code.rst 18.6 KB
Newer Older
1

Martin Reinecke's avatar
Martin Reinecke committed
2
=============
3 4
Code Overview
=============
5

Martin Reinecke's avatar
Martin Reinecke committed
6 7 8 9

Executive summary
=================

10 11
The fundamental building blocks required for IFT computations are best
recognized from a large distance, ignoring all technical details.
12

13 14
From such a perspective,

Martin Reinecke's avatar
tweaks  
Martin Reinecke committed
15
- IFT problems largely consist of the combination of several high dimensional
16
  *minimization* problems.
Philipp Arras's avatar
Philipp Arras committed
17
- Within NIFTy, *operators* are used to define the characteristic equations and
Martin Reinecke's avatar
tweaks  
Martin Reinecke committed
18
  properties of the problems.
Martin Reinecke's avatar
Martin Reinecke committed
19 20
- The equations are built mostly from the application of *linear operators*,
  but there may also be nonlinear functions involved.
21
- The unknowns in the equations represent either continuous physical *fields*,
Martin Reinecke's avatar
Martin Reinecke committed
22 23 24 25
  or they are simply individual measured *data points*.
- Discretized *fields* have geometrical information (like locations and volume
  elements) associated with every entry; this information is called the field's
  *domain*.
26 27 28 29 30 31 32 33 34 35 36 37 38

In the following sections, the concepts briefly presented here will be
discussed in more detail; this is done in reversed order of their introduction,
to avoid forward references.


Domains
=======


Abstract base class
-------------------

39 40
.. currentmodule:: nifty5.domains.domain

Philipp Arras's avatar
Philipp Arras committed
41
One of the fundamental building blocks of the NIFTy5 framework is the *domain*.
42
Its required capabilities are expressed by the abstract :py:class:`Domain` class.
43
A domain must be able to answer the following queries:
44
m
45 46 47 48 49 50 51 52 53 54 55

- its total number of data entries (pixels), which is accessible via the
  :attr:`~Domain.size` property
- the shape of the array that is supposed to hold these data entries
  (obtainable by means of the :attr:`~Domain.shape` property)
- equality comparison to another :class:`Domain` instance


Unstructured domains
--------------------

56 57
.. currentmodule:: nifty5.domains.unstructured_domain

58 59 60 61 62 63 64 65 66 67 68
Domains can be either *structured* (i.e. there is geometrical information
associated with them, like position in space and volume factors),
or *unstructured* (meaning that the data points have no associated manifold).

Unstructured domains can be described by instances of NIFTy's
:class:`UnstructuredDomain` class.


Structured domains
------------------

69 70
.. currentmodule:: nifty5.domains.structured_domain

71 72 73 74 75
In contrast to unstructured domains, these domains have an assigned geometry.
NIFTy requires them to provide the volume elements of their grid cells.
The additional methods are specified in the abstract class
:class:`StructuredDomain`:

Martin Reinecke's avatar
Martin Reinecke committed
76
- The properties :attr:`~StructuredDomain.scalar_dvol`,
77 78 79 80 81 82 83 84 85 86 87 88 89
  :attr:`~StructuredDomain.dvol`, and  :attr:`~StructuredDomain.total_volume`
  provide information about the domain's pixel volume(s) and its total volume.
- The property :attr:`~StructuredDomain.harmonic` specifies whether a domain
  is harmonic (i.e. describes a frequency space) or not
- Iff the domain is harmonic, the methods
  :meth:`~StructuredDomain.get_k_length_array`,
  :meth:`~StructuredDomain.get_unique_k_lengths`, and
  :meth:`~StructuredDomain.get_fft_smoothing_kernel_function` provide absolute
  distances of the individual grid cells from the origin and assist with
  Gaussian convolution.

NIFTy comes with several concrete subclasses of :class:`StructuredDomain`:

90 91 92
.. currentmodule:: nifty5.domains

- :class:`rg_space.RGSpace` represents a regular Cartesian grid with an arbitrary
93
  number of dimensions, which is supposed to be periodic in each dimension.
94 95
- :class:`hp_space.HPSpace` and :class:`gl_space.GLSpace` describe pixelisations of the
  2-sphere; their counterpart in harmonic space is :class:`lm_space.LMSpace`, which
96
  contains spherical harmonic coefficients.
97
- :class:`power_space.PowerSpace` is used to describe one-dimensional power spectra.
98

99 100
Among these, :class:`rg_space.RGSpace` can be harmonic or not (depending on constructor arguments), :class:`gl_space.GLSpace`, :class:`hp_space.HPSpace`, and :class:`power_space.PowerSpace` are
pure position domains (i.e. nonharmonic), and :class:`lm_space.LMSpace` is always
101 102 103 104 105 106 107
harmonic.


Combinations of domains
=======================

The fundamental classes described above are often sufficient to specify the
108 109
domain of a field. In some cases, however, it will be necessary to define the
field on a product of elementary domains instead of a single one.
Philipp Arras's avatar
Philipp Arras committed
110
More sophisticated operators also require a set of several such fields.
111 112 113 114 115
Some examples are:

- sky emission depending on location and energy. This could be represented by
  a product of an :class:`HPSpace` (for location) with an :class:`RGSpace`
  (for energy).
Philipp Arras's avatar
Philipp Arras committed
116
- a polarized field, which could be modeled as a product of any structured
Martin Reinecke's avatar
Martin Reinecke committed
117 118
  domain (representing location) with a four-element
  :class:`UnstructuredDomain` holding Stokes I, Q, U and V components.
Martin Reinecke's avatar
Martin Reinecke committed
119 120 121
- a model for the sky emission, which holds both the current realization
  (on a harmonic domain) and a few inferred model parameters (e.g. on an
  unstructured grid).
122 123 124 125 126 127

Consequently, NIFTy defines a class called :class:`DomainTuple` holding
a sequence of :class:`Domain` objects, which is used to specify full field
domains. In principle, a :class:`DomainTuple` can even be empty, which implies
that the field living on it is a scalar.

Martin Reinecke's avatar
Martin Reinecke committed
128 129 130 131
A :class:`DomainTuple` supports iteration and indexing, and also provides the
properties :attr:`~DomainTuple.shape`, :attr:`~DomainTuple.size` in analogy to
the elementary :class:`Domain`.

Torsten Ensslin's avatar
Torsten Ensslin committed
132
An aggregation of several :class:`DomainTuple` s, each member identified by a
Martin Reinecke's avatar
Martin Reinecke committed
133
name, is described by the :class:`MultiDomain` class.
134 135 136 137

Fields
======

Martin Reinecke's avatar
Martin Reinecke committed
138 139 140
Fields on a single DomainTuple
------------------------------

141 142 143 144 145 146
A :class:`Field` object consists of the following components:

- a domain in form of a :class:`DomainTuple` object
- a data type (e.g. numpy.float64)
- an array containing the actual values

Martin Reinecke's avatar
Martin Reinecke committed
147
Usually, the array is stored in the form of a ``numpy.ndarray``, but for very
Martin Reinecke's avatar
Martin Reinecke committed
148 149 150 151 152 153 154 155 156 157 158 159 160 161
resource-intensive tasks NIFTy also provides an alternative storage method to
be used with distributed memory processing.

Fields support a wide range of arithmetic operations, either involving two
fields with equal domains, or a field and a scalar.
Contractions (like summation, integration, minimum/maximum, computation of
statistical moments) can be carried out either over an entire field (producing
a scalar result) or over sub-domains (resulting in a field living on a smaller
domain). Scalar products of two fields can also be computed easily.

There is also a set of convenience functions to generate fields with constant
values or fields filled with random numbers according to a user-specified
distribution.

Martin Reinecke's avatar
Martin Reinecke committed
162 163 164 165 166
Like almost all NIFTy objects, fields are immutable: their value or any other
attribute cannot be modified after construction. To manipulate a field in ways
that are not covered by the provided standard operations, its data content must
be extracted first, then changed, and a new field has to be created from the
result.
Martin Reinecke's avatar
Martin Reinecke committed
167

168
Fields defined on a MultiDomain
Martin Reinecke's avatar
Martin Reinecke committed
169
-------------------------------
Martin Reinecke's avatar
Martin Reinecke committed
170 171

The :class:`MultiField` class can be seen as a dictionary of individual
172
:class:`Field` s, each identified by a name, which is defined on a
Martin Reinecke's avatar
Martin Reinecke committed
173 174 175 176 177 178 179 180
:class:`MultiDomain`.


Operators
=========

All transformations between different NIFTy fields are expressed  (explicitly
or implicitly) in the form of :class:`Operator` objects. The interface of this
Torsten Ensslin's avatar
Torsten Ensslin committed
181
class is very minimalistic: it has a property called :class:`domain` which returns
182
a :class:`DomainTuple` or :class:`MultiDomain` object specifying the structure of the
Torsten Ensslin's avatar
Torsten Ensslin committed
183
:class:`Field` s or :class:`MultiField` s it expects as input, another property :class:`target`
Martin Reinecke's avatar
Martin Reinecke committed
184 185 186
describing its output, and finally an overloaded `apply` method, which can
take

Martin Reinecke's avatar
Martin Reinecke committed
187
- a :class:`Field`/:class:`MultiField` object, in which case it returns the transformed
Torsten Ensslin's avatar
Torsten Ensslin committed
188 189 190
  :class:`Field`/:class:`MultiField`
- a :class:`Linearization` object, in which case it returns the transformed
  :class:`Linearization`
Martin Reinecke's avatar
Martin Reinecke committed
191

Torsten Ensslin's avatar
Torsten Ensslin committed
192 193
This is the interface that all objects derived from :class:`Operator` must implement.
In addition, :class:`Operator` objects can be added/subtracted, multiplied, chained
Martin Reinecke's avatar
Martin Reinecke committed
194 195
(via the :class:`__call__` method and the `@` operator) and support pointwise
application of functions like
Torsten Ensslin's avatar
Torsten Ensslin committed
196
:class:`exp()`, :class:`log()`, :class:`sqrt()`, :class:`conjugate()` etc.
Martin Reinecke's avatar
Martin Reinecke committed
197

198

Martin Reinecke's avatar
Martin Reinecke committed
199 200 201 202 203 204 205 206 207 208 209
Advanced operators
------------------

NIFTy provides a library of more sophisticated operators which are used for more
specific inference problems. Currently these are:

- :class:`AmplitudeOperator`, which returns a smooth power spectrum.
- :class:`InverseGammaOperator`, which models point sources which follow a inverse gamma distribution.
- :class:`CorrelatedField`, which models a diffuse log-normal field. It takes an amplitude operator to specify the correlation structure of the field.


210 211 212
Linear Operators
================

Philipp Arras's avatar
Philipp Arras committed
213
A linear operator (represented by NIFTy5's abstract :class:`LinearOperator`
Martin Reinecke's avatar
Martin Reinecke committed
214 215
class) is derived from `Operator` and can be interpreted as an
(implicitly defined) matrix. Since its operation is linear, it can provide some
Torsten Ensslin's avatar
Torsten Ensslin committed
216
additional functionality which is not available for the more generic :class:`Operator`
Martin Reinecke's avatar
Martin Reinecke committed
217
class.
218

Martin Reinecke's avatar
Martin Reinecke committed
219

Martin Reinecke's avatar
Martin Reinecke committed
220 221
Linear Operator basics
----------------------
Martin Reinecke's avatar
Martin Reinecke committed
222

223 224
There are four basic ways of applying an operator :math:`A` to a field :math:`f`:

Martin Reinecke's avatar
Martin Reinecke committed
225 226 227 228
- direct application: :math:`A\cdot f`
- adjoint application: :math:`A^\dagger \cdot f`
- inverse application: :math:`A^{-1}\cdot f`
- adjoint inverse application: :math:`(A^\dagger)^{-1}\cdot f`
229

Martin Reinecke's avatar
Martin Reinecke committed
230 231
(Because of the linearity, inverse adjoint and adjoint inverse application
are equivalent.)
232

Martin Reinecke's avatar
Martin Reinecke committed
233 234
These different actions of a linear operator ``Op`` on a field ``f`` can be
invoked in various ways:
235 236 237 238 239 240 241 242 243 244 245 246 247

- direct multiplication: ``Op(f)`` or ``Op.times(f)`` or ``Op.apply(f, Op.TIMES)``
- adjoint multiplication: ``Op.adjoint_times(f)`` or ``Op.apply(f, Op.ADJOINT_TIMES)``
- inverse multiplication: ``Op.inverse_times(f)`` or ``Op.apply(f, Op.INVERSE_TIMES)``
- adjoint inverse multiplication: ``Op.adjoint_inverse_times(f)`` or ``Op.apply(f, Op.ADJOINT_INVERSE_TIMES)``

Operator classes defined in NIFTy may implement an arbitrary subset of these
four operations. This subset can be queried using the
:attr:`~LinearOperator.capability` property.

If needed, the set of supported operations can be enhanced by iterative
inversion methods;
for example, an operator defining direct and adjoint multiplication could be
Martin Reinecke's avatar
Martin Reinecke committed
248
enhanced by this approach to support the complete set. This functionality is
249 250 251 252 253 254 255 256 257 258 259 260 261 262
provided by NIFTy's :class:`InversionEnabler` class, which is itself a linear
operator.

Direct multiplication and adjoint inverse multiplication transform a field
living on the operator's :attr:`~LinearOperator.domain` to one living on the operator's :attr:`~LinearOperator.target`, whereas adjoint multiplication
and inverse multiplication transform from :attr:`~LinearOperator.target` to :attr:`~LinearOperator.domain`.

Operators with identical domain and target can be derived from
:class:`EndomorphicOperator`; typical examples for this category are the :class:`ScalingOperator`, which simply multiplies its input by a scalar
value, and :class:`DiagonalOperator`, which multiplies every value of its input
field with potentially different values.

Further operator classes provided by NIFTy are

Martin Reinecke's avatar
Martin Reinecke committed
263 264
- :class:`HarmonicTransformOperator` for transforms from a harmonic domain to
  its counterpart in position space, and their adjoint
265
- :class:`PowerDistributor` for transforms from a :class:`PowerSpace` to
Martin Reinecke's avatar
Martin Reinecke committed
266
  an associated harmonic domain, and their adjoint
267 268 269 270
- :class:`GeometryRemover`, which transforms from structured domains to
  unstructured ones. This is typically needed when building instrument response
  operators.

Martin Reinecke's avatar
Martin Reinecke committed
271

Martin Reinecke's avatar
Martin Reinecke committed
272 273
Syntactic sugar
---------------
Martin Reinecke's avatar
Martin Reinecke committed
274

Martin Reinecke's avatar
Martin Reinecke committed
275
Nifty5 allows simple and intuitive construction of altered and combined
276 277
operators.
As an example, if ``A``, ``B`` and ``C`` are of type :class:`LinearOperator`
Martin Reinecke's avatar
Martin Reinecke committed
278
and ``f1`` and ``f2`` are of type :class:`Field`, writing::
279

Martin Reinecke's avatar
Martin Reinecke committed
280
    X = A(B.inverse(A.adjoint)) + C
281 282 283 284 285 286 287 288 289 290 291
    f2 = X(f1)

will perform the operation suggested intuitively by the notation, checking
domain compatibility while building the composed operator.
The combined operator infers its domain and target from its constituents,
as well as the set of operations it can support.
The properties :attr:`~LinearOperator.adjoint` and
:attr:`~LinearOperator.inverse` return a new operator which behaves as if it
were the original operator's adjoint or inverse, respectively.


Philipp Arras's avatar
Philipp Arras committed
292 293
Operators
=========
294

Philipp Arras's avatar
Philipp Arras committed
295
Operator classes (represented by NIFTy5's abstract :class:`Operator` class) are used to construct
Martin Reinecke's avatar
tweaks  
Martin Reinecke committed
296
the equations of a specific inference problem.
Philipp Arras's avatar
Philipp Arras committed
297
Most operators are defined via a position, which is a :class:`MultiField` object,
Martin Reinecke's avatar
tweaks  
Martin Reinecke committed
298
their value at this position, which is again a :class:`MultiField` object and a Jacobian derivative,
299
which is a :class:`LinearOperator` and is needed for the minimization procedure.
Martin Reinecke's avatar
tweaks  
Martin Reinecke committed
300

Philipp Arras's avatar
Philipp Arras committed
301
Using the existing basic operator classes one can construct more complicated operators, as
Martin Reinecke's avatar
tweaks  
Martin Reinecke committed
302
NIFTy allows for easy and self-consinstent combination via point-wise multiplication,
Philipp Arras's avatar
Philipp Arras committed
303
addition and subtraction. The operator resulting from these operations then automatically
Martin Reinecke's avatar
tweaks  
Martin Reinecke committed
304 305 306 307
contains the correct Jacobians, positions and values.
Notably, :class:`Constant` and :class:`Variable` allow for an easy way to turn
inference of specific quantities on and off.

Philipp Arras's avatar
Philipp Arras committed
308
The basic operator classes also allow for more complex operations on operators such as
309
the application of :class:`LinearOperators` or local non-linearities.
Philipp Arras's avatar
Philipp Arras committed
310 311
As an example one may consider the following combination of ``x``, which is an operator of type
:class:`Variable` and ``y``, which is an operator of type :class:`Constant`::
312

Martin Reinecke's avatar
tweaks  
Martin Reinecke committed
313
	z = x*x + y
314

Philipp Arras's avatar
Philipp Arras committed
315
``z`` will then be an operator with the following properties::
316 317

	z.value = x.value*x.value + y.value
Martin Reinecke's avatar
Martin Reinecke committed
318
	z.position = Union(x.position, y.position)
319 320 321
	z.jacobian = 2*makeOp(x.value)


Philipp Arras's avatar
Philipp Arras committed
322
Basic operators
Philipp Arras's avatar
Philipp Arras committed
323
---------------
Philipp Arras's avatar
Philipp Arras committed
324
# FIXME All this is outdated!
325

Philipp Arras's avatar
Philipp Arras committed
326
Basic operator classes provided by NIFTy are
327

Martin Reinecke's avatar
tweaks  
Martin Reinecke committed
328
- :class:`Constant` contains a constant value and has a zero valued Jacobian.
Philipp Arras's avatar
Philipp Arras committed
329
  Like other operators, it has a position, but its value does not depend on it.
330 331 332
- :class:`Variable` returns the position as its value, its derivative is one.
- :class:`LinearModel` applies a :class:`LinearOperator` on the model.
- :class:`LocalModel` applies a non-linearity locally on the model.
Martin Reinecke's avatar
Martin Reinecke committed
333
	value and Jacobian are combined into corresponding :class:`MultiFields` and operators.
Martin Reinecke's avatar
tweaks  
Martin Reinecke committed
334

335

Philipp Arras's avatar
Philipp Arras committed
336 337
Advanced operators
------------------
338

Philipp Arras's avatar
Philipp Arras committed
339
NIFTy also provides a library of more sophisticated operators which are used for more
340 341
specific inference problems. Currently these are:

Philipp Arras's avatar
Philipp Arras committed
342 343 344
- :class:`AmplitudeOperator`, which returns a smooth power spectrum.
- :class:`InverseGammaOperator`, which models point sources which follow a inverse gamma distribution.
- :class:`CorrelatedField`, which models a diffuse log-normal field. It takes an amplitude operator
345 346 347
	to specify the correlation structure of the field.


348 349
.. _minimization:

Martin Reinecke's avatar
Martin Reinecke committed
350

351 352 353
Minimization
============

Martin Reinecke's avatar
Martin Reinecke committed
354 355 356 357 358 359
Most problems in IFT are solved by (possibly nested) minimizations of
high-dimensional functions, which are often nonlinear.


Energy functionals
------------------
360

Philipp Arras's avatar
Philipp Arras committed
361
In NIFTy5 such functions are represented by objects of type :class:`Energy`.
Martin Reinecke's avatar
Martin Reinecke committed
362 363
These hold the prescription how to calculate the function's
:attr:`~Energy.value`, :attr:`~Energy.gradient` and
Martin Reinecke's avatar
Martin Reinecke committed
364
(optionally) :attr:`~Energy.metric` at any given :attr:`~Energy.position`
Martin Reinecke's avatar
Martin Reinecke committed
365
in parameter space.
Martin Reinecke's avatar
Martin Reinecke committed
366
Function values are floating-point scalars, gradients have the form of fields
Martin Reinecke's avatar
Martin Reinecke committed
367
living on the energy's position domain, and metrics are represented by
Martin Reinecke's avatar
Martin Reinecke committed
368
linear operator objects.
369

Martin Reinecke's avatar
Martin Reinecke committed
370 371
Energies are classes that typically have to be provided by the user when
tackling new IFT problems.
Philipp Arras's avatar
Philipp Arras committed
372
Some examples of concrete energy classes delivered with NIFTy5 are
Martin Reinecke's avatar
Martin Reinecke committed
373
:class:`QuadraticEnergy` (with position-independent metric, mainly used with
Philipp Arras's avatar
Philipp Arras committed
374
conjugate gradient minimization) and :class:`~nifty5.library.WienerFilterEnergy`.
375

Martin Reinecke's avatar
Martin Reinecke committed
376 377 378

Iteration control
-----------------
Martin Reinecke's avatar
Martin Reinecke committed
379 380

Iterative minimization of an energy reqires some means of
Martin Reinecke's avatar
Martin Reinecke committed
381
checking the quality of the current solution estimate and stopping once
Martin Reinecke's avatar
Martin Reinecke committed
382 383 384
it is sufficiently accurate. In case of numerical problems, the iteration needs
to be terminated as well, returning a suitable error description.

Philipp Arras's avatar
Philipp Arras committed
385
In NIFTy5, this functionality is encapsulated in the abstract
Martin Reinecke's avatar
Martin Reinecke committed
386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401
:class:`IterationController` class, which is provided with the initial energy
object before starting the minimization, and is updated with the improved
energy after every iteration. Based on this information, it can either continue
the minimization or return the current estimate indicating convergence or
failure.

Sensible stopping criteria can vary significantly with the problem being
solved; NIFTy provides one concrete sub-class of :class:`IterationController`
called :class:`GradientNormController`, which should be appropriate in many
circumstances, but users have complete freedom to implement custom sub-classes
for their specific applications.


Minimization algorithms
-----------------------

Martin Reinecke's avatar
Martin Reinecke committed
402 403 404 405 406 407 408 409 410 411 412 413 414
All minimization algorithms in NIFTy inherit from the abstract
:class:`Minimizer` class, which presents a minimalistic interface consisting
only of a :meth:`~Minimizer.__call__` method taking an :class:`Energy` object
and optionally a preconditioning operator, and returning the energy at the
discovered minimum and a status code.

For energies with a quadratic form (i.e. which
can be expressed by means of a :class:`QuadraticEnergy` object), an obvious
choice of algorithm is the :class:`ConjugateGradient` minimizer.

A similar algorithm suited for nonlinear problems is provided by
:class:`NonlinearCG`.

Martin Reinecke's avatar
Martin Reinecke committed
415
Many minimizers for nonlinear problems can be characterized as
Martin Reinecke's avatar
Martin Reinecke committed
416 417 418 419 420 421 422 423 424

- first deciding on a direction for the next step
- then finding a suitable step length along this direction, resulting in the
  next energy estimate.

This family of algorithms is encapsulated in NIFTy's :class:`DescentMinimizer`
class, which currently has three concrete implementations:
:class:`SteepestDescent`, :class:`VL_BFGS`, and :class:`RelaxedNewton`.
Of these algorithms, only :class:`RelaxedNewton` requires the energy object to
Martin Reinecke's avatar
Martin Reinecke committed
425
provide a :attr:`~Energy.metric` property, the others only need energy
Martin Reinecke's avatar
Martin Reinecke committed
426 427
values and gradients.

Martin Reinecke's avatar
Martin Reinecke committed
428 429 430 431 432
The flexibility of NIFTy's design allows using externally provided
minimizers. With only small effort, adapters for two SciPy minimizers were
written; they are available under the names :class:`NewtonCG` and
:class:`L_BFGS_B`.

Martin Reinecke's avatar
Martin Reinecke committed
433 434 435 436

Application to operator inversion
---------------------------------

Martin Reinecke's avatar
Martin Reinecke committed
437 438 439 440 441 442 443 444 445 446
It is important to realize that the machinery presented here cannot only be
used for minimizing IFT Hamiltonians, but also for the numerical inversion of
linear operators, if the desired application mode is not directly available.
A classical example is the information propagator

:math:`D = \left(R^\dagger N^{-1} R + S^{-1}\right)^{-1}`,

which must be applied when calculating a Wiener filter. Only its inverse
application is straightforward; to use it in forward direction, we make use
of NIFTy's :class:`InversionEnabler` class, which internally performs a
Martin Reinecke's avatar
Martin Reinecke committed
447
minimization of a :class:`QuadraticEnergy` by means of the
Martin Reinecke's avatar
Martin Reinecke committed
448
:class:`ConjugateGradient` algorithm.