/usr/share/pyshared/mvpa2/datasets/base.py is in python-mvpa2 2.2.0-4ubuntu2.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 | # emacs: -*- mode: python; py-indent-offset: 4; indent-tabs-mode: nil -*-
# vi: set ft=python sts=4 ts=4 sw=4 et:
### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ##
#
# See COPYING file distributed along with the PyMVPA package for the
# copyright and license terms.
#
### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ### ##
"""PyMVPA's common Dataset container."""
__docformat__ = 'restructuredtext'
import numpy as np
import copy
from mvpa2.base import warning
from mvpa2.base.collections import SampleAttributesCollection, \
FeatureAttributesCollection, DatasetAttributesCollection, \
SampleAttribute, FeatureAttribute, DatasetAttribute
from mvpa2.base.dataset import AttrDataset
from mvpa2.base.dataset import _expand_attribute
from mvpa2.misc.support import idhash as idhash_
from mvpa2.mappers.base import ChainMapper
from mvpa2.featsel.base import StaticFeatureSelection
from mvpa2.mappers.flatten import mask_mapper, FlattenMapper
if __debug__:
from mvpa2.base import debug
class Dataset(AttrDataset):
__doc__ = AttrDataset.__doc__
def get_mapped(self, mapper):
"""Feed this dataset through a trained mapper (forward).
Parameters
----------
mapper : Mapper
This mapper instance has to be trained.
Returns
-------
Dataset
The forward-mapped dataset.
"""
# if we use .forward, no postcall is called... is that
# desired? doesn't seem to have major impact on unittests
# BUT since postcall might change dimensionality/meaning of
# data, it would not be any longer reversible; more over
# since chain of .forwards do not invoke postcalls, also
# forward would lead to different behavior
#mds = mapper(self)
mds = mapper.forward(self)
mds._append_mapper(mapper)
return mds
def _append_mapper(self, mapper):
if not 'mapper' in self.a:
self.a['mapper'] = mapper
return
pmapper = self.a.mapper
# otherwise we have a mapper already, but is it a chain?
if not isinstance(pmapper, ChainMapper):
self.a.mapper = ChainMapper([pmapper])
# is a chain mapper
# merge slicer?
lastmapper = self.a.mapper[-1]
if isinstance(lastmapper, StaticFeatureSelection):
try:
# try whether mappers can be merged
lastmapper += mapper
except TypeError:
# append new one if not
self.a.mapper.append(mapper)
else:
self.a.mapper.append(mapper)
def __getitem__(self, args):
# uniformize for checks below; it is not a tuple if just single slicing
# spec is passed
if not isinstance(args, tuple):
args = (args,)
# if we get an slicing array for feature selection and it is *not* 1D
# try feeding it through the mapper (if there is any)
if len(args) > 1 and isinstance(args[1], np.ndarray) \
and len(args[1].shape) > 1 \
and self.a.has_key('mapper'):
args = list(args)
args[1] = self.a.mapper.forward1(args[1])
args = tuple(args)
# let the base do the work
ds = super(Dataset, self).__getitem__(args)
# and adjusting the mapper (if any)
if len(args) > 1 and 'mapper' in ds.a:
# create matching mapper
# the mapper is just appended to the dataset. It could also be
# actually used to perform the slicing and prevent duplication of
# functionality between the Dataset.__getitem__ and the mapper.
# However, __getitem__ is sometimes more efficient, since it can
# slice samples and feature axis at the same time. Moreover, the
# mvpa2.base.dataset.Dataset has no clue about mappers and should
# be fully functional without them.
subsetmapper = StaticFeatureSelection(args[1],
dshape=self.samples.shape[1:])
# do not-act forward mapping to charge the output shape of the
# slice mapper without having it to train on a full dataset (which
# is most likely more expensive)
subsetmapper.forward(np.zeros((1,) + self.shape[1:], dtype='bool'))
# mapper is ready to use -- simply store
ds._append_mapper(subsetmapper)
return ds
def find_collection(self, attr):
"""Lookup collection that contains an attribute of a given name.
Collections are search in the following order: sample attributes,
feature attributes, dataset attributes. The first collection
containing a matching attribute is returned.
Parameters
----------
attr : str
Attribute name to be looked up.
Returns
-------
Collection
If not matching collection is found a LookupError exception is raised.
"""
if attr in self.sa:
col = self.sa
if __debug__ and (attr in self.fa or attr in self.a):
warning("An attribute with name '%s' is also present "
"in another attribute collection (fa=%s, a=%s) -- make "
"sure that you got the right one (see ``col`` "
"argument)." % (attr, attr in self.fa, attr in self.a))
elif attr in self.fa:
col = self.fa
if __debug__ and attr in self.a:
warning("An attribute with name '%s' is also present "
"in the dataset attribute collection -- make sure "
"that you got the right one (see ``col`` argument)."
% (attr,))
elif attr in self.a:
col = self.a
# we don't need to warn here, since it wouldn't happen
else:
raise LookupError("Cannot find '%s' attribute in any dataset "
"collection." % attr)
return col
def _collection_id2obj(self, col):
if col == 'sa':
col = self.sa
elif col == 'fa':
col = self.fa
elif col == 'a':
col = self.a
else:
raise LookupError("Unknown collection '%s'. Possible values "
"are: 'sa', 'fa', 'a'." % col)
return col
def set_attr(self, name, value):
"""Set an attribute in a collection.
Parameters
----------
name : str
Collection and attribute name. This has to be in the same format as
for ``get_attr()``.
value : array
Value of the attribute.
"""
if '.' in name:
col, name = name.split('.')[0:2]
# translate collection names into collection
col = self._collection_id2obj(col)
else:
# auto-detect collection
col = self.find_collection(name)
col[name] = value
def get_attr(self, name):
"""Return an attribute from a collection.
A collection can be specified, but can also be auto-detected.
Parameters
----------
name : str
Attribute name. The attribute name can also be prefixed with any valid
collection name ('sa', 'fa', or 'a') separated with a '.', e.g.
'sa.targets'. If no collection prefix is found auto-detection of the
collection is attempted.
Returns
-------
(attr, collection)
2-tuple: First element is the requested attribute and the second
element is the collection that contains the attribute. If no matching
attribute can be found a LookupError exception is raised.
"""
if '.' in name:
col, name = name.split('.')[0:2]
# translate collection names into collection
col = self._collection_id2obj(col)
else:
# auto-detect collection
col = self.find_collection(name)
return (col[name], col)
def item(self):
"""Provide the first element of samples array.
Notes
-----
Introduced to provide compatibility with `numpy.asscalar`.
See `numpy.ndarray.item` for more information.
"""
return self.samples.item()
@property
def idhash(self):
"""To verify if dataset is in the same state as when smth else was done
Like if classifier was trained on the same dataset as in question
"""
res = 'self@%s samples@%s' % (idhash_(self), idhash_(self.samples))
for col in (self.a, self.sa, self.fa):
# We cannot count on the order the values in the dict will show up
# with `self._data.value()` and since idhash will be order-dependent
# we have to make it deterministic
keys = col.keys()
keys.sort()
for k in keys:
res += ' %s@%s' % (k, idhash_(col[k].value))
return res
@classmethod
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
@classmethod
def from_channeltimeseries(cls, samples, targets=None, chunks=None,
t0=None, dt=None, channelids=None):
"""Create a dataset from segmented, per-channel timeseries.
Channels are assumes to contain multiple, equally spaced acquisition
timepoints. The dataset will contain additional feature attributes
associating each feature with a specific `channel` and `timepoint`.
Parameters
----------
samples : ndarray
Three-dimensional array: (samples x channels x timepoints).
t0 : float
Reference time of the first timepoint. Can be used to preserve
information about the onset of some stimulation. Preferably in
seconds.
dt : float
Temporal distance between two timepoints. Preferably in seconds.
channelids : list
List of channel names.
targets, chunks
See `Dataset.from_wizard` for documentation about these arguments.
"""
# check samples
if len(samples.shape) != 3:
raise ValueError(
"Input data should be (samples x channels x timepoints. Got: %s"
% samples.shape)
if not t0 is None and not dt is None:
timepoints = np.arange(t0, t0 + samples.shape[2] * dt, dt)
# broadcast over all channels
timepoints = np.vstack([timepoints] * samples.shape[1])
else:
timepoints = None
if not channelids is None:
if len(channelids) != samples.shape[1]:
raise ValueError(
"Number of channel ids does not match channels in the "
"sample data. Expected %i, but got %i"
% (samples.shape[1], len(channelids)))
# broadcast over all timepoints
channelids = np.dstack([channelids] * samples.shape[2])[0]
ds = cls.from_wizard(samples, targets=targets, chunks=chunks)
# add additional attributes
if not timepoints is None:
ds.fa['timepoints'] = ds.a.mapper.forward1(timepoints)
if not channelids is None:
ds.fa['channels'] = ds.a.mapper.forward1(channelids)
return ds
# shortcut properties
S = property(fget=lambda self:self.samples)
targets = property(fget=lambda self:self.sa.targets,
fset=lambda self, v:self.sa.__setattr__('targets', v))
uniquetargets = property(fget=lambda self:self.sa['targets'].unique)
T = targets
UT = property(fget=lambda self:self.sa['targets'].unique)
chunks = property(fget=lambda self:self.sa.chunks,
fset=lambda self, v:self.sa.__setattr__('chunks', v))
uniquechunks = property(fget=lambda self:self.sa['chunks'].unique)
C = chunks
UC = property(fget=lambda self:self.sa['chunks'].unique)
mapper = property(fget=lambda self:self.a.mapper)
O = property(fget=lambda self:self.a.mapper.reverse(self.samples))
# convenience alias
dataset_wizard = Dataset.from_wizard
class HollowSamples(object):
"""Samples container that doesn't store samples.
The purpose of this class is to provide an object that can be used as
``samples`` in a Dataset, without having actual samples. Instead of storing
multiple samples it only maintains a IDs for samples and features it
pretends to contain.
Using this class in a dataset in conjuction will actual attributes, will
yield a lightweight dataset that is compatible with the majority of all
mappers and can be used to 'simulate' processing by mappers. The class
offers acces to the sample and feature IDs via its ``sid`` and ``fid``
members.
"""
def __init__(self, shape=None, sid=None, fid=None, dtype=np.float):
"""
Parameters
----------
shape : 2-tuple or None
Shape of the pretend-sample array (nsamples x nfeatures). Can be
left out if both ``sid`` and ``fid`` are provided.
sid : 1d-array or None
Vector of sample IDs. Can be left out if ``shape`` is provided.
fid : 1d-array or None
Vector of feature IDs. Can be left out if ``shape`` is provided.
dtype : type or str
Pretend-datatype of the non-existing samples.
"""
if shape is None and sid is None and fid is None:
raise ValueError("Either shape or ID vectors have to be given")
if not shape is None and not len(shape) == 2:
raise ValueError("Only two-dimensional shapes are supported")
if sid is None:
self.sid = np.arange(shape[0], dtype='uint')
else:
self.sid = sid
if fid is None:
self.fid = np.arange(shape[1], dtype='uint')
else:
self.fid = fid
self.dtype = dtype
# sanity check
if not shape is None and not len(self.sid) == shape[0] \
and not len(self.fid) == shape[1]:
raise ValueError("Provided ID vectors do not match given `shape`")
def __reduce__(self):
return (self.__class__,
((len(self.sid), len(self.fid)),
self.sid,
self.fid,
self.dtype))
def copy(self, deep=True):
return deep and copy.deepcopy(self) or copy.copy(self)
@property
def shape(self):
return (len(self.sid), len(self.fid))
@property
def samples(self):
return np.zeros((len(self.sid), len(self.fid)), dtype=self.dtype)
def __array__(self, dtype=None):
# come up with a fake array of proper dtype
return np.zeros((len(self.sid), len(self.fid)), dtype=self.dtype)
def __getitem__(self, args):
if not isinstance(args, tuple):
args = (args,)
if len(args) > 2:
raise ValueError("Too many arguments (%i). At most there can be "
"two arguments, one for samples selection and one "
"for features selection" % len(args))
if len(args) == 1:
args = [args[0], slice(None)]
else:
args = [a for a in args]
# ints need to become lists to prevent silent dimensionality changes
# of the arrays when slicing
for i, a in enumerate(args):
if isinstance(a, int):
args[i] = [a]
# apply to vectors
sid = self.sid[args[0]]
fid = self.fid[args[1]]
return HollowSamples((len(sid), len(fid)), sid=sid, fid=fid,
dtype=self.dtype)
def view(self):
"""Return itself"""
return self
|