diagonal_operator.py 7.29 KB
Newer Older
1
2
3
4
5
6
7
8
9
10
11
12
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program.  If not, see <http://www.gnu.org/licenses/>.
Theo Steininger's avatar
Theo Steininger committed
13
#
14
# Copyright(C) 2013-2018 Max-Planck-Society
Theo Steininger's avatar
Theo Steininger committed
15
16
17
#
# NIFTy is being developed at the Max-Planck-Institut fuer Astrophysik
# and financially supported by the Studienstiftung des deutschen Volkes.
18

Martin Reinecke's avatar
Martin Reinecke committed
19
from __future__ import division
20
import numpy as np
Martin Reinecke's avatar
cleanup    
Martin Reinecke committed
21
from ..field import Field
Martin Reinecke's avatar
Martin Reinecke committed
22
23
from ..domain_tuple import DomainTuple
from .endomorphic_operator import EndomorphicOperator
24
from .. import utilities
Martin Reinecke's avatar
Martin Reinecke committed
25
from .. import dobj
Martin Reinecke's avatar
Martin Reinecke committed
26

27
28

class DiagonalOperator(EndomorphicOperator):
Martin Reinecke's avatar
Martin Reinecke committed
29
    """ NIFTy class for diagonal operators.
Theo Steininger's avatar
Theo Steininger committed
30

Martin Reinecke's avatar
Martin Reinecke committed
31
    The NIFTy DiagonalOperator class is a subclass derived from the
Theo Steininger's avatar
Theo Steininger committed
32
33
    EndomorphicOperator. It multiplies an input field pixel-wise with its
    diagonal.
34

35
36
    Parameters
    ----------
Martin Reinecke's avatar
Martin Reinecke committed
37
    diagonal : Field
Martin Reinecke's avatar
docs    
Martin Reinecke committed
38
39
        The diagonal entries of the operator.
    domain : Domain, tuple of Domain or DomainTuple, optional
40
41
        The domain on which the Operator's input Field lives.
        If None, use the domain of "diagonal".
Martin Reinecke's avatar
docs    
Martin Reinecke committed
42
    spaces : int or tuple of int, optional
43
44
        The elements of "domain" on which the operator acts.
        If None, it acts on all elements.
Martin Reinecke's avatar
Martin Reinecke committed
45
46
47
48
49
50
51
52
53
54

    Notes
    -----
    Formally, this operator always supports all operation modes (times,
    adjoint_times, inverse_times and inverse_adjoint_times), even if there
    are diagonal elements with value 0 or infinity. It is the user's
    responsibility to apply the operator only in appropriate ways (e.g. call
    inverse_times only if there are no zeros on the diagonal).

    This shortcoming will hopefully be fixed in the future.
55
56
    """

Martin Reinecke's avatar
Martin Reinecke committed
57
    def __init__(self, diagonal, domain=None, spaces=None):
58
        super(DiagonalOperator, self).__init__()
59

Martin Reinecke's avatar
Martin Reinecke committed
60
61
        if not isinstance(diagonal, Field):
            raise TypeError("Field object required")
62
63
64
65
66
67
68
69
70
        if domain is None:
            self._domain = diagonal.domain
        else:
            self._domain = DomainTuple.make(domain)
        if spaces is None:
            self._spaces = None
            if diagonal.domain != self._domain:
                raise ValueError("domain mismatch")
        else:
71
72
            self._spaces = utilities.parse_spaces(spaces, len(self._domain))
            if len(self._spaces) != len(diagonal.domain):
73
                raise ValueError("spaces and domain must have the same length")
Martin Reinecke's avatar
Martin Reinecke committed
74
            for i, j in enumerate(self._spaces):
75
76
                if diagonal.domain[i] != self._domain[j]:
                    raise ValueError("domain mismatch")
Martin Reinecke's avatar
Martin Reinecke committed
77
            if self._spaces == tuple(range(len(self._domain))):
Martin Reinecke's avatar
tweak    
Martin Reinecke committed
78
79
80
81
82
83
84
                self._spaces = None  # shortcut

        if self._spaces is not None:
            active_axes = []
            for space_index in self._spaces:
                active_axes += self._domain.axes[space_index]

Martin Reinecke's avatar
Martin Reinecke committed
85
            if self._spaces[0] == 0:
Martin Reinecke's avatar
Martin Reinecke committed
86
                self._ldiag = diagonal.local_data
Martin Reinecke's avatar
Martin Reinecke committed
87
            else:
Martin Reinecke's avatar
Martin Reinecke committed
88
                self._ldiag = diagonal.to_global_data()
Martin Reinecke's avatar
Martin Reinecke committed
89
            locshape = dobj.local_shape(self._domain.shape, 0)
Martin Reinecke's avatar
tweak    
Martin Reinecke committed
90
            self._reshaper = [shp if i in active_axes else 1
Martin Reinecke's avatar
Martin Reinecke committed
91
92
93
                              for i, shp in enumerate(locshape)]
            self._ldiag = self._ldiag.reshape(self._reshaper)
        else:
Martin Reinecke's avatar
Martin Reinecke committed
94
            self._ldiag = diagonal.local_data
95
96
97
        self._update_diagmin()

    def _update_diagmin(self):
Martin Reinecke's avatar
Martin Reinecke committed
98
        self._ldiag.flags.writeable = False
99
100
101
        if not np.issubdtype(self._ldiag.dtype, np.complexfloating):
            lmin = self._ldiag.min() if self._ldiag.size > 0 else 1.
            self._diagmin = dobj.np_allreduce_min(np.array(lmin))[()]
Martin Reinecke's avatar
Martin Reinecke committed
102

103
    def _from_ldiag(self, spc, ldiag):
Martin Reinecke's avatar
Martin Reinecke committed
104
105
106
107
108
109
        res = DiagonalOperator.__new__(DiagonalOperator)
        res._domain = self._domain
        if self._spaces is None or spc is None:
            res._spaces = None
        else:
            res._spaces = tuple(set(self._spaces) | set(spc))
110
111
        res._ldiag = ldiag
        res._update_diagmin()
Martin Reinecke's avatar
Martin Reinecke committed
112
113
114
115
116
        return res

    def _scale(self, fct):
        if not np.isscalar(fct):
            raise TypeError("scalar value required")
117
        return self._from_ldiag((), self._ldiag*fct)
Martin Reinecke's avatar
Martin Reinecke committed
118
119
120
121

    def _add(self, sum):
        if not np.isscalar(sum):
            raise TypeError("scalar value required")
122
        return self._from_ldiag((), self._ldiag+sum)
Martin Reinecke's avatar
Martin Reinecke committed
123
124
125
126

    def _combine_prod(self, op):
        if not isinstance(op, DiagonalOperator):
            raise TypeError("DiagonalOperator required")
127
        return self._from_ldiag(op._spaces, self._ldiag*op._ldiag)
Martin Reinecke's avatar
Martin Reinecke committed
128
129
130
131

    def _combine_sum(self, op, selfneg, opneg):
        if not isinstance(op, DiagonalOperator):
            raise TypeError("DiagonalOperator required")
132
133
134
        tdiag = (self._ldiag * (-1 if selfneg else 1) +
                 op._ldiag * (-1 if opneg else 1))
        return self._from_ldiag(op._spaces, tdiag)
135

Martin Reinecke's avatar
Martin Reinecke committed
136
137
    def apply(self, x, mode):
        self._check_input(x, mode)
138

Martin Reinecke's avatar
Martin Reinecke committed
139
140
141
142
143
144
145
146
147
148
149
150
151
152
        if mode == self.TIMES:
            return Field(x.domain, val=x.val*self._ldiag)
        elif mode == self.ADJOINT_TIMES:
            if np.issubdtype(self._ldiag.dtype, np.floating):
                return Field(x.domain, val=x.val*self._ldiag)
            else:
                return Field(x.domain, val=x.val*self._ldiag.conj())
        elif mode == self.INVERSE_TIMES:
            return Field(x.domain, val=x.val/self._ldiag)
        else:
            if np.issubdtype(self._ldiag.dtype, np.floating):
                return Field(x.domain, val=x.val/self._ldiag)
            else:
                return Field(x.domain, val=x.val/self._ldiag.conj())
153

154
155
    @property
    def domain(self):
156
        return self._domain
157

158
    @property
Martin Reinecke's avatar
Martin Reinecke committed
159
    def capability(self):
Martin Reinecke's avatar
Martin Reinecke committed
160
161
        return self._all_ops

162
163
164
165
166
    def _flip_modes(self, trafo):
        ADJ = self.ADJOINT_BIT
        INV = self.INVERSE_BIT

        if trafo == 0:
Martin Reinecke's avatar
Martin Reinecke committed
167
            return self
168
        if trafo == ADJ and np.issubdtype(self._ldiag.dtype, np.floating):
Martin Reinecke's avatar
tweaks    
Martin Reinecke committed
169
            return self
170
        if trafo == ADJ:
171
            return self._from_ldiag((), self._ldiag.conjugate())
172
        elif trafo == INV:
173
            return self._from_ldiag((), 1./self._ldiag)
174
        elif trafo == ADJ | INV:
175
176
            return self._from_ldiag((), 1./self._ldiag.conjugate())
        raise ValueError("invalid operator transformation")
177

178
    def draw_sample(self, from_inverse=False, dtype=np.float64):
179
        if np.issubdtype(self._ldiag.dtype, np.complexfloating):
clienhar's avatar
clienhar committed
180
            raise ValueError("operator not positive definite")
181
182
183
184
        if self._diagmin < 0.:
            raise ValueError("operator not positive definite")
        if self._diagmin == 0. and from_inverse:
            raise ValueError("operator not positive definite")
185
186
        res = Field.from_random(random_type="normal", domain=self._domain,
                                dtype=dtype)
187
188
189
190
        if from_inverse:
            res.local_data[()] /= np.sqrt(self._ldiag)
        else:
            res.local_data[()] *= np.sqrt(self._ldiag)
clienhar's avatar
clienhar committed
191
        return res