power_space.py 9.17 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
15
16
17
#
# Copyright(C) 2013-2017 Max-Planck-Society
#
# NIFTy is being developed at the Max-Planck-Institut fuer Astrophysik
# and financially supported by the Studienstiftung des deutschen Volkes.
Theo Steininger's avatar
Theo Steininger committed
18

19
import numpy as np
Martin Reinecke's avatar
Martin Reinecke committed
20
from .structured_domain import StructuredDomain
Martin Reinecke's avatar
Martin Reinecke committed
21
from .. import dobj
Theo Steininger's avatar
Theo Steininger committed
22
23


Martin Reinecke's avatar
Martin Reinecke committed
24
class PowerSpace(StructuredDomain):
Martin Reinecke's avatar
Martin Reinecke committed
25
    """NIFTy class for spaces of power spectra.
Theo Steininger's avatar
Theo Steininger committed
26

Martin Reinecke's avatar
Martin Reinecke committed
27
    A power space is the result of a projection of a harmonic domain where
Martin Reinecke's avatar
Martin Reinecke committed
28
29
    k-modes of equal length get mapped to one power index.

Theo Steininger's avatar
Theo Steininger committed
30
31
    Parameters
    ----------
Martin Reinecke's avatar
Martin Reinecke committed
32
    harmonic_partner : StructuredDomain
Martin Reinecke's avatar
Martin Reinecke committed
33
34
        The harmonic domain of which this is the power space.
    binbounds : None, or tuple of float (default: None)
Martin Reinecke's avatar
Martin Reinecke committed
35
36
37
        if None:
            There will be as many bins as there are distinct k-vector lengths
            in the harmonic partner space.
Martin Reinecke's avatar
Martin Reinecke committed
38
            The `binbounds` property of the PowerSpace will also be None.
Martin Reinecke's avatar
Martin Reinecke committed
39
40
41
42
        else:
            the bin bounds requested for this PowerSpace. The array
            must be sorted and strictly ascending. The first entry is the right
            boundary of the first bin, and the last entry is the left boundary
Martin Reinecke's avatar
Martin Reinecke committed
43
44
45
            of the last bin, i.e. thee will be `len(binbounds)+1` bins in
            total, with the first and last bins reaching to -+infinity,
            respectively.
Theo Steininger's avatar
Theo Steininger committed
46
    """
47

48
49
    _powerIndexCache = {}

Martin Reinecke's avatar
Martin Reinecke committed
50
    @staticmethod
Martin Reinecke's avatar
PEP8    
Martin Reinecke committed
51
    def linear_binbounds(nbin, first_bound, last_bound):
Martin Reinecke's avatar
Martin Reinecke committed
52
53
        """Produces linearly spaced bin bounds.

Martin Reinecke's avatar
Martin Reinecke committed
54
55
56
57
        This will produce a binbounds array with nbin-1 entries with
        binbounds[0]=first_bound and binbounds[-1]=last_bound and the remaining
        values equidistantly spaced (in linear scale) between these two.

Martin Reinecke's avatar
Martin Reinecke committed
58
        nbin : int
Martin Reinecke's avatar
Martin Reinecke committed
59
            the number of bins
Martin Reinecke's avatar
Martin Reinecke committed
60
        first_bound, last_bound : float
Martin Reinecke's avatar
Martin Reinecke committed
61
62
63
64
65
            the k values for the right boundary of the first bin and the left
            boundary of the last bin, respectively. They are given in length
            units of the harmonic partner space.
        """
        nbin = int(nbin)
66
67
        if nbin < 3:
            raise ValueError("nbin must be at least 3")
Martin Reinecke's avatar
PEP8    
Martin Reinecke committed
68
        return np.linspace(float(first_bound), float(last_bound), nbin-1)
Martin Reinecke's avatar
Martin Reinecke committed
69
70

    @staticmethod
Martin Reinecke's avatar
PEP8    
Martin Reinecke committed
71
    def logarithmic_binbounds(nbin, first_bound, last_bound):
Martin Reinecke's avatar
Martin Reinecke committed
72
73
        """Produces logarithmically spaced bin bounds.

Martin Reinecke's avatar
Martin Reinecke committed
74
75
76
77
78
        This will produce a binbounds array with nbin-1 entries with
        binbounds[0]=first_bound and binbounds[-1]=last_bound and the remaining
        values equidistantly spaced (in natural logarithmic scale)
        between these two.

Martin Reinecke's avatar
Martin Reinecke committed
79
        nbin : int
Martin Reinecke's avatar
Martin Reinecke committed
80
            the number of bins
Martin Reinecke's avatar
Martin Reinecke committed
81
        first_bound, last_bound : float
Martin Reinecke's avatar
Martin Reinecke committed
82
83
84
85
            the k values for the right boundary of the first bin and the left
            boundary of the last bin, respectively. They are given in length
            units of the harmonic partner space.
        """
Martin Reinecke's avatar
Martin Reinecke committed
86
        nbin = int(nbin)
87
88
        if nbin < 3:
            raise ValueError("nbin must be at least 3")
Martin Reinecke's avatar
Martin Reinecke committed
89
90
91
        return np.logspace(np.log(float(first_bound)),
                           np.log(float(last_bound)),
                           nbin-1, base=np.e)
Martin Reinecke's avatar
Martin Reinecke committed
92

93
94
    @staticmethod
    def useful_binbounds(space, logarithmic, nbin=None):
Martin Reinecke's avatar
Martin Reinecke committed
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
        """Produces bin bounds suitable for a given domain.

        This will produce a binbounds array with `nbin-1` entries, if `nbin` is
        supplied, or the maximum number of entries that does not produce empty
        bins, if `nbin` is not supplied.
        The first and last bin boundary are inferred from `space`.

        space : StructuredDomain
            the domain for which the binbounds will be computed.
        logarithmic : bool
            If True bins will have equal size in linear space; otherwise they
            will have equali size in logarithmic space.
        nbin : int, optional
            the number of bins
            If None, the highest possible number of bins will be used
        """
Martin Reinecke's avatar
Martin Reinecke committed
111
        if not (isinstance(space, StructuredDomain) and space.harmonic):
112
113
114
115
116
            raise ValueError("first argument must be a harmonic space.")
        if logarithmic is None and nbin is None:
            return None
        nbin = None if nbin is None else int(nbin)
        logarithmic = bool(logarithmic)
117
        dists = space.get_unique_k_lengths()
118
119
120
121
122
123
124
125
126
127
128
129
        if len(dists) < 3:
            raise ValueError("Space does not have enough unique k lengths")
        lbound = 0.5*(dists[0]+dists[1])
        rbound = 0.5*(dists[-2]+dists[-1])
        dists[0] = lbound
        dists[-1] = rbound
        if logarithmic:
            dists = np.log(dists)
        binsz_min = np.max(np.diff(dists))
        nbin_max = int((dists[-1]-dists[0])/binsz_min)+2
        if nbin is None:
            nbin = nbin_max
130
131
        if nbin < 3:
            raise ValueError("nbin must be at least 3")
132
133
134
135
136
137
138
        if nbin > nbin_max:
            raise ValueError("nbin is too large")
        if logarithmic:
            return PowerSpace.logarithmic_binbounds(nbin, lbound, rbound)
        else:
            return PowerSpace.linear_binbounds(nbin, lbound, rbound)

Martin Reinecke's avatar
Martin Reinecke committed
139
    def __init__(self, harmonic_partner, binbounds=None):
Martin Reinecke's avatar
Martin Reinecke committed
140
        super(PowerSpace, self).__init__()
141
        self._needed_for_hash += ['_harmonic_partner', '_binbounds']
142

Martin Reinecke's avatar
Martin Reinecke committed
143
        if not (isinstance(harmonic_partner, StructuredDomain) and
Martin Reinecke's avatar
Martin Reinecke committed
144
145
                harmonic_partner.harmonic):
            raise ValueError("harmonic_partner must be a harmonic space.")
Martin Reinecke's avatar
Martin Reinecke committed
146
147
148
        if harmonic_partner.scalar_dvol() is None:
            raise ValueError("harmonic partner must have "
                             "scalar volume factors")
149
        self._harmonic_partner = harmonic_partner
Martin Reinecke's avatar
Martin Reinecke committed
150
        pdvol = harmonic_partner.scalar_dvol()
151

Martin Reinecke's avatar
Martin Reinecke committed
152
153
        if binbounds is not None:
            binbounds = tuple(binbounds)
154

Martin Reinecke's avatar
Martin Reinecke committed
155
        key = (harmonic_partner, binbounds)
156
        if self._powerIndexCache.get(key) is None:
157
            k_length_array = self.harmonic_partner.get_k_length_array()
Martin Reinecke's avatar
Martin Reinecke committed
158
159
160
161
162
163
            if binbounds is None:
                tmp = harmonic_partner.get_unique_k_lengths()
                tbb = 0.5*(tmp[:-1]+tmp[1:])
            else:
                tbb = binbounds
            locdat = np.searchsorted(tbb, dobj.local_data(k_length_array.val))
Martin Reinecke's avatar
tweaks    
Martin Reinecke committed
164
            temp_pindex = dobj.from_local_data(
165
166
                k_length_array.val.shape, locdat,
                dobj.distaxis(k_length_array.val))
Martin Reinecke's avatar
Martin Reinecke committed
167
            nbin = len(tbb)+1
Martin Reinecke's avatar
Martin Reinecke committed
168
169
            temp_rho = np.bincount(dobj.local_data(temp_pindex).ravel(),
                                   minlength=nbin)
Martin Reinecke's avatar
Martin Reinecke committed
170
            temp_rho = dobj.np_allreduce_sum(temp_rho)
171
172
            if (temp_rho == 0).any():
                raise ValueError("empty bins detected")
Martin Reinecke's avatar
Martin Reinecke committed
173
174
175
            # The explicit conversion to float64 is necessary because bincount
            # sometimes returns its result as an integer array, even when
            # floating-point weights are present ...
176
177
            temp_k_lengths = np.bincount(
                dobj.local_data(temp_pindex).ravel(),
Martin Reinecke's avatar
Martin Reinecke committed
178
                weights=dobj.local_data(k_length_array.val).ravel(),
Martin Reinecke's avatar
Martin Reinecke committed
179
                minlength=nbin).astype(np.float64, copy=False)
Martin Reinecke's avatar
Martin Reinecke committed
180
            temp_k_lengths = dobj.np_allreduce_sum(temp_k_lengths) / temp_rho
Martin Reinecke's avatar
Martin Reinecke committed
181
            temp_dvol = temp_rho*pdvol
Martin Reinecke's avatar
Martin Reinecke committed
182
183
            self._powerIndexCache[key] = (binbounds, temp_pindex,
                                          temp_k_lengths, temp_dvol)
184

Martin Reinecke's avatar
Martin Reinecke committed
185
        (self._binbounds, self._pindex, self._k_lengths, self._dvol) = \
186
187
            self._powerIndexCache[key]

188
    def __repr__(self):
Martin Reinecke's avatar
stage1    
Martin Reinecke committed
189
190
        return ("PowerSpace(harmonic_partner=%r, binbounds=%r)"
                % (self.harmonic_partner, self._binbounds))
191

192
193
    @property
    def harmonic(self):
Martin Reinecke's avatar
Martin Reinecke committed
194
        """bool : Always False for this class."""
195
        return False
196

197
198
    @property
    def shape(self):
Martin Reinecke's avatar
Martin Reinecke committed
199
        return self.k_lengths.shape
200

201
    @property
Martin Reinecke's avatar
Martin Reinecke committed
202
    def size(self):
203
204
        return self.shape[0]

205
    def scalar_dvol(self):
Martin Reinecke's avatar
Martin Reinecke committed
206
207
208
209
        return None

    def dvol(self):
        return self._dvol
210

211
    @property
212
    def harmonic_partner(self):
Martin Reinecke's avatar
Martin Reinecke committed
213
        """StructuredDomain : the harmonic domain associated with `self`."""
214
        return self._harmonic_partner
215
216

    @property
Martin Reinecke's avatar
Martin Reinecke committed
217
    def binbounds(self):
Martin Reinecke's avatar
Martin Reinecke committed
218
219
220
221
        """None or tuple of float : inner bin boundaries

        The boundaries between bins, starting with the right boundary of the
        first bin, up to the left boundary of the last bin.
Martin Reinecke's avatar
Martin Reinecke committed
222
223

        `None` is used to indicate natural binning.
Martin Reinecke's avatar
Martin Reinecke committed
224
        """
Martin Reinecke's avatar
Martin Reinecke committed
225
        return self._binbounds
226
227
228

    @property
    def pindex(self):
Martin Reinecke's avatar
Martin Reinecke committed
229
230
231
        """data_object : bin indices

        Bin index for every pixel in the harmonic partner.
Theo Steininger's avatar
Theo Steininger committed
232
        """
233
234
235
        return self._pindex

    @property
Martin Reinecke's avatar
Martin Reinecke committed
236
    def k_lengths(self):
Martin Reinecke's avatar
Martin Reinecke committed
237
        """numpy.ndarray(float) : sorted array of all k-vector lengths."""
Martin Reinecke's avatar
Martin Reinecke committed
238
        return self._k_lengths