
if not targets is None:
sa_items['targets'] = _expand_attribute(targets, samples.shape[0],
'targets')
if not chunks is None:
# unlike previous implementation, we do not do magic to do chunks
# if there are none, there are none
sa_items['chunks'] = _expand_attribute(chunks, samples.shape[0],
'chunks')
# common checks should go into __init__
ds = cls(samples, sa=sa_items)
# apply mask through mapper
if mask is None:
# if we have multi-dim data
if len(samples.shape) > 2 and \
((flatten is None and mapper is None) # auto case
or flatten): # bool case
fm = FlattenMapper(shape=samples.shape[1:], space=space)
ds = ds.get_mapped(fm)
else:
mm = mask_mapper(mask, space=space)
mm.train(ds)
ds = ds.get_mapped(mm)
# apply generic mapper
if not mapper is None:
ds = ds.get_mapped(mapper)
return ds
def from_wizard(cls, samples, targets=None, chunks=None, mask=None,
mapper=None, flatten=None, space=None):
"""Convenience method to create dataset.
Datasets can be created from N-dimensional samples. Data arrays with
more than two dimensions are going to be flattened, while preserving
the first axis (separating the samples) and concatenating all other as
the second axis. Optionally, it is possible to specify targets and
chunk attributes for all samples, and masking of the input data (only
selecting elements corresponding to non-zero mask elements
Parameters
----------
samples : ndarray
N-dimensional samples array. The first axis separates individual
samples.
targets : scalar or ndarray, optional
Labels for all samples. If a scalar is provided its values is assigned
as label to all samples.
chunks : scalar or ndarray, optional
Chunks definition for all samples. If a scalar is provided its values
is assigned as chunk of all samples.
mask : ndarray, optional
The shape of the array has to correspond to the shape of a single
sample (shape(samples)[1:] == shape(mask)). Its non-zero elements
are used to mask the input data.
mapper : Mapper instance, optional
A trained mapper instance that is used to forward-map
possibly already flattened (see flatten) and masked samples
upon construction of the dataset. The mapper must have a
simple feature space (samples x features) as output. Use a
`ChainMapper` to achieve that, if necessary.
flatten : None or bool, optional
If None (default) and no mapper provided, data would get flattened.
Bool value would instruct explicitly either to flatten before
possibly passing into the mapper if no mask is given.
space : str, optional
If provided it is assigned to the mapper instance that performs the
initial flattening of the data.
Returns
-------
instance : Dataset
"""
# for all non-ndarray samples you need to go with the constructor
samples = np.asanyarray(samples)
# compile the necessary samples attributes collection
sa_items = {}
if not targets is None:
sa_items['targets'] = _expand_attribute(targets,
samples.shape[0],
'targets')
if not chunks is None:
# unlike previous implementation, we do not do magic to do chunks
# if there are none, there are none
sa_items['chunks'] = _expand_attribute(chunks,
samples.shape[0],
'chunks')
# common checks should go into __init__
ds = cls(samples, sa=sa_items)
# apply mask through mapper
if mask is None:
# if we have multi-dim data
if len(samples.shape) > 2 and \
((flatten is None and mapper is None) # auto case
or flatten): # bool case
fm = FlattenMapper(shape=samples.shape[1:], space=space)
ds = ds.get_mapped(fm)
else:
mm = mask_mapper(mask, space=space)
mm.train(ds)
ds = ds.get_mapped(mm)
# apply generic mapper
if not mapper is None:
ds = ds.get_mapped(mapper)
return ds
def test_mapper_aliases():
mm = mask_mapper(np.ones((3, 4, 2), dtype='bool'))
assert_array_equal(mm.forward(np.ones((2, 3, 4, 2))),
mm.forward(np.ones((2, 3, 4, 2))))
