line |
stmt |
bran |
cond |
sub |
pod |
time |
code |
1
|
|
|
|
|
|
|
|
2
|
|
|
|
|
|
|
# |
3
|
|
|
|
|
|
|
# GENERATED WITH PDLA::PP! Don't modify! |
4
|
|
|
|
|
|
|
# |
5
|
|
|
|
|
|
|
package PDLA::Primitive; |
6
|
|
|
|
|
|
|
|
7
|
|
|
|
|
|
|
@EXPORT_OK = qw( PDLA::PP inner PDLA::PP outer matmult PDLA::PP matmult PDLA::PP innerwt PDLA::PP inner2 PDLA::PP inner2d PDLA::PP inner2t PDLA::PP crossp PDLA::PP norm PDLA::PP indadd PDLA::PP conv1d PDLA::PP in uniq uniqind uniqvec PDLA::PP hclip PDLA::PP lclip clip PDLA::PP clip PDLA::PP wtstat PDLA::PP statsover stats PDLA::PP histogram PDLA::PP whistogram PDLA::PP histogram2d PDLA::PP whistogram2d PDLA::PP fibonacci PDLA::PP append PDLA::PP axisvalues PDLA::PP random PDLA::PP randsym grandom vsearch PDLA::PP vsearch_sample PDLA::PP vsearch_insert_leftmost PDLA::PP vsearch_insert_rightmost PDLA::PP vsearch_match PDLA::PP vsearch_bin_inclusive PDLA::PP vsearch_bin_exclusive PDLA::PP interpolate interpol interpND one2nd PDLA::PP which PDLA::PP which_both where whereND whichND setops intersect ); |
8
|
|
|
|
|
|
|
%EXPORT_TAGS = (Func=>[@EXPORT_OK]); |
9
|
|
|
|
|
|
|
|
10
|
78
|
|
|
78
|
|
560
|
use PDLA::Core; |
|
78
|
|
|
|
|
169
|
|
|
78
|
|
|
|
|
477
|
|
11
|
78
|
|
|
78
|
|
535
|
use PDLA::Exporter; |
|
78
|
|
|
|
|
159
|
|
|
78
|
|
|
|
|
421
|
|
12
|
78
|
|
|
78
|
|
380
|
use DynaLoader; |
|
78
|
|
|
|
|
164
|
|
|
78
|
|
|
|
|
5419
|
|
13
|
|
|
|
|
|
|
|
14
|
|
|
|
|
|
|
|
15
|
|
|
|
|
|
|
|
16
|
|
|
|
|
|
|
|
17
|
|
|
|
|
|
|
@ISA = ( 'PDLA::Exporter','DynaLoader' ); |
18
|
|
|
|
|
|
|
push @PDLA::Core::PP, __PACKAGE__; |
19
|
|
|
|
|
|
|
bootstrap PDLA::Primitive ; |
20
|
|
|
|
|
|
|
|
21
|
|
|
|
|
|
|
|
22
|
|
|
|
|
|
|
|
23
|
|
|
|
|
|
|
|
24
|
|
|
|
|
|
|
|
25
|
78
|
|
|
78
|
|
46727
|
use PDLA::Slices; |
|
78
|
|
|
|
|
215
|
|
|
78
|
|
|
|
|
544
|
|
26
|
78
|
|
|
78
|
|
613
|
use Carp; |
|
78
|
|
|
|
|
173
|
|
|
78
|
|
|
|
|
45654
|
|
27
|
|
|
|
|
|
|
|
28
|
|
|
|
|
|
|
=head1 NAME |
29
|
|
|
|
|
|
|
|
30
|
|
|
|
|
|
|
PDLA::Primitive - primitive operations for pdl |
31
|
|
|
|
|
|
|
|
32
|
|
|
|
|
|
|
=head1 DESCRIPTION |
33
|
|
|
|
|
|
|
|
34
|
|
|
|
|
|
|
This module provides some primitive and useful functions defined |
35
|
|
|
|
|
|
|
using PDLA::PP and able to use the new indexing tricks. |
36
|
|
|
|
|
|
|
|
37
|
|
|
|
|
|
|
See L for how to use indices creatively. |
38
|
|
|
|
|
|
|
For explanation of the signature format, see L. |
39
|
|
|
|
|
|
|
|
40
|
|
|
|
|
|
|
=head1 SYNOPSIS |
41
|
|
|
|
|
|
|
|
42
|
|
|
|
|
|
|
# Pulls in PDLA::Primitive, among other modules. |
43
|
|
|
|
|
|
|
use PDLA; |
44
|
|
|
|
|
|
|
|
45
|
|
|
|
|
|
|
# Only pull in PDLA::Primitive: |
46
|
|
|
|
|
|
|
use PDLA::Primitive; |
47
|
|
|
|
|
|
|
|
48
|
|
|
|
|
|
|
=cut |
49
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
51
|
|
|
|
|
|
|
|
52
|
|
|
|
|
|
|
|
53
|
|
|
|
|
|
|
|
54
|
|
|
|
|
|
|
|
55
|
|
|
|
|
|
|
|
56
|
|
|
|
|
|
|
=head1 FUNCTIONS |
57
|
|
|
|
|
|
|
|
58
|
|
|
|
|
|
|
|
59
|
|
|
|
|
|
|
|
60
|
|
|
|
|
|
|
=cut |
61
|
|
|
|
|
|
|
|
62
|
|
|
|
|
|
|
|
63
|
|
|
|
|
|
|
|
64
|
|
|
|
|
|
|
|
65
|
|
|
|
|
|
|
|
66
|
|
|
|
|
|
|
|
67
|
|
|
|
|
|
|
=head2 inner |
68
|
|
|
|
|
|
|
|
69
|
|
|
|
|
|
|
=for sig |
70
|
|
|
|
|
|
|
|
71
|
|
|
|
|
|
|
Signature: (a(n); b(n); [o]c()) |
72
|
|
|
|
|
|
|
|
73
|
|
|
|
|
|
|
|
74
|
|
|
|
|
|
|
|
75
|
|
|
|
|
|
|
=for ref |
76
|
|
|
|
|
|
|
|
77
|
|
|
|
|
|
|
Inner product over one dimension |
78
|
|
|
|
|
|
|
|
79
|
|
|
|
|
|
|
c = sum_i a_i * b_i |
80
|
|
|
|
|
|
|
|
81
|
|
|
|
|
|
|
|
82
|
|
|
|
|
|
|
|
83
|
|
|
|
|
|
|
=for bad |
84
|
|
|
|
|
|
|
|
85
|
|
|
|
|
|
|
=for bad |
86
|
|
|
|
|
|
|
|
87
|
|
|
|
|
|
|
If C contains only bad data, |
88
|
|
|
|
|
|
|
C is set bad. Otherwise C will have its bad flag cleared, |
89
|
|
|
|
|
|
|
as it will not contain any bad values. |
90
|
|
|
|
|
|
|
|
91
|
|
|
|
|
|
|
|
92
|
|
|
|
|
|
|
|
93
|
|
|
|
|
|
|
=cut |
94
|
|
|
|
|
|
|
|
95
|
|
|
|
|
|
|
|
96
|
|
|
|
|
|
|
|
97
|
|
|
|
|
|
|
|
98
|
|
|
|
|
|
|
|
99
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
*inner = \&PDLA::inner; |
101
|
|
|
|
|
|
|
|
102
|
|
|
|
|
|
|
|
103
|
|
|
|
|
|
|
|
104
|
|
|
|
|
|
|
|
105
|
|
|
|
|
|
|
|
106
|
|
|
|
|
|
|
=head2 outer |
107
|
|
|
|
|
|
|
|
108
|
|
|
|
|
|
|
=for sig |
109
|
|
|
|
|
|
|
|
110
|
|
|
|
|
|
|
Signature: (a(n); b(m); [o]c(n,m)) |
111
|
|
|
|
|
|
|
|
112
|
|
|
|
|
|
|
|
113
|
|
|
|
|
|
|
|
114
|
|
|
|
|
|
|
=for ref |
115
|
|
|
|
|
|
|
|
116
|
|
|
|
|
|
|
outer product over one dimension |
117
|
|
|
|
|
|
|
|
118
|
|
|
|
|
|
|
Naturally, it is possible to achieve the effects of outer |
119
|
|
|
|
|
|
|
product simply by threading over the "C<*>" |
120
|
|
|
|
|
|
|
operator but this function is provided for convenience. |
121
|
|
|
|
|
|
|
|
122
|
|
|
|
|
|
|
|
123
|
|
|
|
|
|
|
|
124
|
|
|
|
|
|
|
=for bad |
125
|
|
|
|
|
|
|
|
126
|
|
|
|
|
|
|
outer processes bad values. |
127
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
128
|
|
|
|
|
|
|
|
129
|
|
|
|
|
|
|
|
130
|
|
|
|
|
|
|
=cut |
131
|
|
|
|
|
|
|
|
132
|
|
|
|
|
|
|
|
133
|
|
|
|
|
|
|
|
134
|
|
|
|
|
|
|
|
135
|
|
|
|
|
|
|
|
136
|
|
|
|
|
|
|
|
137
|
|
|
|
|
|
|
*outer = \&PDLA::outer; |
138
|
|
|
|
|
|
|
|
139
|
|
|
|
|
|
|
|
140
|
|
|
|
|
|
|
|
141
|
|
|
|
|
|
|
|
142
|
|
|
|
|
|
|
=head2 x |
143
|
|
|
|
|
|
|
|
144
|
|
|
|
|
|
|
=for sig |
145
|
|
|
|
|
|
|
|
146
|
|
|
|
|
|
|
Signature: (a(i,z), b(x,i),[o]c(x,z)) |
147
|
|
|
|
|
|
|
|
148
|
|
|
|
|
|
|
=for ref |
149
|
|
|
|
|
|
|
|
150
|
|
|
|
|
|
|
Matrix multiplication |
151
|
|
|
|
|
|
|
|
152
|
|
|
|
|
|
|
PDLA overloads the C operator (normally the repeat operator) for |
153
|
|
|
|
|
|
|
matrix multiplication. The number of columns (size of the 0 |
154
|
|
|
|
|
|
|
dimension) in the left-hand argument must normally equal the number of |
155
|
|
|
|
|
|
|
rows (size of the 1 dimension) in the right-hand argument. |
156
|
|
|
|
|
|
|
|
157
|
|
|
|
|
|
|
Row vectors are represented as (N x 1) two-dimensional PDLAs, or you |
158
|
|
|
|
|
|
|
may be sloppy and use a one-dimensional PDLA. Column vectors are |
159
|
|
|
|
|
|
|
represented as (1 x N) two-dimensional PDLAs. |
160
|
|
|
|
|
|
|
|
161
|
|
|
|
|
|
|
Threading occurs in the usual way, but as both the 0 and 1 dimension |
162
|
|
|
|
|
|
|
(if present) are included in the operation, you must be sure that |
163
|
|
|
|
|
|
|
you don't try to thread over either of those dims. |
164
|
|
|
|
|
|
|
|
165
|
|
|
|
|
|
|
EXAMPLES |
166
|
|
|
|
|
|
|
|
167
|
|
|
|
|
|
|
Here are some simple ways to define vectors and matrices: |
168
|
|
|
|
|
|
|
|
169
|
|
|
|
|
|
|
pdla> $r = pdl(1,2); # A row vector |
170
|
|
|
|
|
|
|
pdla> $c = pdl([[3],[4]]); # A column vector |
171
|
|
|
|
|
|
|
pdla> $c = pdl(3,4)->(*1); # A column vector, using NiceSlice |
172
|
|
|
|
|
|
|
pdla> $m = pdl([[1,2],[3,4]]); # A 2x2 matrix |
173
|
|
|
|
|
|
|
|
174
|
|
|
|
|
|
|
Now that we have a few objects prepared, here is how to |
175
|
|
|
|
|
|
|
matrix-multiply them: |
176
|
|
|
|
|
|
|
|
177
|
|
|
|
|
|
|
pdla> print $r x $m # row x matrix = row |
178
|
|
|
|
|
|
|
[ |
179
|
|
|
|
|
|
|
[ 7 10] |
180
|
|
|
|
|
|
|
] |
181
|
|
|
|
|
|
|
|
182
|
|
|
|
|
|
|
pdla> print $m x $r # matrix x row = ERROR |
183
|
|
|
|
|
|
|
PDLA: Dim mismatch in matmult of [2x2] x [2x1]: 2 != 1 |
184
|
|
|
|
|
|
|
|
185
|
|
|
|
|
|
|
pdla> print $m x $c # matrix x column = column |
186
|
|
|
|
|
|
|
[ |
187
|
|
|
|
|
|
|
[ 5] |
188
|
|
|
|
|
|
|
[11] |
189
|
|
|
|
|
|
|
] |
190
|
|
|
|
|
|
|
|
191
|
|
|
|
|
|
|
pdla> print $m x 2 # Trivial case: scalar mult. |
192
|
|
|
|
|
|
|
[ |
193
|
|
|
|
|
|
|
[2 4] |
194
|
|
|
|
|
|
|
[6 8] |
195
|
|
|
|
|
|
|
] |
196
|
|
|
|
|
|
|
|
197
|
|
|
|
|
|
|
pdla> print $r x $c # row x column = scalar |
198
|
|
|
|
|
|
|
[ |
199
|
|
|
|
|
|
|
[11] |
200
|
|
|
|
|
|
|
] |
201
|
|
|
|
|
|
|
|
202
|
|
|
|
|
|
|
pdla> print $c x $r # column x row = matrix |
203
|
|
|
|
|
|
|
[ |
204
|
|
|
|
|
|
|
[3 6] |
205
|
|
|
|
|
|
|
[4 8] |
206
|
|
|
|
|
|
|
] |
207
|
|
|
|
|
|
|
|
208
|
|
|
|
|
|
|
|
209
|
|
|
|
|
|
|
INTERNALS |
210
|
|
|
|
|
|
|
|
211
|
|
|
|
|
|
|
The mechanics of the multiplication are carried out by the |
212
|
|
|
|
|
|
|
L method. |
213
|
|
|
|
|
|
|
|
214
|
|
|
|
|
|
|
=cut |
215
|
|
|
|
|
|
|
|
216
|
|
|
|
|
|
|
|
217
|
|
|
|
|
|
|
|
218
|
|
|
|
|
|
|
|
219
|
|
|
|
|
|
|
|
220
|
|
|
|
|
|
|
=head2 matmult |
221
|
|
|
|
|
|
|
|
222
|
|
|
|
|
|
|
=for sig |
223
|
|
|
|
|
|
|
|
224
|
|
|
|
|
|
|
Signature: (a(t,h); b(w,t); [o]c(w,h)) |
225
|
|
|
|
|
|
|
|
226
|
|
|
|
|
|
|
=for ref |
227
|
|
|
|
|
|
|
|
228
|
|
|
|
|
|
|
Matrix multiplication |
229
|
|
|
|
|
|
|
|
230
|
|
|
|
|
|
|
Notionally, matrix multiplication $x x $y is equivalent to the |
231
|
|
|
|
|
|
|
threading expression |
232
|
|
|
|
|
|
|
|
233
|
|
|
|
|
|
|
$x->dummy(1)->inner($y->xchg(0,1)->dummy(2),$c); |
234
|
|
|
|
|
|
|
|
235
|
|
|
|
|
|
|
but for large matrices that breaks CPU cache and is slow. Instead, |
236
|
|
|
|
|
|
|
matmult calculates its result in 32x32x32 tiles, to keep the memory |
237
|
|
|
|
|
|
|
footprint within cache as long as possible on most modern CPUs. |
238
|
|
|
|
|
|
|
|
239
|
|
|
|
|
|
|
For usage, see L, a description of the overloaded 'x' operator |
240
|
|
|
|
|
|
|
|
241
|
|
|
|
|
|
|
|
242
|
|
|
|
|
|
|
|
243
|
|
|
|
|
|
|
=for bad |
244
|
|
|
|
|
|
|
|
245
|
|
|
|
|
|
|
matmult ignores the bad-value flag of the input piddles. |
246
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
247
|
|
|
|
|
|
|
|
248
|
|
|
|
|
|
|
|
249
|
|
|
|
|
|
|
=cut |
250
|
|
|
|
|
|
|
|
251
|
|
|
|
|
|
|
|
252
|
|
|
|
|
|
|
|
253
|
|
|
|
|
|
|
|
254
|
|
|
|
|
|
|
sub PDLA::matmult { |
255
|
20
|
|
|
20
|
0
|
935
|
my ($x,$y,$c) = @_; |
256
|
|
|
|
|
|
|
|
257
|
20
|
100
|
|
|
|
34
|
$y = pdl($y) unless eval { $y->isa('PDLA') }; |
|
20
|
|
|
|
|
100
|
|
258
|
20
|
100
|
|
|
|
41
|
$c = PDLA->null unless eval { $c->isa('PDLA') }; |
|
20
|
|
|
|
|
89
|
|
259
|
|
|
|
|
|
|
|
260
|
20
|
|
|
|
|
77
|
while($x->getndims < 2) {$x = $x->dummy(-1)} |
|
2
|
|
|
|
|
8
|
|
261
|
20
|
|
|
|
|
76
|
while($y->getndims < 2) {$y = $y->dummy(-1)} |
|
2
|
|
|
|
|
15
|
|
262
|
|
|
|
|
|
|
|
263
|
20
|
100
|
66
|
|
|
201
|
return ($c .= $x * $y) if( ($x->dim(0)==1 && $x->dim(1)==1) || |
|
|
|
66
|
|
|
|
|
|
|
|
100
|
|
|
|
|
264
|
|
|
|
|
|
|
($y->dim(0)==1 && $y->dim(1)==1) ); |
265
|
18
|
100
|
|
|
|
65
|
if($y->dim(1) != $x->dim(0)) { |
266
|
1
|
|
|
|
|
18
|
barf(sprintf("Dim mismatch in matmult of [%dx%d] x [%dx%d]: %d != %d",$x->dim(0),$x->dim(1),$y->dim(0),$y->dim(1),$x->dim(0),$y->dim(1))); |
267
|
|
|
|
|
|
|
} |
268
|
17
|
|
|
|
|
422
|
PDLA::_matmult_int($x,$y,$c); |
269
|
17
|
|
|
|
|
68
|
$c; |
270
|
|
|
|
|
|
|
} |
271
|
|
|
|
|
|
|
|
272
|
|
|
|
|
|
|
|
273
|
|
|
|
|
|
|
*matmult = \&PDLA::matmult; |
274
|
|
|
|
|
|
|
|
275
|
|
|
|
|
|
|
|
276
|
|
|
|
|
|
|
|
277
|
|
|
|
|
|
|
|
278
|
|
|
|
|
|
|
|
279
|
|
|
|
|
|
|
=head2 innerwt |
280
|
|
|
|
|
|
|
|
281
|
|
|
|
|
|
|
=for sig |
282
|
|
|
|
|
|
|
|
283
|
|
|
|
|
|
|
Signature: (a(n); b(n); c(n); [o]d()) |
284
|
|
|
|
|
|
|
|
285
|
|
|
|
|
|
|
|
286
|
|
|
|
|
|
|
|
287
|
|
|
|
|
|
|
=for ref |
288
|
|
|
|
|
|
|
|
289
|
|
|
|
|
|
|
Weighted (i.e. triple) inner product |
290
|
|
|
|
|
|
|
|
291
|
|
|
|
|
|
|
d = sum_i a(i) b(i) c(i) |
292
|
|
|
|
|
|
|
|
293
|
|
|
|
|
|
|
|
294
|
|
|
|
|
|
|
|
295
|
|
|
|
|
|
|
=for bad |
296
|
|
|
|
|
|
|
|
297
|
|
|
|
|
|
|
innerwt processes bad values. |
298
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
299
|
|
|
|
|
|
|
|
300
|
|
|
|
|
|
|
|
301
|
|
|
|
|
|
|
=cut |
302
|
|
|
|
|
|
|
|
303
|
|
|
|
|
|
|
|
304
|
|
|
|
|
|
|
|
305
|
|
|
|
|
|
|
|
306
|
|
|
|
|
|
|
|
307
|
|
|
|
|
|
|
|
308
|
|
|
|
|
|
|
*innerwt = \&PDLA::innerwt; |
309
|
|
|
|
|
|
|
|
310
|
|
|
|
|
|
|
|
311
|
|
|
|
|
|
|
|
312
|
|
|
|
|
|
|
|
313
|
|
|
|
|
|
|
|
314
|
|
|
|
|
|
|
=head2 inner2 |
315
|
|
|
|
|
|
|
|
316
|
|
|
|
|
|
|
=for sig |
317
|
|
|
|
|
|
|
|
318
|
|
|
|
|
|
|
Signature: (a(n); b(n,m); c(m); [o]d()) |
319
|
|
|
|
|
|
|
|
320
|
|
|
|
|
|
|
|
321
|
|
|
|
|
|
|
|
322
|
|
|
|
|
|
|
=for ref |
323
|
|
|
|
|
|
|
|
324
|
|
|
|
|
|
|
Inner product of two vectors and a matrix |
325
|
|
|
|
|
|
|
|
326
|
|
|
|
|
|
|
d = sum_ij a(i) b(i,j) c(j) |
327
|
|
|
|
|
|
|
|
328
|
|
|
|
|
|
|
Note that you should probably not thread over C and C since that would be |
329
|
|
|
|
|
|
|
very wasteful. Instead, you should use a temporary for C. |
330
|
|
|
|
|
|
|
|
331
|
|
|
|
|
|
|
|
332
|
|
|
|
|
|
|
|
333
|
|
|
|
|
|
|
=for bad |
334
|
|
|
|
|
|
|
|
335
|
|
|
|
|
|
|
inner2 processes bad values. |
336
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
337
|
|
|
|
|
|
|
|
338
|
|
|
|
|
|
|
|
339
|
|
|
|
|
|
|
=cut |
340
|
|
|
|
|
|
|
|
341
|
|
|
|
|
|
|
|
342
|
|
|
|
|
|
|
|
343
|
|
|
|
|
|
|
|
344
|
|
|
|
|
|
|
|
345
|
|
|
|
|
|
|
|
346
|
|
|
|
|
|
|
*inner2 = \&PDLA::inner2; |
347
|
|
|
|
|
|
|
|
348
|
|
|
|
|
|
|
|
349
|
|
|
|
|
|
|
|
350
|
|
|
|
|
|
|
|
351
|
|
|
|
|
|
|
|
352
|
|
|
|
|
|
|
=head2 inner2d |
353
|
|
|
|
|
|
|
|
354
|
|
|
|
|
|
|
=for sig |
355
|
|
|
|
|
|
|
|
356
|
|
|
|
|
|
|
Signature: (a(n,m); b(n,m); [o]c()) |
357
|
|
|
|
|
|
|
|
358
|
|
|
|
|
|
|
|
359
|
|
|
|
|
|
|
|
360
|
|
|
|
|
|
|
=for ref |
361
|
|
|
|
|
|
|
|
362
|
|
|
|
|
|
|
Inner product over 2 dimensions. |
363
|
|
|
|
|
|
|
|
364
|
|
|
|
|
|
|
Equivalent to |
365
|
|
|
|
|
|
|
|
366
|
|
|
|
|
|
|
$c = inner($x->clump(2), $y->clump(2)) |
367
|
|
|
|
|
|
|
|
368
|
|
|
|
|
|
|
|
369
|
|
|
|
|
|
|
|
370
|
|
|
|
|
|
|
=for bad |
371
|
|
|
|
|
|
|
|
372
|
|
|
|
|
|
|
inner2d processes bad values. |
373
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
374
|
|
|
|
|
|
|
|
375
|
|
|
|
|
|
|
|
376
|
|
|
|
|
|
|
=cut |
377
|
|
|
|
|
|
|
|
378
|
|
|
|
|
|
|
|
379
|
|
|
|
|
|
|
|
380
|
|
|
|
|
|
|
|
381
|
|
|
|
|
|
|
|
382
|
|
|
|
|
|
|
|
383
|
|
|
|
|
|
|
*inner2d = \&PDLA::inner2d; |
384
|
|
|
|
|
|
|
|
385
|
|
|
|
|
|
|
|
386
|
|
|
|
|
|
|
|
387
|
|
|
|
|
|
|
|
388
|
|
|
|
|
|
|
|
389
|
|
|
|
|
|
|
=head2 inner2t |
390
|
|
|
|
|
|
|
|
391
|
|
|
|
|
|
|
=for sig |
392
|
|
|
|
|
|
|
|
393
|
|
|
|
|
|
|
Signature: (a(j,n); b(n,m); c(m,k); [t]tmp(n,k); [o]d(j,k))) |
394
|
|
|
|
|
|
|
|
395
|
|
|
|
|
|
|
|
396
|
|
|
|
|
|
|
|
397
|
|
|
|
|
|
|
=for ref |
398
|
|
|
|
|
|
|
|
399
|
|
|
|
|
|
|
Efficient Triple matrix product C |
400
|
|
|
|
|
|
|
|
401
|
|
|
|
|
|
|
Efficiency comes from by using the temporary C. This operation only |
402
|
|
|
|
|
|
|
scales as C whereas threading using L would scale |
403
|
|
|
|
|
|
|
as C. |
404
|
|
|
|
|
|
|
|
405
|
|
|
|
|
|
|
The reason for having this routine is that you do not need to |
406
|
|
|
|
|
|
|
have the same thread-dimensions for C as for the other arguments, |
407
|
|
|
|
|
|
|
which in case of large numbers of matrices makes this much more |
408
|
|
|
|
|
|
|
memory-efficient. |
409
|
|
|
|
|
|
|
|
410
|
|
|
|
|
|
|
It is hoped that things like this could be taken care of as a kind of |
411
|
|
|
|
|
|
|
closures at some point. |
412
|
|
|
|
|
|
|
|
413
|
|
|
|
|
|
|
|
414
|
|
|
|
|
|
|
|
415
|
|
|
|
|
|
|
=for bad |
416
|
|
|
|
|
|
|
|
417
|
|
|
|
|
|
|
inner2t processes bad values. |
418
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
419
|
|
|
|
|
|
|
|
420
|
|
|
|
|
|
|
|
421
|
|
|
|
|
|
|
=cut |
422
|
|
|
|
|
|
|
|
423
|
|
|
|
|
|
|
|
424
|
|
|
|
|
|
|
|
425
|
|
|
|
|
|
|
|
426
|
|
|
|
|
|
|
|
427
|
|
|
|
|
|
|
|
428
|
|
|
|
|
|
|
*inner2t = \&PDLA::inner2t; |
429
|
|
|
|
|
|
|
|
430
|
|
|
|
|
|
|
|
431
|
|
|
|
|
|
|
|
432
|
|
|
|
|
|
|
|
433
|
|
|
|
|
|
|
|
434
|
|
|
|
|
|
|
=head2 crossp |
435
|
|
|
|
|
|
|
|
436
|
|
|
|
|
|
|
=for sig |
437
|
|
|
|
|
|
|
|
438
|
|
|
|
|
|
|
Signature: (a(tri=3); b(tri); [o] c(tri)) |
439
|
|
|
|
|
|
|
|
440
|
|
|
|
|
|
|
|
441
|
|
|
|
|
|
|
=for ref |
442
|
|
|
|
|
|
|
|
443
|
|
|
|
|
|
|
Cross product of two 3D vectors |
444
|
|
|
|
|
|
|
|
445
|
|
|
|
|
|
|
After |
446
|
|
|
|
|
|
|
|
447
|
|
|
|
|
|
|
=for example |
448
|
|
|
|
|
|
|
|
449
|
|
|
|
|
|
|
$c = crossp $x, $y |
450
|
|
|
|
|
|
|
|
451
|
|
|
|
|
|
|
the inner product C<$c*$x> and C<$c*$y> will be zero, i.e. C<$c> is |
452
|
|
|
|
|
|
|
orthogonal to C<$x> and C<$y> |
453
|
|
|
|
|
|
|
|
454
|
|
|
|
|
|
|
|
455
|
|
|
|
|
|
|
|
456
|
|
|
|
|
|
|
=for bad |
457
|
|
|
|
|
|
|
|
458
|
|
|
|
|
|
|
crossp does not process bad values. |
459
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
460
|
|
|
|
|
|
|
|
461
|
|
|
|
|
|
|
|
462
|
|
|
|
|
|
|
=cut |
463
|
|
|
|
|
|
|
|
464
|
|
|
|
|
|
|
|
465
|
|
|
|
|
|
|
|
466
|
|
|
|
|
|
|
|
467
|
|
|
|
|
|
|
|
468
|
|
|
|
|
|
|
|
469
|
|
|
|
|
|
|
*crossp = \&PDLA::crossp; |
470
|
|
|
|
|
|
|
|
471
|
|
|
|
|
|
|
|
472
|
|
|
|
|
|
|
|
473
|
|
|
|
|
|
|
|
474
|
|
|
|
|
|
|
|
475
|
|
|
|
|
|
|
=head2 norm |
476
|
|
|
|
|
|
|
|
477
|
|
|
|
|
|
|
=for sig |
478
|
|
|
|
|
|
|
|
479
|
|
|
|
|
|
|
Signature: (vec(n); [o] norm(n)) |
480
|
|
|
|
|
|
|
|
481
|
|
|
|
|
|
|
=for ref |
482
|
|
|
|
|
|
|
|
483
|
|
|
|
|
|
|
Normalises a vector to unit Euclidean length |
484
|
|
|
|
|
|
|
|
485
|
|
|
|
|
|
|
=for bad |
486
|
|
|
|
|
|
|
|
487
|
|
|
|
|
|
|
norm processes bad values. |
488
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
489
|
|
|
|
|
|
|
|
490
|
|
|
|
|
|
|
|
491
|
|
|
|
|
|
|
=cut |
492
|
|
|
|
|
|
|
|
493
|
|
|
|
|
|
|
|
494
|
|
|
|
|
|
|
|
495
|
|
|
|
|
|
|
|
496
|
|
|
|
|
|
|
|
497
|
|
|
|
|
|
|
|
498
|
|
|
|
|
|
|
*norm = \&PDLA::norm; |
499
|
|
|
|
|
|
|
|
500
|
|
|
|
|
|
|
|
501
|
|
|
|
|
|
|
|
502
|
|
|
|
|
|
|
|
503
|
|
|
|
|
|
|
|
504
|
|
|
|
|
|
|
=head2 indadd |
505
|
|
|
|
|
|
|
|
506
|
|
|
|
|
|
|
=for sig |
507
|
|
|
|
|
|
|
|
508
|
|
|
|
|
|
|
Signature: (a(); indx ind(); [o] sum(m)) |
509
|
|
|
|
|
|
|
|
510
|
|
|
|
|
|
|
|
511
|
|
|
|
|
|
|
|
512
|
|
|
|
|
|
|
=for ref |
513
|
|
|
|
|
|
|
|
514
|
|
|
|
|
|
|
Threaded Index Add: Add C to the C element of C, i.e: |
515
|
|
|
|
|
|
|
|
516
|
|
|
|
|
|
|
sum(ind) += a |
517
|
|
|
|
|
|
|
|
518
|
|
|
|
|
|
|
=for example |
519
|
|
|
|
|
|
|
|
520
|
|
|
|
|
|
|
Simple Example: |
521
|
|
|
|
|
|
|
|
522
|
|
|
|
|
|
|
$x = 2; |
523
|
|
|
|
|
|
|
$ind = 3; |
524
|
|
|
|
|
|
|
$sum = zeroes(10); |
525
|
|
|
|
|
|
|
indadd($x,$ind, $sum); |
526
|
|
|
|
|
|
|
print $sum |
527
|
|
|
|
|
|
|
#Result: ( 2 added to element 3 of $sum) |
528
|
|
|
|
|
|
|
# [0 0 0 2 0 0 0 0 0 0] |
529
|
|
|
|
|
|
|
|
530
|
|
|
|
|
|
|
Threaded Example: |
531
|
|
|
|
|
|
|
|
532
|
|
|
|
|
|
|
$x = pdl( 1,2,3); |
533
|
|
|
|
|
|
|
$ind = pdl( 1,4,6); |
534
|
|
|
|
|
|
|
$sum = zeroes(10); |
535
|
|
|
|
|
|
|
indadd($x,$ind, $sum); |
536
|
|
|
|
|
|
|
print $sum."\n"; |
537
|
|
|
|
|
|
|
#Result: ( 1, 2, and 3 added to elements 1,4,6 $sum) |
538
|
|
|
|
|
|
|
# [0 1 0 0 2 0 3 0 0 0] |
539
|
|
|
|
|
|
|
|
540
|
|
|
|
|
|
|
|
541
|
|
|
|
|
|
|
|
542
|
|
|
|
|
|
|
=for bad |
543
|
|
|
|
|
|
|
|
544
|
|
|
|
|
|
|
=for bad |
545
|
|
|
|
|
|
|
|
546
|
|
|
|
|
|
|
The routine barfs if any of the indices are bad. |
547
|
|
|
|
|
|
|
|
548
|
|
|
|
|
|
|
|
549
|
|
|
|
|
|
|
|
550
|
|
|
|
|
|
|
=cut |
551
|
|
|
|
|
|
|
|
552
|
|
|
|
|
|
|
|
553
|
|
|
|
|
|
|
|
554
|
|
|
|
|
|
|
|
555
|
|
|
|
|
|
|
|
556
|
|
|
|
|
|
|
|
557
|
|
|
|
|
|
|
*indadd = \&PDLA::indadd; |
558
|
|
|
|
|
|
|
|
559
|
|
|
|
|
|
|
|
560
|
|
|
|
|
|
|
|
561
|
|
|
|
|
|
|
|
562
|
|
|
|
|
|
|
|
563
|
|
|
|
|
|
|
=head2 conv1d |
564
|
|
|
|
|
|
|
|
565
|
|
|
|
|
|
|
=for sig |
566
|
|
|
|
|
|
|
|
567
|
|
|
|
|
|
|
Signature: (a(m); kern(p); [o]b(m); int reflect) |
568
|
|
|
|
|
|
|
|
569
|
|
|
|
|
|
|
|
570
|
|
|
|
|
|
|
=for ref |
571
|
|
|
|
|
|
|
|
572
|
|
|
|
|
|
|
1D convolution along first dimension |
573
|
|
|
|
|
|
|
|
574
|
|
|
|
|
|
|
The m-th element of the discrete convolution of an input piddle |
575
|
|
|
|
|
|
|
C<$a> of size C<$M>, and a kernel piddle C<$kern> of size C<$P>, is |
576
|
|
|
|
|
|
|
calculated as |
577
|
|
|
|
|
|
|
|
578
|
|
|
|
|
|
|
n = ($P-1)/2 |
579
|
|
|
|
|
|
|
==== |
580
|
|
|
|
|
|
|
\ |
581
|
|
|
|
|
|
|
($a conv1d $kern)[m] = > $a_ext[m - n] * $kern[n] |
582
|
|
|
|
|
|
|
/ |
583
|
|
|
|
|
|
|
==== |
584
|
|
|
|
|
|
|
n = -($P-1)/2 |
585
|
|
|
|
|
|
|
|
586
|
|
|
|
|
|
|
where C<$a_ext> is either the periodic (or reflected) extension of |
587
|
|
|
|
|
|
|
C<$a> so it is equal to C<$a> on C< 0..$M-1 > and equal to the |
588
|
|
|
|
|
|
|
corresponding periodic/reflected image of C<$a> outside that range. |
589
|
|
|
|
|
|
|
|
590
|
|
|
|
|
|
|
|
591
|
|
|
|
|
|
|
=for example |
592
|
|
|
|
|
|
|
|
593
|
|
|
|
|
|
|
$con = conv1d sequence(10), pdl(-1,0,1); |
594
|
|
|
|
|
|
|
|
595
|
|
|
|
|
|
|
$con = conv1d sequence(10), pdl(-1,0,1), {Boundary => 'reflect'}; |
596
|
|
|
|
|
|
|
|
597
|
|
|
|
|
|
|
By default, periodic boundary conditions are assumed (i.e. wrap around). |
598
|
|
|
|
|
|
|
Alternatively, you can request reflective boundary conditions using |
599
|
|
|
|
|
|
|
the C option: |
600
|
|
|
|
|
|
|
|
601
|
|
|
|
|
|
|
{Boundary => 'reflect'} # case in 'reflect' doesn't matter |
602
|
|
|
|
|
|
|
|
603
|
|
|
|
|
|
|
The convolution is performed along the first dimension. To apply it across |
604
|
|
|
|
|
|
|
another dimension use the slicing routines, e.g. |
605
|
|
|
|
|
|
|
|
606
|
|
|
|
|
|
|
$y = $x->mv(2,0)->conv1d($kernel)->mv(0,2); # along third dim |
607
|
|
|
|
|
|
|
|
608
|
|
|
|
|
|
|
This function is useful for threaded filtering of 1D signals. |
609
|
|
|
|
|
|
|
|
610
|
|
|
|
|
|
|
Compare also L, L, |
611
|
|
|
|
|
|
|
L, L, |
612
|
|
|
|
|
|
|
L |
613
|
|
|
|
|
|
|
|
614
|
|
|
|
|
|
|
=for bad |
615
|
|
|
|
|
|
|
|
616
|
|
|
|
|
|
|
WARNING: C processes bad values in its inputs as |
617
|
|
|
|
|
|
|
the numeric value of C<< $pdl->badvalue >> so it is not |
618
|
|
|
|
|
|
|
recommended for processing pdls with bad values in them |
619
|
|
|
|
|
|
|
unless special care is taken. |
620
|
|
|
|
|
|
|
|
621
|
|
|
|
|
|
|
|
622
|
|
|
|
|
|
|
|
623
|
|
|
|
|
|
|
=for bad |
624
|
|
|
|
|
|
|
|
625
|
|
|
|
|
|
|
conv1d ignores the bad-value flag of the input piddles. |
626
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
627
|
|
|
|
|
|
|
|
628
|
|
|
|
|
|
|
|
629
|
|
|
|
|
|
|
=cut |
630
|
|
|
|
|
|
|
|
631
|
|
|
|
|
|
|
|
632
|
|
|
|
|
|
|
|
633
|
|
|
|
|
|
|
|
634
|
|
|
|
|
|
|
|
635
|
|
|
|
|
|
|
|
636
|
|
|
|
|
|
|
sub PDLA::conv1d { |
637
|
0
|
0
|
|
0
|
0
|
0
|
my $opt = pop @_ if ref($_[$#_]) eq 'HASH'; |
638
|
0
|
0
|
0
|
|
|
0
|
die 'Usage: conv1d( a(m), kern(p), [o]b(m), {Options} )' |
639
|
|
|
|
|
|
|
if $#_<1 || $#_>2; |
640
|
0
|
|
|
|
|
0
|
my($x,$kern) = @_; |
641
|
0
|
0
|
|
|
|
0
|
my $c = $#_ == 2 ? $_[2] : PDLA->null; |
642
|
|
|
|
|
|
|
&PDLA::_conv1d_int($x,$kern,$c, |
643
|
|
|
|
|
|
|
!(defined $opt && exists $$opt{Boundary}) ? 0 : |
644
|
0
|
0
|
0
|
|
|
0
|
lc $$opt{Boundary} eq "reflect"); |
645
|
0
|
|
|
|
|
0
|
return $c; |
646
|
|
|
|
|
|
|
} |
647
|
|
|
|
|
|
|
|
648
|
|
|
|
|
|
|
|
649
|
|
|
|
|
|
|
|
650
|
|
|
|
|
|
|
*conv1d = \&PDLA::conv1d; |
651
|
|
|
|
|
|
|
|
652
|
|
|
|
|
|
|
|
653
|
|
|
|
|
|
|
|
654
|
|
|
|
|
|
|
|
655
|
|
|
|
|
|
|
|
656
|
|
|
|
|
|
|
=head2 in |
657
|
|
|
|
|
|
|
|
658
|
|
|
|
|
|
|
=for sig |
659
|
|
|
|
|
|
|
|
660
|
|
|
|
|
|
|
Signature: (a(); b(n); [o] c()) |
661
|
|
|
|
|
|
|
|
662
|
|
|
|
|
|
|
|
663
|
|
|
|
|
|
|
=for ref |
664
|
|
|
|
|
|
|
|
665
|
|
|
|
|
|
|
test if a is in the set of values b |
666
|
|
|
|
|
|
|
|
667
|
|
|
|
|
|
|
=for example |
668
|
|
|
|
|
|
|
|
669
|
|
|
|
|
|
|
$goodmsk = $labels->in($goodlabels); |
670
|
|
|
|
|
|
|
print pdl(3,1,4,6,2)->in(pdl(2,3,3)); |
671
|
|
|
|
|
|
|
[1 0 0 0 1] |
672
|
|
|
|
|
|
|
|
673
|
|
|
|
|
|
|
C is akin to the I of set theory. In principle, |
674
|
|
|
|
|
|
|
PDLA threading could be used to achieve its functionality by using a |
675
|
|
|
|
|
|
|
construct like |
676
|
|
|
|
|
|
|
|
677
|
|
|
|
|
|
|
$msk = ($labels->dummy(0) == $goodlabels)->orover; |
678
|
|
|
|
|
|
|
|
679
|
|
|
|
|
|
|
However, C doesn't create a (potentially large) intermediate |
680
|
|
|
|
|
|
|
and is generally faster. |
681
|
|
|
|
|
|
|
|
682
|
|
|
|
|
|
|
|
683
|
|
|
|
|
|
|
|
684
|
|
|
|
|
|
|
=for bad |
685
|
|
|
|
|
|
|
|
686
|
|
|
|
|
|
|
in does not process bad values. |
687
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
688
|
|
|
|
|
|
|
|
689
|
|
|
|
|
|
|
|
690
|
|
|
|
|
|
|
=cut |
691
|
|
|
|
|
|
|
|
692
|
|
|
|
|
|
|
|
693
|
|
|
|
|
|
|
|
694
|
|
|
|
|
|
|
|
695
|
|
|
|
|
|
|
|
696
|
|
|
|
|
|
|
|
697
|
|
|
|
|
|
|
*in = \&PDLA::in; |
698
|
|
|
|
|
|
|
|
699
|
|
|
|
|
|
|
|
700
|
|
|
|
|
|
|
|
701
|
|
|
|
|
|
|
|
702
|
|
|
|
|
|
|
=head2 uniq |
703
|
|
|
|
|
|
|
|
704
|
|
|
|
|
|
|
=for ref |
705
|
|
|
|
|
|
|
|
706
|
|
|
|
|
|
|
return all unique elements of a piddle |
707
|
|
|
|
|
|
|
|
708
|
|
|
|
|
|
|
The unique elements are returned in ascending order. |
709
|
|
|
|
|
|
|
|
710
|
|
|
|
|
|
|
=for example |
711
|
|
|
|
|
|
|
|
712
|
|
|
|
|
|
|
PDLA> p pdl(2,2,2,4,0,-1,6,6)->uniq |
713
|
|
|
|
|
|
|
[-1 0 2 4 6] # 0 is returned 2nd (sorted order) |
714
|
|
|
|
|
|
|
|
715
|
|
|
|
|
|
|
PDLA> p pdl(2,2,2,4,nan,-1,6,6)->uniq |
716
|
|
|
|
|
|
|
[-1 2 4 6 nan] # NaN value is returned at end |
717
|
|
|
|
|
|
|
|
718
|
|
|
|
|
|
|
Note: The returned pdl is 1D; any structure of the input |
719
|
|
|
|
|
|
|
piddle is lost. C values are never compare equal to |
720
|
|
|
|
|
|
|
any other values, even themselves. As a result, they are |
721
|
|
|
|
|
|
|
always unique. C returns the NaN values at the end |
722
|
|
|
|
|
|
|
of the result piddle. This follows the Matlab usage. |
723
|
|
|
|
|
|
|
|
724
|
|
|
|
|
|
|
See L if you need the indices of the unique |
725
|
|
|
|
|
|
|
elements rather than the values. |
726
|
|
|
|
|
|
|
|
727
|
|
|
|
|
|
|
=cut |
728
|
|
|
|
|
|
|
|
729
|
|
|
|
|
|
|
|
730
|
|
|
|
|
|
|
|
731
|
|
|
|
|
|
|
|
732
|
|
|
|
|
|
|
=for bad |
733
|
|
|
|
|
|
|
|
734
|
|
|
|
|
|
|
Bad values are not considered unique by uniq and are ignored. |
735
|
|
|
|
|
|
|
|
736
|
|
|
|
|
|
|
$x=sequence(10); |
737
|
|
|
|
|
|
|
$x=$x->setbadif($x%3); |
738
|
|
|
|
|
|
|
print $x->uniq; |
739
|
|
|
|
|
|
|
[0 3 6 9] |
740
|
|
|
|
|
|
|
|
741
|
|
|
|
|
|
|
=cut |
742
|
|
|
|
|
|
|
|
743
|
|
|
|
|
|
|
|
744
|
|
|
|
|
|
|
|
745
|
|
|
|
|
|
|
|
746
|
|
|
|
|
|
|
*uniq = \&PDLA::uniq; |
747
|
|
|
|
|
|
|
# return unique elements of array |
748
|
|
|
|
|
|
|
# find as jumps in the sorted array |
749
|
|
|
|
|
|
|
# flattens in the process |
750
|
|
|
|
|
|
|
sub PDLA::uniq { |
751
|
78
|
|
|
78
|
|
650
|
use PDLA::Core 'barf'; |
|
78
|
|
|
|
|
196
|
|
|
78
|
|
|
|
|
413
|
|
752
|
11
|
|
|
11
|
0
|
17
|
my ($arr) = @_; |
753
|
11
|
50
|
|
|
|
29
|
return $arr if($arr->nelem == 0); # The null list is unique (CED) |
754
|
11
|
|
|
|
|
25
|
my $srt = $arr->clump(-1)->where($arr==$arr)->qsort; # no NaNs or BADs for qsort |
755
|
11
|
|
|
|
|
85
|
my $nans = $arr->clump(-1)->where($arr!=$arr); |
756
|
11
|
50
|
|
|
|
236
|
my $uniq = ($srt->nelem > 0) ? $srt->where($srt != $srt->rotate(-1)) : $srt; |
757
|
|
|
|
|
|
|
# make sure we return something if there is only one value |
758
|
11
|
|
|
|
|
63
|
my $answ = $nans; # NaN values always uniq |
759
|
11
|
50
|
|
|
|
56
|
if ( $uniq->nelem > 0 ) { |
760
|
11
|
|
|
|
|
132
|
$answ = $uniq->append($answ); |
761
|
|
|
|
|
|
|
} else { |
762
|
0
|
0
|
|
|
|
0
|
$answ = ( ($srt->nelem == 0) ? $srt : PDLA::pdl( ref($srt), [$srt->index(0)] ) )->append($answ); |
763
|
|
|
|
|
|
|
} |
764
|
11
|
|
|
|
|
89
|
return $answ; |
765
|
|
|
|
|
|
|
} |
766
|
|
|
|
|
|
|
|
767
|
|
|
|
|
|
|
|
768
|
|
|
|
|
|
|
|
769
|
|
|
|
|
|
|
|
770
|
|
|
|
|
|
|
=head2 uniqind |
771
|
|
|
|
|
|
|
|
772
|
|
|
|
|
|
|
=for ref |
773
|
|
|
|
|
|
|
|
774
|
|
|
|
|
|
|
Return the indices of all unique elements of a piddle |
775
|
|
|
|
|
|
|
The order is in the order of the values to be consistent |
776
|
|
|
|
|
|
|
with uniq. C values never compare equal with any |
777
|
|
|
|
|
|
|
other value and so are always unique. This follows the |
778
|
|
|
|
|
|
|
Matlab usage. |
779
|
|
|
|
|
|
|
|
780
|
|
|
|
|
|
|
=for example |
781
|
|
|
|
|
|
|
|
782
|
|
|
|
|
|
|
PDLA> p pdl(2,2,2,4,0,-1,6,6)->uniqind |
783
|
|
|
|
|
|
|
[5 4 1 3 6] # the 0 at index 4 is returned 2nd, but... |
784
|
|
|
|
|
|
|
|
785
|
|
|
|
|
|
|
PDLA> p pdl(2,2,2,4,nan,-1,6,6)->uniqind |
786
|
|
|
|
|
|
|
[5 1 3 6 4] # ...the NaN at index 4 is returned at end |
787
|
|
|
|
|
|
|
|
788
|
|
|
|
|
|
|
|
789
|
|
|
|
|
|
|
Note: The returned pdl is 1D; any structure of the input |
790
|
|
|
|
|
|
|
piddle is lost. |
791
|
|
|
|
|
|
|
|
792
|
|
|
|
|
|
|
See L if you want the unique values instead of the |
793
|
|
|
|
|
|
|
indices. |
794
|
|
|
|
|
|
|
|
795
|
|
|
|
|
|
|
=cut |
796
|
|
|
|
|
|
|
|
797
|
|
|
|
|
|
|
|
798
|
|
|
|
|
|
|
|
799
|
|
|
|
|
|
|
|
800
|
|
|
|
|
|
|
=for bad |
801
|
|
|
|
|
|
|
|
802
|
|
|
|
|
|
|
Bad values are not considered unique by uniqind and are ignored. |
803
|
|
|
|
|
|
|
|
804
|
|
|
|
|
|
|
=cut |
805
|
|
|
|
|
|
|
|
806
|
|
|
|
|
|
|
|
807
|
|
|
|
|
|
|
|
808
|
|
|
|
|
|
|
|
809
|
|
|
|
|
|
|
*uniqind = \&PDLA::uniqind; |
810
|
|
|
|
|
|
|
# return unique elements of array |
811
|
|
|
|
|
|
|
# find as jumps in the sorted array |
812
|
|
|
|
|
|
|
# flattens in the process |
813
|
|
|
|
|
|
|
sub PDLA::uniqind { |
814
|
78
|
|
|
78
|
|
612
|
use PDLA::Core 'barf'; |
|
78
|
|
|
|
|
186
|
|
|
78
|
|
|
|
|
352
|
|
815
|
2
|
|
|
2
|
0
|
10
|
my ($arr) = @_; |
816
|
2
|
50
|
|
|
|
10
|
return $arr if($arr->nelem == 0); # The null list is unique (CED) |
817
|
|
|
|
|
|
|
# Different from uniq we sort and store the result in an intermediary |
818
|
2
|
|
|
|
|
5
|
my $aflat = $arr->flat; |
819
|
2
|
|
|
|
|
40
|
my $nanind = which($aflat!=$aflat); # NaN indexes |
820
|
2
|
|
|
|
|
12
|
my $good = $aflat->sequence->long->where($aflat==$aflat); # good indexes |
821
|
2
|
|
|
|
|
27
|
my $i_srt = $aflat->where($aflat==$aflat)->qsorti; # no BAD or NaN values for qsorti |
822
|
2
|
|
|
|
|
30
|
my $srt = $aflat->where($aflat==$aflat)->index($i_srt); |
823
|
2
|
|
|
|
|
8
|
my $uniqind; |
824
|
2
|
50
|
|
|
|
14
|
if ($srt->nelem > 0) { |
825
|
2
|
|
|
|
|
39
|
$uniqind = which($srt != $srt->rotate(-1)); |
826
|
2
|
100
|
|
|
|
19
|
$uniqind = $i_srt->slice('0') if $uniqind->isempty; |
827
|
|
|
|
|
|
|
} else { |
828
|
0
|
|
|
|
|
0
|
$uniqind = which($srt); |
829
|
|
|
|
|
|
|
} |
830
|
|
|
|
|
|
|
# Now map back to the original space |
831
|
2
|
|
|
|
|
4
|
my $ansind = $nanind; |
832
|
2
|
50
|
|
|
|
10
|
if ( $uniqind->nelem > 0 ) { |
833
|
2
|
|
|
|
|
51
|
$ansind = ($good->index($i_srt->index($uniqind)))->append($ansind); |
834
|
|
|
|
|
|
|
} else { |
835
|
0
|
|
|
|
|
0
|
$ansind = $uniqind->append($ansind); |
836
|
|
|
|
|
|
|
} |
837
|
2
|
|
|
|
|
43
|
return $ansind; |
838
|
|
|
|
|
|
|
} |
839
|
|
|
|
|
|
|
|
840
|
|
|
|
|
|
|
|
841
|
|
|
|
|
|
|
|
842
|
|
|
|
|
|
|
|
843
|
|
|
|
|
|
|
=head2 uniqvec |
844
|
|
|
|
|
|
|
|
845
|
|
|
|
|
|
|
=for ref |
846
|
|
|
|
|
|
|
|
847
|
|
|
|
|
|
|
Return all unique vectors out of a collection |
848
|
|
|
|
|
|
|
|
849
|
|
|
|
|
|
|
NOTE: If any vectors in the input piddle have NaN values |
850
|
|
|
|
|
|
|
they are returned at the end of the non-NaN ones. This is |
851
|
|
|
|
|
|
|
because, by definition, NaN values never compare equal with |
852
|
|
|
|
|
|
|
any other value. |
853
|
|
|
|
|
|
|
|
854
|
|
|
|
|
|
|
NOTE: The current implementation does not sort the vectors |
855
|
|
|
|
|
|
|
containing NaN values. |
856
|
|
|
|
|
|
|
|
857
|
|
|
|
|
|
|
The unique vectors are returned in lexicographically sorted |
858
|
|
|
|
|
|
|
ascending order. The 0th dimension of the input PDLA is treated |
859
|
|
|
|
|
|
|
as a dimensional index within each vector, and the 1st and any |
860
|
|
|
|
|
|
|
higher dimensions are taken to run across vectors. The return |
861
|
|
|
|
|
|
|
value is always 2D; any structure of the input PDLA (beyond using |
862
|
|
|
|
|
|
|
the 0th dimension for vector index) is lost. |
863
|
|
|
|
|
|
|
|
864
|
|
|
|
|
|
|
See also L for a unique list of scalars; and |
865
|
|
|
|
|
|
|
L for sorting a list of vectors |
866
|
|
|
|
|
|
|
lexicographcally. |
867
|
|
|
|
|
|
|
|
868
|
|
|
|
|
|
|
=cut |
869
|
|
|
|
|
|
|
|
870
|
|
|
|
|
|
|
|
871
|
|
|
|
|
|
|
|
872
|
|
|
|
|
|
|
|
873
|
|
|
|
|
|
|
=for bad |
874
|
|
|
|
|
|
|
|
875
|
|
|
|
|
|
|
If a vector contains all bad values, it is ignored as in L. |
876
|
|
|
|
|
|
|
If some of the values are good, it is treated as a normal vector. For |
877
|
|
|
|
|
|
|
example, [1 2 BAD] and [BAD 2 3] could be returned, but [BAD BAD BAD] |
878
|
|
|
|
|
|
|
could not. Vectors containing BAD values will be returned after any |
879
|
|
|
|
|
|
|
non-NaN and non-BAD containing vectors, followed by the NaN vectors. |
880
|
|
|
|
|
|
|
|
881
|
|
|
|
|
|
|
|
882
|
|
|
|
|
|
|
=cut |
883
|
|
|
|
|
|
|
|
884
|
|
|
|
|
|
|
|
885
|
|
|
|
|
|
|
|
886
|
|
|
|
|
|
|
|
887
|
|
|
|
|
|
|
sub PDLA::uniqvec { |
888
|
|
|
|
|
|
|
|
889
|
9
|
|
|
9
|
0
|
1804
|
my($pdl) = shift; |
890
|
|
|
|
|
|
|
|
891
|
9
|
50
|
33
|
|
|
88
|
return $pdl if ( $pdl->nelem == 0 || $pdl->ndims < 2 ); |
892
|
9
|
100
|
|
|
|
36
|
return $pdl if ( $pdl->slice("(0)")->nelem < 2 ); # slice isn't cheap but uniqvec isn't either |
893
|
|
|
|
|
|
|
|
894
|
8
|
|
|
|
|
51
|
my $pdl2d = null; |
895
|
8
|
|
|
|
|
65
|
$pdl2d = $pdl->mv(0,-1)->clump($pdl->ndims-1)->mv(-1,0); # clump all but dim(0) |
896
|
|
|
|
|
|
|
|
897
|
8
|
|
|
|
|
35
|
my $ngood = null; |
898
|
8
|
|
|
|
|
32
|
$ngood = $pdl2d->ones->sumover; |
899
|
8
|
50
|
33
|
|
|
119
|
$ngood = $pdl2d->ngoodover if ($PDLA::Bad::Status && $pdl->badflag); # number of good values each vector |
900
|
8
|
|
|
|
|
35
|
my $ngood2 = null; |
901
|
8
|
|
|
|
|
35
|
$ngood2 = $ngood->where($ngood); # number of good values with no all-BADs |
902
|
|
|
|
|
|
|
|
903
|
8
|
|
|
|
|
60
|
$pdl2d = $pdl2d->mv(0,-1)->dice($ngood->which)->mv(-1,0); # remove all-BAD vectors |
904
|
|
|
|
|
|
|
|
905
|
|
|
|
|
|
|
|
906
|
8
|
|
|
|
|
39
|
my $numnan = null; |
907
|
8
|
|
|
|
|
12714
|
$numnan = ($pdl2d!=$pdl2d)->sumover; # works since no all-BADs to confuse |
908
|
|
|
|
|
|
|
|
909
|
8
|
|
|
|
|
62
|
my $presrt = null; |
910
|
8
|
|
|
|
|
292
|
$presrt = $pdl2d->mv(0,-1)->dice($numnan->not->which)->mv(0,-1); # remove vectors with any NaN values |
911
|
8
|
|
|
|
|
45
|
my $nanvec = null; |
912
|
8
|
|
|
|
|
85
|
$nanvec = $pdl2d->mv(0,-1)->dice($numnan->which)->mv(0,-1); # the vectors with any NaN values |
913
|
|
|
|
|
|
|
|
914
|
|
|
|
|
|
|
# use dice instead of nslice since qsortvec might be packing |
915
|
|
|
|
|
|
|
# the badvals to the front of the array instead of the end like |
916
|
|
|
|
|
|
|
# the docs say. If that is the case and it gets fixed, it won't |
917
|
|
|
|
|
|
|
# bust uniqvec. DAL 14-March 2006 |
918
|
|
|
|
|
|
|
|
919
|
8
|
|
|
|
|
41
|
my $srt = null; |
920
|
8
|
|
|
|
|
50938
|
$srt = $presrt->qsortvec->mv(0,-1); # BADs are sorted by qsortvec |
921
|
8
|
|
|
|
|
49
|
my $srtdice = $srt; |
922
|
8
|
|
|
|
|
37
|
my $somebad = null; |
923
|
8
|
50
|
33
|
|
|
86
|
if ($PDLA::Bad::Status && $srt->badflag) { |
924
|
0
|
|
|
|
|
0
|
$srtdice = $srt->dice($srt->mv(0,-1)->nbadover->not->which); |
925
|
0
|
|
|
|
|
0
|
$somebad = $srt->dice($srt->mv(0,-1)->nbadover->which); |
926
|
|
|
|
|
|
|
} |
927
|
|
|
|
|
|
|
|
928
|
8
|
|
|
|
|
21
|
my $uniq = null; |
929
|
8
|
50
|
|
|
|
48
|
if ($srtdice->nelem > 0) { |
930
|
8
|
|
|
|
|
18273
|
$uniq = ($srtdice != $srtdice->rotate(-1))->mv(0,-1)->orover->which; |
931
|
|
|
|
|
|
|
} else { |
932
|
0
|
|
|
|
|
0
|
$uniq = $srtdice->orover->which; |
933
|
|
|
|
|
|
|
} |
934
|
|
|
|
|
|
|
|
935
|
8
|
|
|
|
|
595
|
my $ans = null; |
936
|
8
|
100
|
|
|
|
40
|
if ( $uniq->nelem > 0 ) { |
937
|
1
|
|
|
|
|
4
|
$ans = $srtdice->dice($uniq); |
938
|
|
|
|
|
|
|
} else { |
939
|
7
|
50
|
|
|
|
49
|
$ans = ($srtdice->nelem > 0) ? $srtdice->slice("0,:") : $srtdice; |
940
|
|
|
|
|
|
|
} |
941
|
8
|
|
|
|
|
1217
|
return $ans->append($somebad)->append($nanvec->mv(0,-1))->mv(0,-1); |
942
|
|
|
|
|
|
|
} |
943
|
|
|
|
|
|
|
|
944
|
|
|
|
|
|
|
|
945
|
|
|
|
|
|
|
|
946
|
|
|
|
|
|
|
|
947
|
|
|
|
|
|
|
|
948
|
|
|
|
|
|
|
=head2 hclip |
949
|
|
|
|
|
|
|
|
950
|
|
|
|
|
|
|
=for sig |
951
|
|
|
|
|
|
|
|
952
|
|
|
|
|
|
|
Signature: (a(); b(); [o] c()) |
953
|
|
|
|
|
|
|
|
954
|
|
|
|
|
|
|
=for ref |
955
|
|
|
|
|
|
|
|
956
|
|
|
|
|
|
|
clip (threshold) C<$a> by C<$b> (C<$b> is upper bound) |
957
|
|
|
|
|
|
|
|
958
|
|
|
|
|
|
|
=for bad |
959
|
|
|
|
|
|
|
|
960
|
|
|
|
|
|
|
hclip processes bad values. |
961
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
962
|
|
|
|
|
|
|
|
963
|
|
|
|
|
|
|
|
964
|
|
|
|
|
|
|
=cut |
965
|
|
|
|
|
|
|
|
966
|
|
|
|
|
|
|
|
967
|
|
|
|
|
|
|
|
968
|
|
|
|
|
|
|
|
969
|
|
|
|
|
|
|
sub PDLA::hclip { |
970
|
2
|
|
|
2
|
0
|
7
|
my ($x,$y) = @_; |
971
|
2
|
|
|
|
|
5
|
my $c; |
972
|
2
|
50
|
|
|
|
16
|
if ($x->is_inplace) { |
|
|
50
|
|
|
|
|
|
973
|
0
|
|
|
|
|
0
|
$x->set_inplace(0); $c = $x; |
|
0
|
|
|
|
|
0
|
|
974
|
0
|
|
|
|
|
0
|
} elsif ($#_ > 1) {$c=$_[2]} else {$c=PDLA->nullcreate($x)} |
|
2
|
|
|
|
|
8
|
|
975
|
2
|
|
|
|
|
107
|
&PDLA::_hclip_int($x,$y,$c); |
976
|
2
|
|
|
|
|
25
|
return $c; |
977
|
|
|
|
|
|
|
} |
978
|
|
|
|
|
|
|
|
979
|
|
|
|
|
|
|
|
980
|
|
|
|
|
|
|
*hclip = \&PDLA::hclip; |
981
|
|
|
|
|
|
|
|
982
|
|
|
|
|
|
|
|
983
|
|
|
|
|
|
|
|
984
|
|
|
|
|
|
|
|
985
|
|
|
|
|
|
|
|
986
|
|
|
|
|
|
|
=head2 lclip |
987
|
|
|
|
|
|
|
|
988
|
|
|
|
|
|
|
=for sig |
989
|
|
|
|
|
|
|
|
990
|
|
|
|
|
|
|
Signature: (a(); b(); [o] c()) |
991
|
|
|
|
|
|
|
|
992
|
|
|
|
|
|
|
=for ref |
993
|
|
|
|
|
|
|
|
994
|
|
|
|
|
|
|
clip (threshold) C<$a> by C<$b> (C<$b> is lower bound) |
995
|
|
|
|
|
|
|
|
996
|
|
|
|
|
|
|
=for bad |
997
|
|
|
|
|
|
|
|
998
|
|
|
|
|
|
|
lclip processes bad values. |
999
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
1000
|
|
|
|
|
|
|
|
1001
|
|
|
|
|
|
|
|
1002
|
|
|
|
|
|
|
=cut |
1003
|
|
|
|
|
|
|
|
1004
|
|
|
|
|
|
|
|
1005
|
|
|
|
|
|
|
|
1006
|
|
|
|
|
|
|
|
1007
|
|
|
|
|
|
|
sub PDLA::lclip { |
1008
|
2
|
|
|
2
|
0
|
349
|
my ($x,$y) = @_; |
1009
|
2
|
|
|
|
|
4
|
my $c; |
1010
|
2
|
50
|
|
|
|
15
|
if ($x->is_inplace) { |
|
|
50
|
|
|
|
|
|
1011
|
0
|
|
|
|
|
0
|
$x->set_inplace(0); $c = $x; |
|
0
|
|
|
|
|
0
|
|
1012
|
0
|
|
|
|
|
0
|
} elsif ($#_ > 1) {$c=$_[2]} else {$c=PDLA->nullcreate($x)} |
|
2
|
|
|
|
|
7
|
|
1013
|
2
|
|
|
|
|
100
|
&PDLA::_lclip_int($x,$y,$c); |
1014
|
2
|
|
|
|
|
22
|
return $c; |
1015
|
|
|
|
|
|
|
} |
1016
|
|
|
|
|
|
|
|
1017
|
|
|
|
|
|
|
|
1018
|
|
|
|
|
|
|
*lclip = \&PDLA::lclip; |
1019
|
|
|
|
|
|
|
|
1020
|
|
|
|
|
|
|
|
1021
|
|
|
|
|
|
|
|
1022
|
|
|
|
|
|
|
|
1023
|
|
|
|
|
|
|
=head2 clip |
1024
|
|
|
|
|
|
|
|
1025
|
|
|
|
|
|
|
=for ref |
1026
|
|
|
|
|
|
|
|
1027
|
|
|
|
|
|
|
Clip (threshold) a piddle by (optional) upper or lower bounds. |
1028
|
|
|
|
|
|
|
|
1029
|
|
|
|
|
|
|
=for usage |
1030
|
|
|
|
|
|
|
|
1031
|
|
|
|
|
|
|
$y = $x->clip(0,3); |
1032
|
|
|
|
|
|
|
$c = $x->clip(undef, $x); |
1033
|
|
|
|
|
|
|
|
1034
|
|
|
|
|
|
|
=cut |
1035
|
|
|
|
|
|
|
|
1036
|
|
|
|
|
|
|
|
1037
|
|
|
|
|
|
|
|
1038
|
|
|
|
|
|
|
|
1039
|
|
|
|
|
|
|
=for bad |
1040
|
|
|
|
|
|
|
|
1041
|
|
|
|
|
|
|
clip handles bad values since it is just a |
1042
|
|
|
|
|
|
|
wrapper around L and |
1043
|
|
|
|
|
|
|
L. |
1044
|
|
|
|
|
|
|
|
1045
|
|
|
|
|
|
|
=cut |
1046
|
|
|
|
|
|
|
|
1047
|
|
|
|
|
|
|
|
1048
|
|
|
|
|
|
|
|
1049
|
|
|
|
|
|
|
|
1050
|
|
|
|
|
|
|
|
1051
|
|
|
|
|
|
|
=head2 clip |
1052
|
|
|
|
|
|
|
|
1053
|
|
|
|
|
|
|
=for sig |
1054
|
|
|
|
|
|
|
|
1055
|
|
|
|
|
|
|
Signature: (a(); l(); h(); [o] c()) |
1056
|
|
|
|
|
|
|
|
1057
|
|
|
|
|
|
|
|
1058
|
|
|
|
|
|
|
=for ref |
1059
|
|
|
|
|
|
|
|
1060
|
|
|
|
|
|
|
info not available |
1061
|
|
|
|
|
|
|
|
1062
|
|
|
|
|
|
|
|
1063
|
|
|
|
|
|
|
=for bad |
1064
|
|
|
|
|
|
|
|
1065
|
|
|
|
|
|
|
clip processes bad values. |
1066
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
1067
|
|
|
|
|
|
|
|
1068
|
|
|
|
|
|
|
|
1069
|
|
|
|
|
|
|
=cut |
1070
|
|
|
|
|
|
|
|
1071
|
|
|
|
|
|
|
|
1072
|
|
|
|
|
|
|
|
1073
|
|
|
|
|
|
|
|
1074
|
|
|
|
|
|
|
*clip = \&PDLA::clip; |
1075
|
|
|
|
|
|
|
sub PDLA::clip { |
1076
|
2
|
|
|
2
|
0
|
671
|
my($x, $l, $h) = @_; |
1077
|
2
|
|
|
|
|
4
|
my $d; |
1078
|
2
|
50
|
33
|
|
|
10
|
unless(defined($l) || defined($h)) { |
1079
|
|
|
|
|
|
|
# Deal with pathological case |
1080
|
0
|
0
|
|
|
|
0
|
if($x->is_inplace) { |
1081
|
0
|
|
|
|
|
0
|
$x->set_inplace(0); |
1082
|
0
|
|
|
|
|
0
|
return $x; |
1083
|
|
|
|
|
|
|
} else { |
1084
|
0
|
|
|
|
|
0
|
return $x->copy; |
1085
|
|
|
|
|
|
|
} |
1086
|
|
|
|
|
|
|
} |
1087
|
|
|
|
|
|
|
|
1088
|
2
|
50
|
|
|
|
15
|
if($x->is_inplace) { |
|
|
50
|
|
|
|
|
|
1089
|
0
|
|
|
|
|
0
|
$x->set_inplace(0); $d = $x |
|
0
|
|
|
|
|
0
|
|
1090
|
|
|
|
|
|
|
} elsif ($#_ > 2) { |
1091
|
0
|
|
|
|
|
0
|
$d=$_[3] |
1092
|
|
|
|
|
|
|
} else { |
1093
|
2
|
|
|
|
|
9
|
$d = PDLA->nullcreate($x); |
1094
|
|
|
|
|
|
|
} |
1095
|
2
|
50
|
33
|
|
|
17
|
if(defined($l) && defined($h)) { |
|
|
0
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
1096
|
2
|
|
|
|
|
112
|
&PDLA::_clip_int($x,$l,$h,$d); |
1097
|
|
|
|
|
|
|
} elsif( defined($l) ) { |
1098
|
0
|
|
|
|
|
0
|
&PDLA::_lclip_int($x,$l,$d); |
1099
|
|
|
|
|
|
|
} elsif( defined($h) ) { |
1100
|
0
|
|
|
|
|
0
|
&PDLA::_hclip_int($x,$h,$d); |
1101
|
|
|
|
|
|
|
} else { |
1102
|
0
|
|
|
|
|
0
|
die "This can't happen (clip contingency) - file a bug"; |
1103
|
|
|
|
|
|
|
} |
1104
|
|
|
|
|
|
|
|
1105
|
2
|
|
|
|
|
27
|
return $d; |
1106
|
|
|
|
|
|
|
} |
1107
|
|
|
|
|
|
|
|
1108
|
|
|
|
|
|
|
|
1109
|
|
|
|
|
|
|
*clip = \&PDLA::clip; |
1110
|
|
|
|
|
|
|
|
1111
|
|
|
|
|
|
|
|
1112
|
|
|
|
|
|
|
|
1113
|
|
|
|
|
|
|
|
1114
|
|
|
|
|
|
|
|
1115
|
|
|
|
|
|
|
=head2 wtstat |
1116
|
|
|
|
|
|
|
|
1117
|
|
|
|
|
|
|
=for sig |
1118
|
|
|
|
|
|
|
|
1119
|
|
|
|
|
|
|
Signature: (a(n); wt(n); avg(); [o]b(); int deg) |
1120
|
|
|
|
|
|
|
|
1121
|
|
|
|
|
|
|
|
1122
|
|
|
|
|
|
|
|
1123
|
|
|
|
|
|
|
=for ref |
1124
|
|
|
|
|
|
|
|
1125
|
|
|
|
|
|
|
Weighted statistical moment of given degree |
1126
|
|
|
|
|
|
|
|
1127
|
|
|
|
|
|
|
This calculates a weighted statistic over the vector C. |
1128
|
|
|
|
|
|
|
The formula is |
1129
|
|
|
|
|
|
|
|
1130
|
|
|
|
|
|
|
b() = (sum_i wt_i * (a_i ** degree - avg)) / (sum_i wt_i) |
1131
|
|
|
|
|
|
|
|
1132
|
|
|
|
|
|
|
|
1133
|
|
|
|
|
|
|
|
1134
|
|
|
|
|
|
|
=for bad |
1135
|
|
|
|
|
|
|
|
1136
|
|
|
|
|
|
|
=for bad |
1137
|
|
|
|
|
|
|
|
1138
|
|
|
|
|
|
|
Bad values are ignored in any calculation; C<$b> will only |
1139
|
|
|
|
|
|
|
have its bad flag set if the output contains any bad data. |
1140
|
|
|
|
|
|
|
|
1141
|
|
|
|
|
|
|
|
1142
|
|
|
|
|
|
|
|
1143
|
|
|
|
|
|
|
=cut |
1144
|
|
|
|
|
|
|
|
1145
|
|
|
|
|
|
|
|
1146
|
|
|
|
|
|
|
|
1147
|
|
|
|
|
|
|
|
1148
|
|
|
|
|
|
|
|
1149
|
|
|
|
|
|
|
|
1150
|
|
|
|
|
|
|
*wtstat = \&PDLA::wtstat; |
1151
|
|
|
|
|
|
|
|
1152
|
|
|
|
|
|
|
|
1153
|
|
|
|
|
|
|
|
1154
|
|
|
|
|
|
|
|
1155
|
|
|
|
|
|
|
|
1156
|
|
|
|
|
|
|
=head2 statsover |
1157
|
|
|
|
|
|
|
|
1158
|
|
|
|
|
|
|
=for sig |
1159
|
|
|
|
|
|
|
|
1160
|
|
|
|
|
|
|
Signature: (a(n); w(n); float+ [o]avg(); float+ [o]prms(); int+ [o]median(); int+ [o]min(); int+ [o]max(); float+ [o]adev(); float+ [o]rms()) |
1161
|
|
|
|
|
|
|
|
1162
|
|
|
|
|
|
|
|
1163
|
|
|
|
|
|
|
|
1164
|
|
|
|
|
|
|
=for ref |
1165
|
|
|
|
|
|
|
|
1166
|
|
|
|
|
|
|
Calculate useful statistics over a dimension of a piddle |
1167
|
|
|
|
|
|
|
|
1168
|
|
|
|
|
|
|
=for usage |
1169
|
|
|
|
|
|
|
|
1170
|
|
|
|
|
|
|
($mean,$prms,$median,$min,$max,$adev,$rms) = statsover($piddle, $weights); |
1171
|
|
|
|
|
|
|
|
1172
|
|
|
|
|
|
|
This utility function calculates various useful |
1173
|
|
|
|
|
|
|
quantities of a piddle. These are: |
1174
|
|
|
|
|
|
|
|
1175
|
|
|
|
|
|
|
=over 3 |
1176
|
|
|
|
|
|
|
|
1177
|
|
|
|
|
|
|
=item * the mean: |
1178
|
|
|
|
|
|
|
|
1179
|
|
|
|
|
|
|
MEAN = sum (x)/ N |
1180
|
|
|
|
|
|
|
|
1181
|
|
|
|
|
|
|
with C being the number of elements in x |
1182
|
|
|
|
|
|
|
|
1183
|
|
|
|
|
|
|
=item * the population RMS deviation from the mean: |
1184
|
|
|
|
|
|
|
|
1185
|
|
|
|
|
|
|
PRMS = sqrt( sum( (x-mean(x))^2 )/(N-1) |
1186
|
|
|
|
|
|
|
|
1187
|
|
|
|
|
|
|
The population deviation is the best-estimate of the deviation |
1188
|
|
|
|
|
|
|
of the population from which a sample is drawn. |
1189
|
|
|
|
|
|
|
|
1190
|
|
|
|
|
|
|
=item * the median |
1191
|
|
|
|
|
|
|
|
1192
|
|
|
|
|
|
|
The median is the 50th percentile data value. Median is found by |
1193
|
|
|
|
|
|
|
L, so WEIGHTING IS IGNORED FOR THE MEDIAN CALCULATION. |
1194
|
|
|
|
|
|
|
|
1195
|
|
|
|
|
|
|
=item * the minimum |
1196
|
|
|
|
|
|
|
|
1197
|
|
|
|
|
|
|
=item * the maximum |
1198
|
|
|
|
|
|
|
|
1199
|
|
|
|
|
|
|
=item * the average absolute deviation: |
1200
|
|
|
|
|
|
|
|
1201
|
|
|
|
|
|
|
AADEV = sum( abs(x-mean(x)) )/N |
1202
|
|
|
|
|
|
|
|
1203
|
|
|
|
|
|
|
=item * RMS deviation from the mean: |
1204
|
|
|
|
|
|
|
|
1205
|
|
|
|
|
|
|
RMS = sqrt(sum( (x-mean(x))^2 )/N) |
1206
|
|
|
|
|
|
|
|
1207
|
|
|
|
|
|
|
(also known as the root-mean-square deviation, or the square root of the |
1208
|
|
|
|
|
|
|
variance) |
1209
|
|
|
|
|
|
|
|
1210
|
|
|
|
|
|
|
=back |
1211
|
|
|
|
|
|
|
|
1212
|
|
|
|
|
|
|
This operator is a projection operator so the calculation |
1213
|
|
|
|
|
|
|
will take place over the final dimension. Thus if the input |
1214
|
|
|
|
|
|
|
is N-dimensional each returned value will be N-1 dimensional, |
1215
|
|
|
|
|
|
|
to calculate the statistics for the entire piddle either |
1216
|
|
|
|
|
|
|
use C directly on the piddle or call C. |
1217
|
|
|
|
|
|
|
|
1218
|
|
|
|
|
|
|
|
1219
|
|
|
|
|
|
|
|
1220
|
|
|
|
|
|
|
=for bad |
1221
|
|
|
|
|
|
|
|
1222
|
|
|
|
|
|
|
=for bad |
1223
|
|
|
|
|
|
|
|
1224
|
|
|
|
|
|
|
Bad values are simply ignored in the calculation, effectively reducing |
1225
|
|
|
|
|
|
|
the sample size. If all data are bad then the output data are marked bad. |
1226
|
|
|
|
|
|
|
|
1227
|
|
|
|
|
|
|
|
1228
|
|
|
|
|
|
|
|
1229
|
|
|
|
|
|
|
=cut |
1230
|
|
|
|
|
|
|
|
1231
|
|
|
|
|
|
|
|
1232
|
|
|
|
|
|
|
|
1233
|
|
|
|
|
|
|
|
1234
|
|
|
|
|
|
|
|
1235
|
|
|
|
|
|
|
|
1236
|
|
|
|
|
|
|
sub PDLA::statsover { |
1237
|
9
|
50
|
|
9
|
0
|
31
|
barf('Usage: ($mean,[$prms, $median, $min, $max, $adev, $rms]) = statsover($data,[$weights])') if $#_>1; |
1238
|
9
|
|
|
|
|
24
|
my ($data, $weights) = @_; |
1239
|
9
|
100
|
|
|
|
37
|
$weights = $data->ones() if !defined($weights); |
1240
|
|
|
|
|
|
|
|
1241
|
9
|
|
|
|
|
316
|
my $median = $data->medover(); |
1242
|
9
|
|
|
|
|
69
|
my $mean = PDLA->nullcreate($data); |
1243
|
9
|
|
|
|
|
28
|
my $rms = PDLA->nullcreate($data); |
1244
|
9
|
|
|
|
|
25
|
my $min = PDLA->nullcreate($data); |
1245
|
9
|
|
|
|
|
24
|
my $max = PDLA->nullcreate($data); |
1246
|
9
|
|
|
|
|
35
|
my $adev = PDLA->nullcreate($data); |
1247
|
9
|
|
|
|
|
23
|
my $prms = PDLA->nullcreate($data); |
1248
|
9
|
|
|
|
|
425
|
&PDLA::_statsover_int($data, $weights, $mean, $prms, $median, $min, $max, $adev, $rms); |
1249
|
|
|
|
|
|
|
|
1250
|
9
|
100
|
|
|
|
77
|
return $mean unless wantarray; |
1251
|
5
|
|
|
|
|
50
|
return ($mean, $prms, $median, $min, $max, $adev, $rms); |
1252
|
|
|
|
|
|
|
} |
1253
|
|
|
|
|
|
|
|
1254
|
|
|
|
|
|
|
|
1255
|
|
|
|
|
|
|
|
1256
|
|
|
|
|
|
|
*statsover = \&PDLA::statsover; |
1257
|
|
|
|
|
|
|
|
1258
|
|
|
|
|
|
|
|
1259
|
|
|
|
|
|
|
|
1260
|
|
|
|
|
|
|
|
1261
|
|
|
|
|
|
|
=head2 stats |
1262
|
|
|
|
|
|
|
|
1263
|
|
|
|
|
|
|
=for ref |
1264
|
|
|
|
|
|
|
|
1265
|
|
|
|
|
|
|
Calculates useful statistics on a piddle |
1266
|
|
|
|
|
|
|
|
1267
|
|
|
|
|
|
|
=for usage |
1268
|
|
|
|
|
|
|
|
1269
|
|
|
|
|
|
|
($mean,$prms,$median,$min,$max,$adev,$rms) = stats($piddle,[$weights]); |
1270
|
|
|
|
|
|
|
|
1271
|
|
|
|
|
|
|
This utility calculates all the most useful quantities in one call. |
1272
|
|
|
|
|
|
|
It works the same way as L, except that the quantities are |
1273
|
|
|
|
|
|
|
calculated considering the entire input PDLA as a single sample, rather |
1274
|
|
|
|
|
|
|
than as a collection of rows. See L for definitions of the |
1275
|
|
|
|
|
|
|
returned quantities. |
1276
|
|
|
|
|
|
|
|
1277
|
|
|
|
|
|
|
=cut |
1278
|
|
|
|
|
|
|
|
1279
|
|
|
|
|
|
|
|
1280
|
|
|
|
|
|
|
|
1281
|
|
|
|
|
|
|
|
1282
|
|
|
|
|
|
|
=for bad |
1283
|
|
|
|
|
|
|
|
1284
|
|
|
|
|
|
|
Bad values are handled; if all input values are bad, then all of the output |
1285
|
|
|
|
|
|
|
values are flagged bad. |
1286
|
|
|
|
|
|
|
|
1287
|
|
|
|
|
|
|
=cut |
1288
|
|
|
|
|
|
|
|
1289
|
|
|
|
|
|
|
|
1290
|
|
|
|
|
|
|
|
1291
|
|
|
|
|
|
|
*stats = \&PDLA::stats; |
1292
|
|
|
|
|
|
|
sub PDLA::stats { |
1293
|
7
|
50
|
|
7
|
0
|
161
|
barf('Usage: ($mean,[$rms]) = stats($data,[$weights])') if $#_>1; |
1294
|
7
|
|
|
|
|
83
|
my ($data,$weights) = @_; |
1295
|
|
|
|
|
|
|
|
1296
|
|
|
|
|
|
|
# Ensure that $weights is properly threaded over; this could be |
1297
|
|
|
|
|
|
|
# done rather more efficiently... |
1298
|
7
|
100
|
|
|
|
24
|
if(defined $weights) { |
1299
|
1
|
50
|
|
|
|
5
|
$weights = pdl($weights) unless UNIVERSAL::isa($weights,'PDLA'); |
1300
|
1
|
50
|
33
|
|
|
9
|
if( ($weights->ndims != $data->ndims) or |
1301
|
|
|
|
|
|
|
(pdl($weights->dims) != pdl($data->dims))->or |
1302
|
|
|
|
|
|
|
) { |
1303
|
0
|
|
|
|
|
0
|
$weights = $weights + zeroes($data) |
1304
|
|
|
|
|
|
|
} |
1305
|
1
|
|
|
|
|
7
|
$weights = $weights->flat; |
1306
|
|
|
|
|
|
|
} |
1307
|
|
|
|
|
|
|
|
1308
|
7
|
|
|
|
|
32
|
return PDLA::statsover($data->flat,$weights); |
1309
|
|
|
|
|
|
|
} |
1310
|
|
|
|
|
|
|
|
1311
|
|
|
|
|
|
|
|
1312
|
|
|
|
|
|
|
|
1313
|
|
|
|
|
|
|
|
1314
|
|
|
|
|
|
|
=head2 histogram |
1315
|
|
|
|
|
|
|
|
1316
|
|
|
|
|
|
|
=for sig |
1317
|
|
|
|
|
|
|
|
1318
|
|
|
|
|
|
|
Signature: (in(n); int+[o] hist(m); double step; double min; int msize => m) |
1319
|
|
|
|
|
|
|
|
1320
|
|
|
|
|
|
|
|
1321
|
|
|
|
|
|
|
=for ref |
1322
|
|
|
|
|
|
|
|
1323
|
|
|
|
|
|
|
Calculates a histogram for given stepsize and minimum. |
1324
|
|
|
|
|
|
|
|
1325
|
|
|
|
|
|
|
=for usage |
1326
|
|
|
|
|
|
|
|
1327
|
|
|
|
|
|
|
$h = histogram($data, $step, $min, $numbins); |
1328
|
|
|
|
|
|
|
$hist = zeroes $numbins; # Put histogram in existing piddle. |
1329
|
|
|
|
|
|
|
histogram($data, $hist, $step, $min, $numbins); |
1330
|
|
|
|
|
|
|
|
1331
|
|
|
|
|
|
|
The histogram will contain C<$numbins> bins starting from C<$min>, each |
1332
|
|
|
|
|
|
|
C<$step> wide. The value in each bin is the number of |
1333
|
|
|
|
|
|
|
values in C<$data> that lie within the bin limits. |
1334
|
|
|
|
|
|
|
|
1335
|
|
|
|
|
|
|
|
1336
|
|
|
|
|
|
|
Data below the lower limit is put in the first bin, and data above the |
1337
|
|
|
|
|
|
|
upper limit is put in the last bin. |
1338
|
|
|
|
|
|
|
|
1339
|
|
|
|
|
|
|
The output is reset in a different threadloop so that you |
1340
|
|
|
|
|
|
|
can take a histogram of C<$a(10,12)> into C<$b(15)> and get the result |
1341
|
|
|
|
|
|
|
you want. |
1342
|
|
|
|
|
|
|
|
1343
|
|
|
|
|
|
|
For a higher-level interface, see L. |
1344
|
|
|
|
|
|
|
|
1345
|
|
|
|
|
|
|
=for example |
1346
|
|
|
|
|
|
|
|
1347
|
|
|
|
|
|
|
pdla> p histogram(pdl(1,1,2),1,0,3) |
1348
|
|
|
|
|
|
|
[0 2 1] |
1349
|
|
|
|
|
|
|
|
1350
|
|
|
|
|
|
|
|
1351
|
|
|
|
|
|
|
|
1352
|
|
|
|
|
|
|
=for bad |
1353
|
|
|
|
|
|
|
|
1354
|
|
|
|
|
|
|
histogram processes bad values. |
1355
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
1356
|
|
|
|
|
|
|
|
1357
|
|
|
|
|
|
|
|
1358
|
|
|
|
|
|
|
=cut |
1359
|
|
|
|
|
|
|
|
1360
|
|
|
|
|
|
|
|
1361
|
|
|
|
|
|
|
|
1362
|
|
|
|
|
|
|
|
1363
|
|
|
|
|
|
|
|
1364
|
|
|
|
|
|
|
|
1365
|
|
|
|
|
|
|
*histogram = \&PDLA::histogram; |
1366
|
|
|
|
|
|
|
|
1367
|
|
|
|
|
|
|
|
1368
|
|
|
|
|
|
|
|
1369
|
|
|
|
|
|
|
|
1370
|
|
|
|
|
|
|
|
1371
|
|
|
|
|
|
|
=head2 whistogram |
1372
|
|
|
|
|
|
|
|
1373
|
|
|
|
|
|
|
=for sig |
1374
|
|
|
|
|
|
|
|
1375
|
|
|
|
|
|
|
Signature: (in(n); float+ wt(n);float+[o] hist(m); double step; double min; int msize => m) |
1376
|
|
|
|
|
|
|
|
1377
|
|
|
|
|
|
|
|
1378
|
|
|
|
|
|
|
=for ref |
1379
|
|
|
|
|
|
|
|
1380
|
|
|
|
|
|
|
Calculates a histogram from weighted data for given stepsize and minimum. |
1381
|
|
|
|
|
|
|
|
1382
|
|
|
|
|
|
|
=for usage |
1383
|
|
|
|
|
|
|
|
1384
|
|
|
|
|
|
|
$h = whistogram($data, $weights, $step, $min, $numbins); |
1385
|
|
|
|
|
|
|
$hist = zeroes $numbins; # Put histogram in existing piddle. |
1386
|
|
|
|
|
|
|
whistogram($data, $weights, $hist, $step, $min, $numbins); |
1387
|
|
|
|
|
|
|
|
1388
|
|
|
|
|
|
|
The histogram will contain C<$numbins> bins starting from C<$min>, each |
1389
|
|
|
|
|
|
|
C<$step> wide. The value in each bin is the sum of the values in C<$weights> |
1390
|
|
|
|
|
|
|
that correspond to values in C<$data> that lie within the bin limits. |
1391
|
|
|
|
|
|
|
|
1392
|
|
|
|
|
|
|
Data below the lower limit is put in the first bin, and data above the |
1393
|
|
|
|
|
|
|
upper limit is put in the last bin. |
1394
|
|
|
|
|
|
|
|
1395
|
|
|
|
|
|
|
The output is reset in a different threadloop so that you |
1396
|
|
|
|
|
|
|
can take a histogram of C<$a(10,12)> into C<$b(15)> and get the result |
1397
|
|
|
|
|
|
|
you want. |
1398
|
|
|
|
|
|
|
|
1399
|
|
|
|
|
|
|
=for example |
1400
|
|
|
|
|
|
|
|
1401
|
|
|
|
|
|
|
pdla> p whistogram(pdl(1,1,2), pdl(0.1,0.1,0.5), 1, 0, 4) |
1402
|
|
|
|
|
|
|
[0 0.2 0.5 0] |
1403
|
|
|
|
|
|
|
|
1404
|
|
|
|
|
|
|
|
1405
|
|
|
|
|
|
|
|
1406
|
|
|
|
|
|
|
=for bad |
1407
|
|
|
|
|
|
|
|
1408
|
|
|
|
|
|
|
whistogram processes bad values. |
1409
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
1410
|
|
|
|
|
|
|
|
1411
|
|
|
|
|
|
|
|
1412
|
|
|
|
|
|
|
=cut |
1413
|
|
|
|
|
|
|
|
1414
|
|
|
|
|
|
|
|
1415
|
|
|
|
|
|
|
|
1416
|
|
|
|
|
|
|
|
1417
|
|
|
|
|
|
|
|
1418
|
|
|
|
|
|
|
|
1419
|
|
|
|
|
|
|
*whistogram = \&PDLA::whistogram; |
1420
|
|
|
|
|
|
|
|
1421
|
|
|
|
|
|
|
|
1422
|
|
|
|
|
|
|
|
1423
|
|
|
|
|
|
|
|
1424
|
|
|
|
|
|
|
|
1425
|
|
|
|
|
|
|
=head2 histogram2d |
1426
|
|
|
|
|
|
|
|
1427
|
|
|
|
|
|
|
=for sig |
1428
|
|
|
|
|
|
|
|
1429
|
|
|
|
|
|
|
Signature: (ina(n); inb(n); int+[o] hist(ma,mb); double stepa; double mina; int masize => ma; |
1430
|
|
|
|
|
|
|
double stepb; double minb; int mbsize => mb;) |
1431
|
|
|
|
|
|
|
|
1432
|
|
|
|
|
|
|
|
1433
|
|
|
|
|
|
|
=for ref |
1434
|
|
|
|
|
|
|
|
1435
|
|
|
|
|
|
|
Calculates a 2d histogram. |
1436
|
|
|
|
|
|
|
|
1437
|
|
|
|
|
|
|
=for usage |
1438
|
|
|
|
|
|
|
|
1439
|
|
|
|
|
|
|
$h = histogram2d($datax, $datay, $stepx, $minx, |
1440
|
|
|
|
|
|
|
$nbinx, $stepy, $miny, $nbiny); |
1441
|
|
|
|
|
|
|
$hist = zeroes $nbinx, $nbiny; # Put histogram in existing piddle. |
1442
|
|
|
|
|
|
|
histogram2d($datax, $datay, $hist, $stepx, $minx, |
1443
|
|
|
|
|
|
|
$nbinx, $stepy, $miny, $nbiny); |
1444
|
|
|
|
|
|
|
|
1445
|
|
|
|
|
|
|
The histogram will contain C<$nbinx> x C<$nbiny> bins, with the lower |
1446
|
|
|
|
|
|
|
limits of the first one at C<($minx, $miny)>, and with bin size |
1447
|
|
|
|
|
|
|
C<($stepx, $stepy)>. |
1448
|
|
|
|
|
|
|
The value in each bin is the number of |
1449
|
|
|
|
|
|
|
values in C<$datax> and C<$datay> that lie within the bin limits. |
1450
|
|
|
|
|
|
|
|
1451
|
|
|
|
|
|
|
Data below the lower limit is put in the first bin, and data above the |
1452
|
|
|
|
|
|
|
upper limit is put in the last bin. |
1453
|
|
|
|
|
|
|
|
1454
|
|
|
|
|
|
|
=for example |
1455
|
|
|
|
|
|
|
|
1456
|
|
|
|
|
|
|
pdla> p histogram2d(pdl(1,1,1,2,2),pdl(2,1,1,1,1),1,0,3,1,0,3) |
1457
|
|
|
|
|
|
|
[ |
1458
|
|
|
|
|
|
|
[0 0 0] |
1459
|
|
|
|
|
|
|
[0 2 2] |
1460
|
|
|
|
|
|
|
[0 1 0] |
1461
|
|
|
|
|
|
|
] |
1462
|
|
|
|
|
|
|
|
1463
|
|
|
|
|
|
|
|
1464
|
|
|
|
|
|
|
|
1465
|
|
|
|
|
|
|
=for bad |
1466
|
|
|
|
|
|
|
|
1467
|
|
|
|
|
|
|
histogram2d processes bad values. |
1468
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
1469
|
|
|
|
|
|
|
|
1470
|
|
|
|
|
|
|
|
1471
|
|
|
|
|
|
|
=cut |
1472
|
|
|
|
|
|
|
|
1473
|
|
|
|
|
|
|
|
1474
|
|
|
|
|
|
|
|
1475
|
|
|
|
|
|
|
|
1476
|
|
|
|
|
|
|
|
1477
|
|
|
|
|
|
|
|
1478
|
|
|
|
|
|
|
*histogram2d = \&PDLA::histogram2d; |
1479
|
|
|
|
|
|
|
|
1480
|
|
|
|
|
|
|
|
1481
|
|
|
|
|
|
|
|
1482
|
|
|
|
|
|
|
|
1483
|
|
|
|
|
|
|
|
1484
|
|
|
|
|
|
|
=head2 whistogram2d |
1485
|
|
|
|
|
|
|
|
1486
|
|
|
|
|
|
|
=for sig |
1487
|
|
|
|
|
|
|
|
1488
|
|
|
|
|
|
|
Signature: (ina(n); inb(n); float+ wt(n);float+[o] hist(ma,mb); double stepa; double mina; int masize => ma; |
1489
|
|
|
|
|
|
|
double stepb; double minb; int mbsize => mb;) |
1490
|
|
|
|
|
|
|
|
1491
|
|
|
|
|
|
|
|
1492
|
|
|
|
|
|
|
=for ref |
1493
|
|
|
|
|
|
|
|
1494
|
|
|
|
|
|
|
Calculates a 2d histogram from weighted data. |
1495
|
|
|
|
|
|
|
|
1496
|
|
|
|
|
|
|
=for usage |
1497
|
|
|
|
|
|
|
|
1498
|
|
|
|
|
|
|
$h = whistogram2d($datax, $datay, $weights, |
1499
|
|
|
|
|
|
|
$stepx, $minx, $nbinx, $stepy, $miny, $nbiny); |
1500
|
|
|
|
|
|
|
$hist = zeroes $nbinx, $nbiny; # Put histogram in existing piddle. |
1501
|
|
|
|
|
|
|
whistogram2d($datax, $datay, $weights, $hist, |
1502
|
|
|
|
|
|
|
$stepx, $minx, $nbinx, $stepy, $miny, $nbiny); |
1503
|
|
|
|
|
|
|
|
1504
|
|
|
|
|
|
|
The histogram will contain C<$nbinx> x C<$nbiny> bins, with the lower |
1505
|
|
|
|
|
|
|
limits of the first one at C<($minx, $miny)>, and with bin size |
1506
|
|
|
|
|
|
|
C<($stepx, $stepy)>. |
1507
|
|
|
|
|
|
|
The value in each bin is the sum of the values in |
1508
|
|
|
|
|
|
|
C<$weights> that correspond to values in C<$datax> and C<$datay> that lie within the bin limits. |
1509
|
|
|
|
|
|
|
|
1510
|
|
|
|
|
|
|
Data below the lower limit is put in the first bin, and data above the |
1511
|
|
|
|
|
|
|
upper limit is put in the last bin. |
1512
|
|
|
|
|
|
|
|
1513
|
|
|
|
|
|
|
=for example |
1514
|
|
|
|
|
|
|
|
1515
|
|
|
|
|
|
|
pdla> p whistogram2d(pdl(1,1,1,2,2),pdl(2,1,1,1,1),pdl(0.1,0.2,0.3,0.4,0.5),1,0,3,1,0,3) |
1516
|
|
|
|
|
|
|
[ |
1517
|
|
|
|
|
|
|
[ 0 0 0] |
1518
|
|
|
|
|
|
|
[ 0 0.5 0.9] |
1519
|
|
|
|
|
|
|
[ 0 0.1 0] |
1520
|
|
|
|
|
|
|
] |
1521
|
|
|
|
|
|
|
|
1522
|
|
|
|
|
|
|
|
1523
|
|
|
|
|
|
|
|
1524
|
|
|
|
|
|
|
|
1525
|
|
|
|
|
|
|
=for bad |
1526
|
|
|
|
|
|
|
|
1527
|
|
|
|
|
|
|
whistogram2d processes bad values. |
1528
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
1529
|
|
|
|
|
|
|
|
1530
|
|
|
|
|
|
|
|
1531
|
|
|
|
|
|
|
=cut |
1532
|
|
|
|
|
|
|
|
1533
|
|
|
|
|
|
|
|
1534
|
|
|
|
|
|
|
|
1535
|
|
|
|
|
|
|
|
1536
|
|
|
|
|
|
|
|
1537
|
|
|
|
|
|
|
|
1538
|
|
|
|
|
|
|
*whistogram2d = \&PDLA::whistogram2d; |
1539
|
|
|
|
|
|
|
|
1540
|
|
|
|
|
|
|
|
1541
|
|
|
|
|
|
|
|
1542
|
|
|
|
|
|
|
|
1543
|
|
|
|
|
|
|
|
1544
|
|
|
|
|
|
|
=head2 fibonacci |
1545
|
|
|
|
|
|
|
|
1546
|
|
|
|
|
|
|
=for sig |
1547
|
|
|
|
|
|
|
|
1548
|
|
|
|
|
|
|
Signature: ([o]x(n)) |
1549
|
|
|
|
|
|
|
|
1550
|
|
|
|
|
|
|
=for ref |
1551
|
|
|
|
|
|
|
|
1552
|
|
|
|
|
|
|
Constructor - a vector with Fibonacci's sequence |
1553
|
|
|
|
|
|
|
|
1554
|
|
|
|
|
|
|
=for bad |
1555
|
|
|
|
|
|
|
|
1556
|
|
|
|
|
|
|
fibonacci does not process bad values. |
1557
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
1558
|
|
|
|
|
|
|
|
1559
|
|
|
|
|
|
|
|
1560
|
|
|
|
|
|
|
=cut |
1561
|
|
|
|
|
|
|
|
1562
|
|
|
|
|
|
|
|
1563
|
|
|
|
|
|
|
|
1564
|
|
|
|
|
|
|
|
1565
|
1
|
50
|
33
|
1
|
1
|
383
|
sub fibonacci { ref($_[0]) && ref($_[0]) ne 'PDLA::Type' ? $_[0]->fibonacci : PDLA->fibonacci(@_) } |
1566
|
|
|
|
|
|
|
sub PDLA::fibonacci{ |
1567
|
1
|
|
|
1
|
0
|
3
|
my $class = shift; |
1568
|
1
|
50
|
|
|
|
7
|
my $x = scalar(@_)? $class->new_from_specification(@_) : $class->new_or_inplace; |
1569
|
1
|
|
|
|
|
4
|
&PDLA::_fibonacci_int($x->clump(-1)); |
1570
|
1
|
|
|
|
|
9
|
return $x; |
1571
|
|
|
|
|
|
|
} |
1572
|
|
|
|
|
|
|
|
1573
|
|
|
|
|
|
|
|
1574
|
|
|
|
|
|
|
|
1575
|
|
|
|
|
|
|
|
1576
|
|
|
|
|
|
|
|
1577
|
|
|
|
|
|
|
|
1578
|
|
|
|
|
|
|
|
1579
|
|
|
|
|
|
|
=head2 append |
1580
|
|
|
|
|
|
|
|
1581
|
|
|
|
|
|
|
=for sig |
1582
|
|
|
|
|
|
|
|
1583
|
|
|
|
|
|
|
Signature: (a(n); b(m); [o] c(mn)) |
1584
|
|
|
|
|
|
|
|
1585
|
|
|
|
|
|
|
|
1586
|
|
|
|
|
|
|
|
1587
|
|
|
|
|
|
|
=for ref |
1588
|
|
|
|
|
|
|
|
1589
|
|
|
|
|
|
|
append two piddles by concatenating along their first dimensions |
1590
|
|
|
|
|
|
|
|
1591
|
|
|
|
|
|
|
=for example |
1592
|
|
|
|
|
|
|
|
1593
|
|
|
|
|
|
|
$x = ones(2,4,7); |
1594
|
|
|
|
|
|
|
$y = sequence 5; |
1595
|
|
|
|
|
|
|
$c = $x->append($y); # size of $c is now (7,4,7) (a jumbo-piddle ;) |
1596
|
|
|
|
|
|
|
|
1597
|
|
|
|
|
|
|
C appends two piddles along their first dimensions. The rest of the |
1598
|
|
|
|
|
|
|
dimensions must be compatible in the threading sense. The resulting |
1599
|
|
|
|
|
|
|
size of the first dimension is the sum of the sizes of the first dimensions |
1600
|
|
|
|
|
|
|
of the two argument piddles - i.e. C. |
1601
|
|
|
|
|
|
|
|
1602
|
|
|
|
|
|
|
Similar functions include L (below), which can append more |
1603
|
|
|
|
|
|
|
than two piddles along an arbitrary dimension, and |
1604
|
|
|
|
|
|
|
L, which can append more than two piddles that all |
1605
|
|
|
|
|
|
|
have the same sized dimensions. |
1606
|
|
|
|
|
|
|
|
1607
|
|
|
|
|
|
|
|
1608
|
|
|
|
|
|
|
|
1609
|
|
|
|
|
|
|
=for bad |
1610
|
|
|
|
|
|
|
|
1611
|
|
|
|
|
|
|
append does not process bad values. |
1612
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
1613
|
|
|
|
|
|
|
|
1614
|
|
|
|
|
|
|
|
1615
|
|
|
|
|
|
|
=cut |
1616
|
|
|
|
|
|
|
|
1617
|
|
|
|
|
|
|
|
1618
|
|
|
|
|
|
|
|
1619
|
|
|
|
|
|
|
|
1620
|
|
|
|
|
|
|
|
1621
|
|
|
|
|
|
|
|
1622
|
|
|
|
|
|
|
*append = \&PDLA::append; |
1623
|
|
|
|
|
|
|
|
1624
|
|
|
|
|
|
|
|
1625
|
|
|
|
|
|
|
|
1626
|
|
|
|
|
|
|
|
1627
|
|
|
|
|
|
|
=head2 glue |
1628
|
|
|
|
|
|
|
|
1629
|
|
|
|
|
|
|
=for usage |
1630
|
|
|
|
|
|
|
|
1631
|
|
|
|
|
|
|
$c = $x->glue(,$y,...) |
1632
|
|
|
|
|
|
|
|
1633
|
|
|
|
|
|
|
=for ref |
1634
|
|
|
|
|
|
|
|
1635
|
|
|
|
|
|
|
Glue two or more PDLAs together along an arbitrary dimension |
1636
|
|
|
|
|
|
|
(N-D L). |
1637
|
|
|
|
|
|
|
|
1638
|
|
|
|
|
|
|
Sticks $x, $y, and all following arguments together along the |
1639
|
|
|
|
|
|
|
specified dimension. All other dimensions must be compatible in the |
1640
|
|
|
|
|
|
|
threading sense. |
1641
|
|
|
|
|
|
|
|
1642
|
|
|
|
|
|
|
Glue is permissive, in the sense that every PDLA is treated as having an |
1643
|
|
|
|
|
|
|
infinite number of trivial dimensions of order 1 -- so C<< $x->glue(3,$y) >> |
1644
|
|
|
|
|
|
|
works, even if $x and $y are only one dimensional. |
1645
|
|
|
|
|
|
|
|
1646
|
|
|
|
|
|
|
If one of the PDLAs has no elements, it is ignored. Likewise, if one |
1647
|
|
|
|
|
|
|
of them is actually the undefined value, it is treated as if it had no |
1648
|
|
|
|
|
|
|
elements. |
1649
|
|
|
|
|
|
|
|
1650
|
|
|
|
|
|
|
If the first parameter is a defined perl scalar rather than a pdl, |
1651
|
|
|
|
|
|
|
then it is taken as a dimension along which to glue everything else, |
1652
|
|
|
|
|
|
|
so you can say C<$cube = PDLA::glue(3,@image_list);> if you like. |
1653
|
|
|
|
|
|
|
|
1654
|
|
|
|
|
|
|
C is implemented in pdl, using a combination of L and |
1655
|
|
|
|
|
|
|
L. It should probably be updated (one day) to a pure PP |
1656
|
|
|
|
|
|
|
function. |
1657
|
|
|
|
|
|
|
|
1658
|
|
|
|
|
|
|
Similar functions include L (above), which appends |
1659
|
|
|
|
|
|
|
only two piddles along their first dimension, and |
1660
|
|
|
|
|
|
|
L, which can append more than two piddles that all |
1661
|
|
|
|
|
|
|
have the same sized dimensions. |
1662
|
|
|
|
|
|
|
|
1663
|
|
|
|
|
|
|
=cut |
1664
|
|
|
|
|
|
|
|
1665
|
|
|
|
|
|
|
sub PDLA::glue{ |
1666
|
2
|
|
|
2
|
0
|
70
|
my($x) = shift; |
1667
|
2
|
|
|
|
|
4
|
my($dim) = shift; |
1668
|
|
|
|
|
|
|
|
1669
|
2
|
50
|
33
|
|
|
12
|
if(defined $x && !(ref $x)) { |
1670
|
0
|
|
|
|
|
0
|
my $y = $dim; |
1671
|
0
|
|
|
|
|
0
|
$dim = $x; |
1672
|
0
|
|
|
|
|
0
|
$x = $y; |
1673
|
|
|
|
|
|
|
} |
1674
|
|
|
|
|
|
|
|
1675
|
2
|
50
|
33
|
|
|
16
|
if(!defined $x || $x->nelem==0) { |
1676
|
0
|
0
|
|
|
|
0
|
return $x unless(@_); |
1677
|
0
|
0
|
|
|
|
0
|
return shift() if(@_<=1); |
1678
|
0
|
|
|
|
|
0
|
$x=shift; |
1679
|
0
|
|
|
|
|
0
|
return PDLA::glue($x,$dim,@_); |
1680
|
|
|
|
|
|
|
} |
1681
|
|
|
|
|
|
|
|
1682
|
2
|
50
|
|
|
|
9
|
if($dim - $x->dim(0) > 100) { |
1683
|
0
|
|
|
|
|
0
|
print STDERR "warning:: PDLA::glue allocating >100 dimensions!\n"; |
1684
|
|
|
|
|
|
|
} |
1685
|
2
|
|
|
|
|
12
|
while($dim >= $x->ndims) { |
1686
|
0
|
|
|
|
|
0
|
$x = $x->dummy(-1,1); |
1687
|
|
|
|
|
|
|
} |
1688
|
2
|
|
|
|
|
14
|
$x = $x->xchg(0,$dim); |
1689
|
|
|
|
|
|
|
|
1690
|
2
|
|
|
|
|
6
|
while(scalar(@_)){ |
1691
|
4
|
|
|
|
|
15
|
my $y = shift; |
1692
|
4
|
50
|
33
|
|
|
22
|
next unless(defined $y && $y->nelem); |
1693
|
|
|
|
|
|
|
|
1694
|
4
|
|
|
|
|
12
|
while($dim >= $y->ndims) { |
1695
|
0
|
|
|
|
|
0
|
$y = $y->dummy(-1,1); |
1696
|
|
|
|
|
|
|
} |
1697
|
4
|
|
|
|
|
18
|
$y = $y->xchg(0,$dim); |
1698
|
4
|
|
|
|
|
85
|
$x = $x->append($y); |
1699
|
|
|
|
|
|
|
} |
1700
|
2
|
|
|
|
|
21
|
$x->xchg(0,$dim); |
1701
|
|
|
|
|
|
|
} |
1702
|
|
|
|
|
|
|
|
1703
|
|
|
|
|
|
|
|
1704
|
|
|
|
|
|
|
|
1705
|
|
|
|
|
|
|
|
1706
|
|
|
|
|
|
|
|
1707
|
|
|
|
|
|
|
|
1708
|
|
|
|
|
|
|
|
1709
|
|
|
|
|
|
|
|
1710
|
|
|
|
|
|
|
=head2 axisvalues |
1711
|
|
|
|
|
|
|
|
1712
|
|
|
|
|
|
|
=for sig |
1713
|
|
|
|
|
|
|
|
1714
|
|
|
|
|
|
|
Signature: ([o,nc]a(n)) |
1715
|
|
|
|
|
|
|
|
1716
|
|
|
|
|
|
|
|
1717
|
|
|
|
|
|
|
|
1718
|
|
|
|
|
|
|
=for ref |
1719
|
|
|
|
|
|
|
|
1720
|
|
|
|
|
|
|
Internal routine |
1721
|
|
|
|
|
|
|
|
1722
|
|
|
|
|
|
|
C is the internal primitive that implements |
1723
|
|
|
|
|
|
|
L |
1724
|
|
|
|
|
|
|
and alters its argument. |
1725
|
|
|
|
|
|
|
|
1726
|
|
|
|
|
|
|
|
1727
|
|
|
|
|
|
|
|
1728
|
|
|
|
|
|
|
=for bad |
1729
|
|
|
|
|
|
|
|
1730
|
|
|
|
|
|
|
axisvalues does not process bad values. |
1731
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
1732
|
|
|
|
|
|
|
|
1733
|
|
|
|
|
|
|
|
1734
|
|
|
|
|
|
|
=cut |
1735
|
|
|
|
|
|
|
|
1736
|
|
|
|
|
|
|
|
1737
|
|
|
|
|
|
|
|
1738
|
|
|
|
|
|
|
|
1739
|
|
|
|
|
|
|
|
1740
|
|
|
|
|
|
|
|
1741
|
|
|
|
|
|
|
*axisvalues = \&PDLA::axisvalues; |
1742
|
|
|
|
|
|
|
|
1743
|
|
|
|
|
|
|
|
1744
|
|
|
|
|
|
|
|
1745
|
|
|
|
|
|
|
|
1746
|
|
|
|
|
|
|
=head2 random |
1747
|
|
|
|
|
|
|
|
1748
|
|
|
|
|
|
|
=for ref |
1749
|
|
|
|
|
|
|
|
1750
|
|
|
|
|
|
|
Constructor which returns piddle of random numbers |
1751
|
|
|
|
|
|
|
|
1752
|
|
|
|
|
|
|
=for usage |
1753
|
|
|
|
|
|
|
|
1754
|
|
|
|
|
|
|
$x = random([type], $nx, $ny, $nz,...); |
1755
|
|
|
|
|
|
|
$x = random $y; |
1756
|
|
|
|
|
|
|
|
1757
|
|
|
|
|
|
|
etc (see L). |
1758
|
|
|
|
|
|
|
|
1759
|
|
|
|
|
|
|
This is the uniform distribution between 0 and 1 (assumedly |
1760
|
|
|
|
|
|
|
excluding 1 itself). The arguments are the same as C |
1761
|
|
|
|
|
|
|
(q.v.) - i.e. one can specify dimensions, types or give |
1762
|
|
|
|
|
|
|
a template. |
1763
|
|
|
|
|
|
|
|
1764
|
|
|
|
|
|
|
You can use the perl function L to seed the random |
1765
|
|
|
|
|
|
|
generator. For further details consult Perl's L |
1766
|
|
|
|
|
|
|
documentation. |
1767
|
|
|
|
|
|
|
|
1768
|
|
|
|
|
|
|
=head2 randsym |
1769
|
|
|
|
|
|
|
|
1770
|
|
|
|
|
|
|
=for ref |
1771
|
|
|
|
|
|
|
|
1772
|
|
|
|
|
|
|
Constructor which returns piddle of random numbers |
1773
|
|
|
|
|
|
|
|
1774
|
|
|
|
|
|
|
=for usage |
1775
|
|
|
|
|
|
|
|
1776
|
|
|
|
|
|
|
$x = randsym([type], $nx, $ny, $nz,...); |
1777
|
|
|
|
|
|
|
$x = randsym $y; |
1778
|
|
|
|
|
|
|
|
1779
|
|
|
|
|
|
|
etc (see L). |
1780
|
|
|
|
|
|
|
|
1781
|
|
|
|
|
|
|
This is the uniform distribution between 0 and 1 (excluding both 0 and |
1782
|
|
|
|
|
|
|
1, cf L). The arguments are the same as C (q.v.) - |
1783
|
|
|
|
|
|
|
i.e. one can specify dimensions, types or give a template. |
1784
|
|
|
|
|
|
|
|
1785
|
|
|
|
|
|
|
You can use the perl function L to seed the random |
1786
|
|
|
|
|
|
|
generator. For further details consult Perl's L |
1787
|
|
|
|
|
|
|
documentation. |
1788
|
|
|
|
|
|
|
|
1789
|
|
|
|
|
|
|
=cut |
1790
|
|
|
|
|
|
|
|
1791
|
|
|
|
|
|
|
|
1792
|
|
|
|
|
|
|
|
1793
|
9
|
100
|
66
|
9
|
1
|
2918
|
sub random { ref($_[0]) && ref($_[0]) ne 'PDLA::Type' ? $_[0]->random : PDLA->random(@_) } |
1794
|
|
|
|
|
|
|
sub PDLA::random { |
1795
|
9
|
|
|
9
|
0
|
20
|
my $class = shift; |
1796
|
9
|
100
|
|
|
|
45
|
my $x = scalar(@_)? $class->new_from_specification(@_) : $class->new_or_inplace; |
1797
|
9
|
|
|
|
|
105
|
&PDLA::_random_int($x); |
1798
|
9
|
|
|
|
|
178
|
return $x; |
1799
|
|
|
|
|
|
|
} |
1800
|
|
|
|
|
|
|
|
1801
|
|
|
|
|
|
|
|
1802
|
|
|
|
|
|
|
|
1803
|
|
|
|
|
|
|
|
1804
|
|
|
|
|
|
|
|
1805
|
2
|
50
|
33
|
2
|
1
|
14
|
sub randsym { ref($_[0]) && ref($_[0]) ne 'PDLA::Type' ? $_[0]->randsym : PDLA->randsym(@_) } |
1806
|
|
|
|
|
|
|
sub PDLA::randsym { |
1807
|
2
|
|
|
2
|
0
|
4
|
my $class = shift; |
1808
|
2
|
50
|
|
|
|
11
|
my $x = scalar(@_)? $class->new_from_specification(@_) : $class->new_or_inplace; |
1809
|
2
|
|
|
|
|
18
|
&PDLA::_randsym_int($x); |
1810
|
2
|
|
|
|
|
84
|
return $x; |
1811
|
|
|
|
|
|
|
} |
1812
|
|
|
|
|
|
|
|
1813
|
|
|
|
|
|
|
|
1814
|
|
|
|
|
|
|
|
1815
|
|
|
|
|
|
|
|
1816
|
|
|
|
|
|
|
|
1817
|
|
|
|
|
|
|
|
1818
|
|
|
|
|
|
|
=head2 grandom |
1819
|
|
|
|
|
|
|
|
1820
|
|
|
|
|
|
|
=for ref |
1821
|
|
|
|
|
|
|
|
1822
|
|
|
|
|
|
|
Constructor which returns piddle of Gaussian random numbers |
1823
|
|
|
|
|
|
|
|
1824
|
|
|
|
|
|
|
=for usage |
1825
|
|
|
|
|
|
|
|
1826
|
|
|
|
|
|
|
$x = grandom([type], $nx, $ny, $nz,...); |
1827
|
|
|
|
|
|
|
$x = grandom $y; |
1828
|
|
|
|
|
|
|
|
1829
|
|
|
|
|
|
|
etc (see L). |
1830
|
|
|
|
|
|
|
|
1831
|
|
|
|
|
|
|
This is generated using the math library routine C. |
1832
|
|
|
|
|
|
|
|
1833
|
|
|
|
|
|
|
Mean = 0, Stddev = 1 |
1834
|
|
|
|
|
|
|
|
1835
|
|
|
|
|
|
|
|
1836
|
|
|
|
|
|
|
You can use the perl function L to seed the random |
1837
|
|
|
|
|
|
|
generator. For further details consult Perl's L |
1838
|
|
|
|
|
|
|
documentation. |
1839
|
|
|
|
|
|
|
|
1840
|
|
|
|
|
|
|
=cut |
1841
|
|
|
|
|
|
|
|
1842
|
2
|
50
|
33
|
2
|
1
|
301
|
sub grandom { ref($_[0]) && ref($_[0]) ne 'PDLA::Type' ? $_[0]->grandom : PDLA->grandom(@_) } |
1843
|
|
|
|
|
|
|
sub PDLA::grandom { |
1844
|
2
|
|
|
2
|
0
|
5
|
my $class = shift; |
1845
|
2
|
50
|
|
|
|
8
|
my $x = scalar(@_)? $class->new_from_specification(@_) : $class->new_or_inplace; |
1846
|
78
|
|
|
78
|
|
40662
|
use PDLA::Math 'ndtri'; |
|
78
|
|
|
|
|
283
|
|
|
78
|
|
|
|
|
642
|
|
1847
|
2
|
|
|
|
|
6
|
$x .= ndtri(randsym($x)); |
1848
|
2
|
|
|
|
|
15
|
return $x; |
1849
|
|
|
|
|
|
|
} |
1850
|
|
|
|
|
|
|
|
1851
|
|
|
|
|
|
|
|
1852
|
|
|
|
|
|
|
|
1853
|
|
|
|
|
|
|
|
1854
|
|
|
|
|
|
|
=head2 vsearch |
1855
|
|
|
|
|
|
|
|
1856
|
|
|
|
|
|
|
=for sig |
1857
|
|
|
|
|
|
|
|
1858
|
|
|
|
|
|
|
Signature: ( vals(); xs(n); [o] indx(); [\%options] ) |
1859
|
|
|
|
|
|
|
|
1860
|
|
|
|
|
|
|
=for ref |
1861
|
|
|
|
|
|
|
|
1862
|
|
|
|
|
|
|
Efficiently search for values in a sorted piddle, returning indices. |
1863
|
|
|
|
|
|
|
|
1864
|
|
|
|
|
|
|
=for usage |
1865
|
|
|
|
|
|
|
|
1866
|
|
|
|
|
|
|
$idx = vsearch( $vals, $x, [\%options] ); |
1867
|
|
|
|
|
|
|
vsearch( $vals, $x, $idx, [\%options ] ); |
1868
|
|
|
|
|
|
|
|
1869
|
|
|
|
|
|
|
B performs a binary search in the ordered piddle C<$x>, |
1870
|
|
|
|
|
|
|
for the values from C<$vals> piddle, returning indices into C<$x>. |
1871
|
|
|
|
|
|
|
What is a "match", and the meaning of the returned indices, are determined |
1872
|
|
|
|
|
|
|
by the options. |
1873
|
|
|
|
|
|
|
|
1874
|
|
|
|
|
|
|
The C option indicates which method of searching to use, and may |
1875
|
|
|
|
|
|
|
be one of: |
1876
|
|
|
|
|
|
|
|
1877
|
|
|
|
|
|
|
=over |
1878
|
|
|
|
|
|
|
|
1879
|
|
|
|
|
|
|
=item C |
1880
|
|
|
|
|
|
|
|
1881
|
|
|
|
|
|
|
invoke B, returning indices appropriate for sampling |
1882
|
|
|
|
|
|
|
within a distribution. |
1883
|
|
|
|
|
|
|
|
1884
|
|
|
|
|
|
|
=item C |
1885
|
|
|
|
|
|
|
|
1886
|
|
|
|
|
|
|
invoke B, returning the left-most possible |
1887
|
|
|
|
|
|
|
insertion point which still leaves the piddle sorted. |
1888
|
|
|
|
|
|
|
|
1889
|
|
|
|
|
|
|
=item C |
1890
|
|
|
|
|
|
|
|
1891
|
|
|
|
|
|
|
invoke B, returning the right-most possible |
1892
|
|
|
|
|
|
|
insertion point which still leaves the piddle sorted. |
1893
|
|
|
|
|
|
|
|
1894
|
|
|
|
|
|
|
=item C |
1895
|
|
|
|
|
|
|
|
1896
|
|
|
|
|
|
|
invoke B, returning the index of a matching element, |
1897
|
|
|
|
|
|
|
else -(insertion point + 1) |
1898
|
|
|
|
|
|
|
|
1899
|
|
|
|
|
|
|
=item C |
1900
|
|
|
|
|
|
|
|
1901
|
|
|
|
|
|
|
invoke B, returning an index appropriate for binning |
1902
|
|
|
|
|
|
|
on a grid where the left bin edges are I of the bin. See |
1903
|
|
|
|
|
|
|
below for further explanation of the bin. |
1904
|
|
|
|
|
|
|
|
1905
|
|
|
|
|
|
|
=item C |
1906
|
|
|
|
|
|
|
|
1907
|
|
|
|
|
|
|
invoke B, returning an index appropriate for binning |
1908
|
|
|
|
|
|
|
on a grid where the left bin edges are I of the bin. See |
1909
|
|
|
|
|
|
|
below for further explanation of the bin. |
1910
|
|
|
|
|
|
|
|
1911
|
|
|
|
|
|
|
=back |
1912
|
|
|
|
|
|
|
|
1913
|
|
|
|
|
|
|
The default value of C is C. |
1914
|
|
|
|
|
|
|
|
1915
|
|
|
|
|
|
|
=cut |
1916
|
|
|
|
|
|
|
|
1917
|
|
|
|
|
|
|
sub vsearch { |
1918
|
149
|
100
|
|
149
|
1
|
47581
|
my $opt = 'HASH' eq ref $_[-1] |
1919
|
|
|
|
|
|
|
? pop |
1920
|
|
|
|
|
|
|
: { mode => 'sample' }; |
1921
|
|
|
|
|
|
|
|
1922
|
|
|
|
|
|
|
croak( "unknown options to vsearch\n" ) |
1923
|
149
|
50
|
33
|
|
|
873
|
if ( ! defined $opt->{mode} && keys %$opt ) |
|
|
|
33
|
|
|
|
|
1924
|
|
|
|
|
|
|
|| keys %$opt > 1; |
1925
|
|
|
|
|
|
|
|
1926
|
149
|
|
|
|
|
271
|
my $mode = $opt->{mode}; |
1927
|
|
|
|
|
|
|
goto |
1928
|
149
|
50
|
|
|
|
9498
|
$mode eq 'sample' ? \&vsearch_sample |
|
|
100
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
1929
|
|
|
|
|
|
|
: $mode eq 'insert_leftmost' ? \&vsearch_insert_leftmost |
1930
|
|
|
|
|
|
|
: $mode eq 'insert_rightmost' ? \&vsearch_insert_rightmost |
1931
|
|
|
|
|
|
|
: $mode eq 'match' ? \&vsearch_match |
1932
|
|
|
|
|
|
|
: $mode eq 'bin_inclusive' ? \&vsearch_bin_inclusive |
1933
|
|
|
|
|
|
|
: $mode eq 'bin_exclusive' ? \&vsearch_bin_exclusive |
1934
|
|
|
|
|
|
|
: croak( "unknown vsearch mode: $mode\n" ); |
1935
|
|
|
|
|
|
|
} |
1936
|
|
|
|
|
|
|
|
1937
|
|
|
|
|
|
|
*PDLA::vsearch = \&vsearch; |
1938
|
|
|
|
|
|
|
|
1939
|
|
|
|
|
|
|
|
1940
|
|
|
|
|
|
|
|
1941
|
|
|
|
|
|
|
|
1942
|
|
|
|
|
|
|
|
1943
|
|
|
|
|
|
|
=head2 vsearch_sample |
1944
|
|
|
|
|
|
|
|
1945
|
|
|
|
|
|
|
=for sig |
1946
|
|
|
|
|
|
|
|
1947
|
|
|
|
|
|
|
Signature: (vals(); x(n); indx [o]idx()) |
1948
|
|
|
|
|
|
|
|
1949
|
|
|
|
|
|
|
|
1950
|
|
|
|
|
|
|
=for ref |
1951
|
|
|
|
|
|
|
|
1952
|
|
|
|
|
|
|
Search for values in a sorted array, return index appropriate for sampling from a distribution |
1953
|
|
|
|
|
|
|
|
1954
|
|
|
|
|
|
|
=for usage |
1955
|
|
|
|
|
|
|
|
1956
|
|
|
|
|
|
|
$idx = vsearch_sample($vals, $x); |
1957
|
|
|
|
|
|
|
|
1958
|
|
|
|
|
|
|
C<$x> must be sorted, but may be in decreasing or increasing |
1959
|
|
|
|
|
|
|
order. |
1960
|
|
|
|
|
|
|
|
1961
|
|
|
|
|
|
|
|
1962
|
|
|
|
|
|
|
|
1963
|
|
|
|
|
|
|
B returns an index I for each value I of C<$vals> appropriate |
1964
|
|
|
|
|
|
|
for sampling C<$vals> |
1965
|
|
|
|
|
|
|
|
1966
|
|
|
|
|
|
|
|
1967
|
|
|
|
|
|
|
|
1968
|
|
|
|
|
|
|
|
1969
|
|
|
|
|
|
|
|
1970
|
|
|
|
|
|
|
I has the following properties: |
1971
|
|
|
|
|
|
|
|
1972
|
|
|
|
|
|
|
=over |
1973
|
|
|
|
|
|
|
|
1974
|
|
|
|
|
|
|
=item * |
1975
|
|
|
|
|
|
|
|
1976
|
|
|
|
|
|
|
if C<$x> is sorted in increasing order |
1977
|
|
|
|
|
|
|
|
1978
|
|
|
|
|
|
|
|
1979
|
|
|
|
|
|
|
V <= x[0] : I = 0 |
1980
|
|
|
|
|
|
|
x[0] < V <= x[-1] : I s.t. x[I-1] < V <= x[I] |
1981
|
|
|
|
|
|
|
x[-1] < V : I = $x->nelem -1 |
1982
|
|
|
|
|
|
|
|
1983
|
|
|
|
|
|
|
|
1984
|
|
|
|
|
|
|
|
1985
|
|
|
|
|
|
|
=item * |
1986
|
|
|
|
|
|
|
|
1987
|
|
|
|
|
|
|
if C<$x> is sorted in decreasing order |
1988
|
|
|
|
|
|
|
|
1989
|
|
|
|
|
|
|
|
1990
|
|
|
|
|
|
|
V > x[0] : I = 0 |
1991
|
|
|
|
|
|
|
x[0] >= V > x[-1] : I s.t. x[I] >= V > x[I+1] |
1992
|
|
|
|
|
|
|
x[-1] >= V : I = $x->nelem - 1 |
1993
|
|
|
|
|
|
|
|
1994
|
|
|
|
|
|
|
|
1995
|
|
|
|
|
|
|
|
1996
|
|
|
|
|
|
|
=back |
1997
|
|
|
|
|
|
|
|
1998
|
|
|
|
|
|
|
|
1999
|
|
|
|
|
|
|
|
2000
|
|
|
|
|
|
|
|
2001
|
|
|
|
|
|
|
If all elements of C<$x> are equal, I<< I = $x->nelem - 1 >>. |
2002
|
|
|
|
|
|
|
|
2003
|
|
|
|
|
|
|
If C<$x> contains duplicated elements, I is the index of the |
2004
|
|
|
|
|
|
|
leftmost (by position in array) duplicate if I matches. |
2005
|
|
|
|
|
|
|
|
2006
|
|
|
|
|
|
|
=for example |
2007
|
|
|
|
|
|
|
|
2008
|
|
|
|
|
|
|
This function is useful e.g. when you have a list of probabilities |
2009
|
|
|
|
|
|
|
for events and want to generate indices to events: |
2010
|
|
|
|
|
|
|
|
2011
|
|
|
|
|
|
|
$x = pdl(.01,.86,.93,1); # Barnsley IFS probabilities cumulatively |
2012
|
|
|
|
|
|
|
$y = random 20; |
2013
|
|
|
|
|
|
|
$c = vsearch_sample($y, $x); # Now, $c will have the appropriate distr. |
2014
|
|
|
|
|
|
|
|
2015
|
|
|
|
|
|
|
It is possible to use the L |
2016
|
|
|
|
|
|
|
function to obtain cumulative probabilities from absolute probabilities. |
2017
|
|
|
|
|
|
|
|
2018
|
|
|
|
|
|
|
|
2019
|
|
|
|
|
|
|
|
2020
|
|
|
|
|
|
|
|
2021
|
|
|
|
|
|
|
|
2022
|
|
|
|
|
|
|
|
2023
|
|
|
|
|
|
|
|
2024
|
|
|
|
|
|
|
=for bad |
2025
|
|
|
|
|
|
|
|
2026
|
|
|
|
|
|
|
needs major (?) work to handles bad values |
2027
|
|
|
|
|
|
|
|
2028
|
|
|
|
|
|
|
=cut |
2029
|
|
|
|
|
|
|
|
2030
|
|
|
|
|
|
|
|
2031
|
|
|
|
|
|
|
|
2032
|
|
|
|
|
|
|
|
2033
|
|
|
|
|
|
|
|
2034
|
|
|
|
|
|
|
|
2035
|
|
|
|
|
|
|
*vsearch_sample = \&PDLA::vsearch_sample; |
2036
|
|
|
|
|
|
|
|
2037
|
|
|
|
|
|
|
|
2038
|
|
|
|
|
|
|
|
2039
|
|
|
|
|
|
|
|
2040
|
|
|
|
|
|
|
|
2041
|
|
|
|
|
|
|
=head2 vsearch_insert_leftmost |
2042
|
|
|
|
|
|
|
|
2043
|
|
|
|
|
|
|
=for sig |
2044
|
|
|
|
|
|
|
|
2045
|
|
|
|
|
|
|
Signature: (vals(); x(n); indx [o]idx()) |
2046
|
|
|
|
|
|
|
|
2047
|
|
|
|
|
|
|
|
2048
|
|
|
|
|
|
|
=for ref |
2049
|
|
|
|
|
|
|
|
2050
|
|
|
|
|
|
|
Determine the insertion point for values in a sorted array, inserting before duplicates. |
2051
|
|
|
|
|
|
|
|
2052
|
|
|
|
|
|
|
=for usage |
2053
|
|
|
|
|
|
|
|
2054
|
|
|
|
|
|
|
$idx = vsearch_insert_leftmost($vals, $x); |
2055
|
|
|
|
|
|
|
|
2056
|
|
|
|
|
|
|
C<$x> must be sorted, but may be in decreasing or increasing |
2057
|
|
|
|
|
|
|
order. |
2058
|
|
|
|
|
|
|
|
2059
|
|
|
|
|
|
|
|
2060
|
|
|
|
|
|
|
|
2061
|
|
|
|
|
|
|
B returns an index I for each value I of |
2062
|
|
|
|
|
|
|
C<$vals> equal to the leftmost position (by index in array) within |
2063
|
|
|
|
|
|
|
C<$x> that I may be inserted and still maintain the order in |
2064
|
|
|
|
|
|
|
C<$x>. |
2065
|
|
|
|
|
|
|
|
2066
|
|
|
|
|
|
|
Insertion at index I involves shifting elements I and higher of |
2067
|
|
|
|
|
|
|
C<$x> to the right by one and setting the now empty element at index |
2068
|
|
|
|
|
|
|
I to I. |
2069
|
|
|
|
|
|
|
|
2070
|
|
|
|
|
|
|
|
2071
|
|
|
|
|
|
|
|
2072
|
|
|
|
|
|
|
|
2073
|
|
|
|
|
|
|
|
2074
|
|
|
|
|
|
|
I has the following properties: |
2075
|
|
|
|
|
|
|
|
2076
|
|
|
|
|
|
|
=over |
2077
|
|
|
|
|
|
|
|
2078
|
|
|
|
|
|
|
=item * |
2079
|
|
|
|
|
|
|
|
2080
|
|
|
|
|
|
|
if C<$x> is sorted in increasing order |
2081
|
|
|
|
|
|
|
|
2082
|
|
|
|
|
|
|
|
2083
|
|
|
|
|
|
|
V <= x[0] : I = 0 |
2084
|
|
|
|
|
|
|
x[0] < V <= x[-1] : I s.t. x[I-1] < V <= x[I] |
2085
|
|
|
|
|
|
|
x[-1] < V : I = $x->nelem |
2086
|
|
|
|
|
|
|
|
2087
|
|
|
|
|
|
|
|
2088
|
|
|
|
|
|
|
|
2089
|
|
|
|
|
|
|
=item * |
2090
|
|
|
|
|
|
|
|
2091
|
|
|
|
|
|
|
if C<$x> is sorted in decreasing order |
2092
|
|
|
|
|
|
|
|
2093
|
|
|
|
|
|
|
|
2094
|
|
|
|
|
|
|
V > x[0] : I = -1 |
2095
|
|
|
|
|
|
|
x[0] >= V >= x[-1] : I s.t. x[I] >= V > x[I+1] |
2096
|
|
|
|
|
|
|
x[-1] >= V : I = $x->nelem -1 |
2097
|
|
|
|
|
|
|
|
2098
|
|
|
|
|
|
|
|
2099
|
|
|
|
|
|
|
|
2100
|
|
|
|
|
|
|
=back |
2101
|
|
|
|
|
|
|
|
2102
|
|
|
|
|
|
|
|
2103
|
|
|
|
|
|
|
|
2104
|
|
|
|
|
|
|
|
2105
|
|
|
|
|
|
|
If all elements of C<$x> are equal, |
2106
|
|
|
|
|
|
|
|
2107
|
|
|
|
|
|
|
i = 0 |
2108
|
|
|
|
|
|
|
|
2109
|
|
|
|
|
|
|
If C<$x> contains duplicated elements, I is the index of the |
2110
|
|
|
|
|
|
|
leftmost (by index in array) duplicate if I matches. |
2111
|
|
|
|
|
|
|
|
2112
|
|
|
|
|
|
|
|
2113
|
|
|
|
|
|
|
|
2114
|
|
|
|
|
|
|
|
2115
|
|
|
|
|
|
|
|
2116
|
|
|
|
|
|
|
|
2117
|
|
|
|
|
|
|
|
2118
|
|
|
|
|
|
|
=for bad |
2119
|
|
|
|
|
|
|
|
2120
|
|
|
|
|
|
|
needs major (?) work to handles bad values |
2121
|
|
|
|
|
|
|
|
2122
|
|
|
|
|
|
|
=cut |
2123
|
|
|
|
|
|
|
|
2124
|
|
|
|
|
|
|
|
2125
|
|
|
|
|
|
|
|
2126
|
|
|
|
|
|
|
|
2127
|
|
|
|
|
|
|
|
2128
|
|
|
|
|
|
|
|
2129
|
|
|
|
|
|
|
*vsearch_insert_leftmost = \&PDLA::vsearch_insert_leftmost; |
2130
|
|
|
|
|
|
|
|
2131
|
|
|
|
|
|
|
|
2132
|
|
|
|
|
|
|
|
2133
|
|
|
|
|
|
|
|
2134
|
|
|
|
|
|
|
|
2135
|
|
|
|
|
|
|
=head2 vsearch_insert_rightmost |
2136
|
|
|
|
|
|
|
|
2137
|
|
|
|
|
|
|
=for sig |
2138
|
|
|
|
|
|
|
|
2139
|
|
|
|
|
|
|
Signature: (vals(); x(n); indx [o]idx()) |
2140
|
|
|
|
|
|
|
|
2141
|
|
|
|
|
|
|
|
2142
|
|
|
|
|
|
|
=for ref |
2143
|
|
|
|
|
|
|
|
2144
|
|
|
|
|
|
|
Determine the insertion point for values in a sorted array, inserting after duplicates. |
2145
|
|
|
|
|
|
|
|
2146
|
|
|
|
|
|
|
=for usage |
2147
|
|
|
|
|
|
|
|
2148
|
|
|
|
|
|
|
$idx = vsearch_insert_rightmost($vals, $x); |
2149
|
|
|
|
|
|
|
|
2150
|
|
|
|
|
|
|
C<$x> must be sorted, but may be in decreasing or increasing |
2151
|
|
|
|
|
|
|
order. |
2152
|
|
|
|
|
|
|
|
2153
|
|
|
|
|
|
|
|
2154
|
|
|
|
|
|
|
|
2155
|
|
|
|
|
|
|
B returns an index I for each value I of |
2156
|
|
|
|
|
|
|
C<$vals> equal to the rightmost position (by index in array) within |
2157
|
|
|
|
|
|
|
C<$x> that I may be inserted and still maintain the order in |
2158
|
|
|
|
|
|
|
C<$x>. |
2159
|
|
|
|
|
|
|
|
2160
|
|
|
|
|
|
|
Insertion at index I involves shifting elements I and higher of |
2161
|
|
|
|
|
|
|
C<$x> to the right by one and setting the now empty element at index |
2162
|
|
|
|
|
|
|
I to I. |
2163
|
|
|
|
|
|
|
|
2164
|
|
|
|
|
|
|
|
2165
|
|
|
|
|
|
|
|
2166
|
|
|
|
|
|
|
|
2167
|
|
|
|
|
|
|
|
2168
|
|
|
|
|
|
|
I has the following properties: |
2169
|
|
|
|
|
|
|
|
2170
|
|
|
|
|
|
|
=over |
2171
|
|
|
|
|
|
|
|
2172
|
|
|
|
|
|
|
=item * |
2173
|
|
|
|
|
|
|
|
2174
|
|
|
|
|
|
|
if C<$x> is sorted in increasing order |
2175
|
|
|
|
|
|
|
|
2176
|
|
|
|
|
|
|
|
2177
|
|
|
|
|
|
|
V < x[0] : I = 0 |
2178
|
|
|
|
|
|
|
x[0] <= V < x[-1] : I s.t. x[I-1] <= V < x[I] |
2179
|
|
|
|
|
|
|
x[-1] <= V : I = $x->nelem |
2180
|
|
|
|
|
|
|
|
2181
|
|
|
|
|
|
|
|
2182
|
|
|
|
|
|
|
|
2183
|
|
|
|
|
|
|
=item * |
2184
|
|
|
|
|
|
|
|
2185
|
|
|
|
|
|
|
if C<$x> is sorted in decreasing order |
2186
|
|
|
|
|
|
|
|
2187
|
|
|
|
|
|
|
|
2188
|
|
|
|
|
|
|
V >= x[0] : I = -1 |
2189
|
|
|
|
|
|
|
x[0] > V >= x[-1] : I s.t. x[I] >= V > x[I+1] |
2190
|
|
|
|
|
|
|
x[-1] > V : I = $x->nelem -1 |
2191
|
|
|
|
|
|
|
|
2192
|
|
|
|
|
|
|
|
2193
|
|
|
|
|
|
|
|
2194
|
|
|
|
|
|
|
=back |
2195
|
|
|
|
|
|
|
|
2196
|
|
|
|
|
|
|
|
2197
|
|
|
|
|
|
|
|
2198
|
|
|
|
|
|
|
|
2199
|
|
|
|
|
|
|
If all elements of C<$x> are equal, |
2200
|
|
|
|
|
|
|
|
2201
|
|
|
|
|
|
|
i = $x->nelem - 1 |
2202
|
|
|
|
|
|
|
|
2203
|
|
|
|
|
|
|
If C<$x> contains duplicated elements, I is the index of the |
2204
|
|
|
|
|
|
|
leftmost (by index in array) duplicate if I matches. |
2205
|
|
|
|
|
|
|
|
2206
|
|
|
|
|
|
|
|
2207
|
|
|
|
|
|
|
|
2208
|
|
|
|
|
|
|
|
2209
|
|
|
|
|
|
|
|
2210
|
|
|
|
|
|
|
|
2211
|
|
|
|
|
|
|
|
2212
|
|
|
|
|
|
|
=for bad |
2213
|
|
|
|
|
|
|
|
2214
|
|
|
|
|
|
|
needs major (?) work to handles bad values |
2215
|
|
|
|
|
|
|
|
2216
|
|
|
|
|
|
|
=cut |
2217
|
|
|
|
|
|
|
|
2218
|
|
|
|
|
|
|
|
2219
|
|
|
|
|
|
|
|
2220
|
|
|
|
|
|
|
|
2221
|
|
|
|
|
|
|
|
2222
|
|
|
|
|
|
|
|
2223
|
|
|
|
|
|
|
*vsearch_insert_rightmost = \&PDLA::vsearch_insert_rightmost; |
2224
|
|
|
|
|
|
|
|
2225
|
|
|
|
|
|
|
|
2226
|
|
|
|
|
|
|
|
2227
|
|
|
|
|
|
|
|
2228
|
|
|
|
|
|
|
|
2229
|
|
|
|
|
|
|
=head2 vsearch_match |
2230
|
|
|
|
|
|
|
|
2231
|
|
|
|
|
|
|
=for sig |
2232
|
|
|
|
|
|
|
|
2233
|
|
|
|
|
|
|
Signature: (vals(); x(n); indx [o]idx()) |
2234
|
|
|
|
|
|
|
|
2235
|
|
|
|
|
|
|
|
2236
|
|
|
|
|
|
|
=for ref |
2237
|
|
|
|
|
|
|
|
2238
|
|
|
|
|
|
|
Match values against a sorted array. |
2239
|
|
|
|
|
|
|
|
2240
|
|
|
|
|
|
|
=for usage |
2241
|
|
|
|
|
|
|
|
2242
|
|
|
|
|
|
|
$idx = vsearch_match($vals, $x); |
2243
|
|
|
|
|
|
|
|
2244
|
|
|
|
|
|
|
C<$x> must be sorted, but may be in decreasing or increasing |
2245
|
|
|
|
|
|
|
order. |
2246
|
|
|
|
|
|
|
|
2247
|
|
|
|
|
|
|
|
2248
|
|
|
|
|
|
|
|
2249
|
|
|
|
|
|
|
B returns an index I for each value I of |
2250
|
|
|
|
|
|
|
C<$vals>. If I matches an element in C<$x>, I is the |
2251
|
|
|
|
|
|
|
index of that element, otherwise it is I<-( insertion_point + 1 )>, |
2252
|
|
|
|
|
|
|
where I is an index in C<$x> where I may be |
2253
|
|
|
|
|
|
|
inserted while maintaining the order in C<$x>. If C<$x> has |
2254
|
|
|
|
|
|
|
duplicated values, I may refer to any of them. |
2255
|
|
|
|
|
|
|
|
2256
|
|
|
|
|
|
|
|
2257
|
|
|
|
|
|
|
|
2258
|
|
|
|
|
|
|
|
2259
|
|
|
|
|
|
|
|
2260
|
|
|
|
|
|
|
|
2261
|
|
|
|
|
|
|
|
2262
|
|
|
|
|
|
|
|
2263
|
|
|
|
|
|
|
|
2264
|
|
|
|
|
|
|
=for bad |
2265
|
|
|
|
|
|
|
|
2266
|
|
|
|
|
|
|
needs major (?) work to handles bad values |
2267
|
|
|
|
|
|
|
|
2268
|
|
|
|
|
|
|
=cut |
2269
|
|
|
|
|
|
|
|
2270
|
|
|
|
|
|
|
|
2271
|
|
|
|
|
|
|
|
2272
|
|
|
|
|
|
|
|
2273
|
|
|
|
|
|
|
|
2274
|
|
|
|
|
|
|
|
2275
|
|
|
|
|
|
|
*vsearch_match = \&PDLA::vsearch_match; |
2276
|
|
|
|
|
|
|
|
2277
|
|
|
|
|
|
|
|
2278
|
|
|
|
|
|
|
|
2279
|
|
|
|
|
|
|
|
2280
|
|
|
|
|
|
|
|
2281
|
|
|
|
|
|
|
=head2 vsearch_bin_inclusive |
2282
|
|
|
|
|
|
|
|
2283
|
|
|
|
|
|
|
=for sig |
2284
|
|
|
|
|
|
|
|
2285
|
|
|
|
|
|
|
Signature: (vals(); x(n); indx [o]idx()) |
2286
|
|
|
|
|
|
|
|
2287
|
|
|
|
|
|
|
|
2288
|
|
|
|
|
|
|
=for ref |
2289
|
|
|
|
|
|
|
|
2290
|
|
|
|
|
|
|
Determine the index for values in a sorted array of bins, lower bound inclusive. |
2291
|
|
|
|
|
|
|
|
2292
|
|
|
|
|
|
|
=for usage |
2293
|
|
|
|
|
|
|
|
2294
|
|
|
|
|
|
|
$idx = vsearch_bin_inclusive($vals, $x); |
2295
|
|
|
|
|
|
|
|
2296
|
|
|
|
|
|
|
C<$x> must be sorted, but may be in decreasing or increasing |
2297
|
|
|
|
|
|
|
order. |
2298
|
|
|
|
|
|
|
|
2299
|
|
|
|
|
|
|
|
2300
|
|
|
|
|
|
|
|
2301
|
|
|
|
|
|
|
C<$x> represents the edges of contiguous bins, with the first and |
2302
|
|
|
|
|
|
|
last elements representing the outer edges of the outer bins, and the |
2303
|
|
|
|
|
|
|
inner elements the shared bin edges. |
2304
|
|
|
|
|
|
|
|
2305
|
|
|
|
|
|
|
The lower bound of a bin is inclusive to the bin, its outer bound is exclusive to it. |
2306
|
|
|
|
|
|
|
B returns an index I for each value I of C<$vals> |
2307
|
|
|
|
|
|
|
|
2308
|
|
|
|
|
|
|
|
2309
|
|
|
|
|
|
|
|
2310
|
|
|
|
|
|
|
|
2311
|
|
|
|
|
|
|
|
2312
|
|
|
|
|
|
|
I has the following properties: |
2313
|
|
|
|
|
|
|
|
2314
|
|
|
|
|
|
|
=over |
2315
|
|
|
|
|
|
|
|
2316
|
|
|
|
|
|
|
=item * |
2317
|
|
|
|
|
|
|
|
2318
|
|
|
|
|
|
|
if C<$x> is sorted in increasing order |
2319
|
|
|
|
|
|
|
|
2320
|
|
|
|
|
|
|
|
2321
|
|
|
|
|
|
|
V < x[0] : I = -1 |
2322
|
|
|
|
|
|
|
x[0] <= V < x[-1] : I s.t. x[I] <= V < x[I+1] |
2323
|
|
|
|
|
|
|
x[-1] <= V : I = $x->nelem - 1 |
2324
|
|
|
|
|
|
|
|
2325
|
|
|
|
|
|
|
|
2326
|
|
|
|
|
|
|
|
2327
|
|
|
|
|
|
|
=item * |
2328
|
|
|
|
|
|
|
|
2329
|
|
|
|
|
|
|
if C<$x> is sorted in decreasing order |
2330
|
|
|
|
|
|
|
|
2331
|
|
|
|
|
|
|
|
2332
|
|
|
|
|
|
|
V >= x[0] : I = 0 |
2333
|
|
|
|
|
|
|
x[0] > V >= x[-1] : I s.t. x[I+1] > V >= x[I] |
2334
|
|
|
|
|
|
|
x[-1] > V : I = $x->nelem |
2335
|
|
|
|
|
|
|
|
2336
|
|
|
|
|
|
|
|
2337
|
|
|
|
|
|
|
|
2338
|
|
|
|
|
|
|
=back |
2339
|
|
|
|
|
|
|
|
2340
|
|
|
|
|
|
|
|
2341
|
|
|
|
|
|
|
|
2342
|
|
|
|
|
|
|
|
2343
|
|
|
|
|
|
|
If all elements of C<$x> are equal, |
2344
|
|
|
|
|
|
|
|
2345
|
|
|
|
|
|
|
i = $x->nelem - 1 |
2346
|
|
|
|
|
|
|
|
2347
|
|
|
|
|
|
|
If C<$x> contains duplicated elements, I is the index of the |
2348
|
|
|
|
|
|
|
righmost (by index in array) duplicate if I matches. |
2349
|
|
|
|
|
|
|
|
2350
|
|
|
|
|
|
|
|
2351
|
|
|
|
|
|
|
|
2352
|
|
|
|
|
|
|
|
2353
|
|
|
|
|
|
|
|
2354
|
|
|
|
|
|
|
|
2355
|
|
|
|
|
|
|
|
2356
|
|
|
|
|
|
|
=for bad |
2357
|
|
|
|
|
|
|
|
2358
|
|
|
|
|
|
|
needs major (?) work to handles bad values |
2359
|
|
|
|
|
|
|
|
2360
|
|
|
|
|
|
|
=cut |
2361
|
|
|
|
|
|
|
|
2362
|
|
|
|
|
|
|
|
2363
|
|
|
|
|
|
|
|
2364
|
|
|
|
|
|
|
|
2365
|
|
|
|
|
|
|
|
2366
|
|
|
|
|
|
|
|
2367
|
|
|
|
|
|
|
*vsearch_bin_inclusive = \&PDLA::vsearch_bin_inclusive; |
2368
|
|
|
|
|
|
|
|
2369
|
|
|
|
|
|
|
|
2370
|
|
|
|
|
|
|
|
2371
|
|
|
|
|
|
|
|
2372
|
|
|
|
|
|
|
|
2373
|
|
|
|
|
|
|
=head2 vsearch_bin_exclusive |
2374
|
|
|
|
|
|
|
|
2375
|
|
|
|
|
|
|
=for sig |
2376
|
|
|
|
|
|
|
|
2377
|
|
|
|
|
|
|
Signature: (vals(); x(n); indx [o]idx()) |
2378
|
|
|
|
|
|
|
|
2379
|
|
|
|
|
|
|
|
2380
|
|
|
|
|
|
|
=for ref |
2381
|
|
|
|
|
|
|
|
2382
|
|
|
|
|
|
|
Determine the index for values in a sorted array of bins, lower bound exclusive. |
2383
|
|
|
|
|
|
|
|
2384
|
|
|
|
|
|
|
=for usage |
2385
|
|
|
|
|
|
|
|
2386
|
|
|
|
|
|
|
$idx = vsearch_bin_exclusive($vals, $x); |
2387
|
|
|
|
|
|
|
|
2388
|
|
|
|
|
|
|
C<$x> must be sorted, but may be in decreasing or increasing |
2389
|
|
|
|
|
|
|
order. |
2390
|
|
|
|
|
|
|
|
2391
|
|
|
|
|
|
|
|
2392
|
|
|
|
|
|
|
|
2393
|
|
|
|
|
|
|
C<$x> represents the edges of contiguous bins, with the first and |
2394
|
|
|
|
|
|
|
last elements representing the outer edges of the outer bins, and the |
2395
|
|
|
|
|
|
|
inner elements the shared bin edges. |
2396
|
|
|
|
|
|
|
|
2397
|
|
|
|
|
|
|
The lower bound of a bin is exclusive to the bin, its upper bound is inclusive to it. |
2398
|
|
|
|
|
|
|
B returns an index I for each value I of C<$vals>. |
2399
|
|
|
|
|
|
|
|
2400
|
|
|
|
|
|
|
|
2401
|
|
|
|
|
|
|
|
2402
|
|
|
|
|
|
|
|
2403
|
|
|
|
|
|
|
|
2404
|
|
|
|
|
|
|
I has the following properties: |
2405
|
|
|
|
|
|
|
|
2406
|
|
|
|
|
|
|
=over |
2407
|
|
|
|
|
|
|
|
2408
|
|
|
|
|
|
|
=item * |
2409
|
|
|
|
|
|
|
|
2410
|
|
|
|
|
|
|
if C<$x> is sorted in increasing order |
2411
|
|
|
|
|
|
|
|
2412
|
|
|
|
|
|
|
|
2413
|
|
|
|
|
|
|
V <= x[0] : I = -1 |
2414
|
|
|
|
|
|
|
x[0] < V <= x[-1] : I s.t. x[I] < V <= x[I+1] |
2415
|
|
|
|
|
|
|
x[-1] < V : I = $x->nelem - 1 |
2416
|
|
|
|
|
|
|
|
2417
|
|
|
|
|
|
|
|
2418
|
|
|
|
|
|
|
|
2419
|
|
|
|
|
|
|
=item * |
2420
|
|
|
|
|
|
|
|
2421
|
|
|
|
|
|
|
if C<$x> is sorted in decreasing order |
2422
|
|
|
|
|
|
|
|
2423
|
|
|
|
|
|
|
|
2424
|
|
|
|
|
|
|
V > x[0] : I = 0 |
2425
|
|
|
|
|
|
|
x[0] >= V > x[-1] : I s.t. x[I-1] >= V > x[I] |
2426
|
|
|
|
|
|
|
x[-1] >= V : I = $x->nelem |
2427
|
|
|
|
|
|
|
|
2428
|
|
|
|
|
|
|
|
2429
|
|
|
|
|
|
|
|
2430
|
|
|
|
|
|
|
=back |
2431
|
|
|
|
|
|
|
|
2432
|
|
|
|
|
|
|
|
2433
|
|
|
|
|
|
|
|
2434
|
|
|
|
|
|
|
|
2435
|
|
|
|
|
|
|
If all elements of C<$x> are equal, |
2436
|
|
|
|
|
|
|
|
2437
|
|
|
|
|
|
|
i = $x->nelem - 1 |
2438
|
|
|
|
|
|
|
|
2439
|
|
|
|
|
|
|
If C<$x> contains duplicated elements, I is the index of the |
2440
|
|
|
|
|
|
|
righmost (by index in array) duplicate if I matches. |
2441
|
|
|
|
|
|
|
|
2442
|
|
|
|
|
|
|
|
2443
|
|
|
|
|
|
|
|
2444
|
|
|
|
|
|
|
|
2445
|
|
|
|
|
|
|
|
2446
|
|
|
|
|
|
|
|
2447
|
|
|
|
|
|
|
|
2448
|
|
|
|
|
|
|
=for bad |
2449
|
|
|
|
|
|
|
|
2450
|
|
|
|
|
|
|
needs major (?) work to handles bad values |
2451
|
|
|
|
|
|
|
|
2452
|
|
|
|
|
|
|
=cut |
2453
|
|
|
|
|
|
|
|
2454
|
|
|
|
|
|
|
|
2455
|
|
|
|
|
|
|
|
2456
|
|
|
|
|
|
|
|
2457
|
|
|
|
|
|
|
|
2458
|
|
|
|
|
|
|
|
2459
|
|
|
|
|
|
|
*vsearch_bin_exclusive = \&PDLA::vsearch_bin_exclusive; |
2460
|
|
|
|
|
|
|
|
2461
|
|
|
|
|
|
|
|
2462
|
|
|
|
|
|
|
|
2463
|
|
|
|
|
|
|
|
2464
|
|
|
|
|
|
|
|
2465
|
|
|
|
|
|
|
=head2 interpolate |
2466
|
|
|
|
|
|
|
|
2467
|
|
|
|
|
|
|
=for sig |
2468
|
|
|
|
|
|
|
|
2469
|
|
|
|
|
|
|
Signature: (xi(); x(n); y(n); [o] yi(); int [o] err()) |
2470
|
|
|
|
|
|
|
|
2471
|
|
|
|
|
|
|
|
2472
|
|
|
|
|
|
|
=for ref |
2473
|
|
|
|
|
|
|
|
2474
|
|
|
|
|
|
|
routine for 1D linear interpolation |
2475
|
|
|
|
|
|
|
|
2476
|
|
|
|
|
|
|
=for usage |
2477
|
|
|
|
|
|
|
|
2478
|
|
|
|
|
|
|
( $yi, $err ) = interpolate($xi, $x, $y) |
2479
|
|
|
|
|
|
|
|
2480
|
|
|
|
|
|
|
Given a set of points C<($x,$y)>, use linear interpolation |
2481
|
|
|
|
|
|
|
to find the values C<$yi> at a set of points C<$xi>. |
2482
|
|
|
|
|
|
|
|
2483
|
|
|
|
|
|
|
C uses a binary search to find the suspects, er..., |
2484
|
|
|
|
|
|
|
interpolation indices and therefore abscissas (ie C<$x>) |
2485
|
|
|
|
|
|
|
have to be I ordered (increasing or decreasing). |
2486
|
|
|
|
|
|
|
For interpolation at lots of |
2487
|
|
|
|
|
|
|
closely spaced abscissas an approach that uses the last index found as |
2488
|
|
|
|
|
|
|
a start for the next search can be faster (compare Numerical Recipes |
2489
|
|
|
|
|
|
|
C routine). Feel free to implement that on top of the binary |
2490
|
|
|
|
|
|
|
search if you like. For out of bounds values it just does a linear |
2491
|
|
|
|
|
|
|
extrapolation and sets the corresponding element of C<$err> to 1, |
2492
|
|
|
|
|
|
|
which is otherwise 0. |
2493
|
|
|
|
|
|
|
|
2494
|
|
|
|
|
|
|
See also L, which uses the same routine, |
2495
|
|
|
|
|
|
|
differing only in the handling of extrapolation - an error message |
2496
|
|
|
|
|
|
|
is printed rather than returning an error piddle. |
2497
|
|
|
|
|
|
|
|
2498
|
|
|
|
|
|
|
|
2499
|
|
|
|
|
|
|
|
2500
|
|
|
|
|
|
|
=for bad |
2501
|
|
|
|
|
|
|
|
2502
|
|
|
|
|
|
|
needs major (?) work to handles bad values |
2503
|
|
|
|
|
|
|
|
2504
|
|
|
|
|
|
|
=cut |
2505
|
|
|
|
|
|
|
|
2506
|
|
|
|
|
|
|
|
2507
|
|
|
|
|
|
|
|
2508
|
|
|
|
|
|
|
|
2509
|
|
|
|
|
|
|
|
2510
|
|
|
|
|
|
|
|
2511
|
|
|
|
|
|
|
*interpolate = \&PDLA::interpolate; |
2512
|
|
|
|
|
|
|
|
2513
|
|
|
|
|
|
|
|
2514
|
|
|
|
|
|
|
|
2515
|
|
|
|
|
|
|
|
2516
|
|
|
|
|
|
|
=head2 interpol |
2517
|
|
|
|
|
|
|
|
2518
|
|
|
|
|
|
|
=for sig |
2519
|
|
|
|
|
|
|
|
2520
|
|
|
|
|
|
|
Signature: (xi(); x(n); y(n); [o] yi()) |
2521
|
|
|
|
|
|
|
|
2522
|
|
|
|
|
|
|
=for ref |
2523
|
|
|
|
|
|
|
|
2524
|
|
|
|
|
|
|
routine for 1D linear interpolation |
2525
|
|
|
|
|
|
|
|
2526
|
|
|
|
|
|
|
=for usage |
2527
|
|
|
|
|
|
|
|
2528
|
|
|
|
|
|
|
$yi = interpol($xi, $x, $y) |
2529
|
|
|
|
|
|
|
|
2530
|
|
|
|
|
|
|
C uses the same search method as L, |
2531
|
|
|
|
|
|
|
hence C<$x> must be I ordered (either increasing or decreasing). |
2532
|
|
|
|
|
|
|
The difference occurs in the handling of out-of-bounds values; here |
2533
|
|
|
|
|
|
|
an error message is printed. |
2534
|
|
|
|
|
|
|
|
2535
|
|
|
|
|
|
|
=cut |
2536
|
|
|
|
|
|
|
|
2537
|
|
|
|
|
|
|
# kept in for backwards compatability |
2538
|
|
|
|
|
|
|
sub interpol ($$$;$) { |
2539
|
1
|
|
|
1
|
1
|
7
|
my $xi = shift; |
2540
|
1
|
|
|
|
|
3
|
my $x = shift; |
2541
|
1
|
|
|
|
|
2
|
my $y = shift; |
2542
|
|
|
|
|
|
|
|
2543
|
1
|
|
|
|
|
2
|
my $yi; |
2544
|
1
|
50
|
|
|
|
4
|
if ( $#_ == 0 ) { $yi = $_[0]; } |
|
0
|
|
|
|
|
0
|
|
2545
|
1
|
|
|
|
|
5
|
else { $yi = PDLA->null; } |
2546
|
|
|
|
|
|
|
|
2547
|
1
|
|
|
|
|
5
|
interpolate( $xi, $x, $y, $yi, my $err = PDLA->null ); |
2548
|
1
|
50
|
|
|
|
8
|
print "some values had to be extrapolated\n" |
2549
|
|
|
|
|
|
|
if any $err; |
2550
|
|
|
|
|
|
|
|
2551
|
1
|
50
|
|
|
|
9
|
return $yi if $#_ == -1; |
2552
|
|
|
|
|
|
|
|
2553
|
|
|
|
|
|
|
} # sub: interpol() |
2554
|
|
|
|
|
|
|
*PDLA::interpol = \&interpol; |
2555
|
|
|
|
|
|
|
|
2556
|
|
|
|
|
|
|
|
2557
|
|
|
|
|
|
|
|
2558
|
|
|
|
|
|
|
|
2559
|
|
|
|
|
|
|
=head2 interpND |
2560
|
|
|
|
|
|
|
|
2561
|
|
|
|
|
|
|
=for ref |
2562
|
|
|
|
|
|
|
|
2563
|
|
|
|
|
|
|
Interpolate values from an N-D piddle, with switchable method |
2564
|
|
|
|
|
|
|
|
2565
|
|
|
|
|
|
|
=for example |
2566
|
|
|
|
|
|
|
|
2567
|
|
|
|
|
|
|
$source = 10*xvals(10,10) + yvals(10,10); |
2568
|
|
|
|
|
|
|
$index = pdl([[2.2,3.5],[4.1,5.0]],[[6.0,7.4],[8,9]]); |
2569
|
|
|
|
|
|
|
print $source->interpND( $index ); |
2570
|
|
|
|
|
|
|
|
2571
|
|
|
|
|
|
|
InterpND acts like L, |
2572
|
|
|
|
|
|
|
collapsing C<$index> by lookup |
2573
|
|
|
|
|
|
|
into C<$source>; but it does interpolation rather than direct sampling. |
2574
|
|
|
|
|
|
|
The interpolation method and boundary condition are switchable via |
2575
|
|
|
|
|
|
|
an options hash. |
2576
|
|
|
|
|
|
|
|
2577
|
|
|
|
|
|
|
By default, linear or sample interpolation is used, with constant |
2578
|
|
|
|
|
|
|
value outside the boundaries of the source pdl. No dataflow occurs, |
2579
|
|
|
|
|
|
|
because in general the output is computed rather than indexed. |
2580
|
|
|
|
|
|
|
|
2581
|
|
|
|
|
|
|
All the interpolation methods treat the pixels as value-centered, so |
2582
|
|
|
|
|
|
|
the C method will return C<< $a->(0) >> for coordinate values on |
2583
|
|
|
|
|
|
|
the set [-0.5,0.5), and all methods will return C<< $a->(1) >> for |
2584
|
|
|
|
|
|
|
a coordinate value of exactly 1. |
2585
|
|
|
|
|
|
|
|
2586
|
|
|
|
|
|
|
|
2587
|
|
|
|
|
|
|
Recognized options: |
2588
|
|
|
|
|
|
|
|
2589
|
|
|
|
|
|
|
=over 3 |
2590
|
|
|
|
|
|
|
|
2591
|
|
|
|
|
|
|
=item method |
2592
|
|
|
|
|
|
|
|
2593
|
|
|
|
|
|
|
Values can be: |
2594
|
|
|
|
|
|
|
|
2595
|
|
|
|
|
|
|
=over 3 |
2596
|
|
|
|
|
|
|
|
2597
|
|
|
|
|
|
|
=item * 0, s, sample, Sample (default for integer source types) |
2598
|
|
|
|
|
|
|
|
2599
|
|
|
|
|
|
|
The nearest value is taken. Pixels are regarded as centered on their |
2600
|
|
|
|
|
|
|
respective integer coordinates (no offset from the linear case). |
2601
|
|
|
|
|
|
|
|
2602
|
|
|
|
|
|
|
=item * 1, l, linear, Linear (default for floating point source types) |
2603
|
|
|
|
|
|
|
|
2604
|
|
|
|
|
|
|
The values are N-linearly interpolated from an N-dimensional cube of size 2. |
2605
|
|
|
|
|
|
|
|
2606
|
|
|
|
|
|
|
=item * 3, c, cube, cubic, Cubic |
2607
|
|
|
|
|
|
|
|
2608
|
|
|
|
|
|
|
The values are interpolated using a local cubic fit to the data. The |
2609
|
|
|
|
|
|
|
fit is constrained to match the original data and its derivative at the |
2610
|
|
|
|
|
|
|
data points. The second derivative of the fit is not continuous at the |
2611
|
|
|
|
|
|
|
data points. Multidimensional datasets are interpolated by the |
2612
|
|
|
|
|
|
|
successive-collapse method. |
2613
|
|
|
|
|
|
|
|
2614
|
|
|
|
|
|
|
(Note that the constraint on the first derivative causes a small amount |
2615
|
|
|
|
|
|
|
of ringing around sudden features such as step functions). |
2616
|
|
|
|
|
|
|
|
2617
|
|
|
|
|
|
|
=item * f, fft, fourier, Fourier |
2618
|
|
|
|
|
|
|
|
2619
|
|
|
|
|
|
|
The source is Fourier transformed, and the interpolated values are |
2620
|
|
|
|
|
|
|
explicitly calculated from the coefficients. The boundary condition |
2621
|
|
|
|
|
|
|
option is ignored -- periodic boundaries are imposed. |
2622
|
|
|
|
|
|
|
|
2623
|
|
|
|
|
|
|
If you pass in the option "fft", and it is a list (ARRAY) ref, then it |
2624
|
|
|
|
|
|
|
is a stash for the magnitude and phase of the source FFT. If the list |
2625
|
|
|
|
|
|
|
has two elements then they are taken as already computed; otherwise |
2626
|
|
|
|
|
|
|
they are calculated and put in the stash. |
2627
|
|
|
|
|
|
|
|
2628
|
|
|
|
|
|
|
=back |
2629
|
|
|
|
|
|
|
|
2630
|
|
|
|
|
|
|
=item b, bound, boundary, Boundary |
2631
|
|
|
|
|
|
|
|
2632
|
|
|
|
|
|
|
This option is passed unmodified into L, |
2633
|
|
|
|
|
|
|
which is used as the indexing engine for the interpolation. |
2634
|
|
|
|
|
|
|
Some current allowed values are 'extend', 'periodic', 'truncate', and 'mirror' |
2635
|
|
|
|
|
|
|
(default is 'truncate'). |
2636
|
|
|
|
|
|
|
|
2637
|
|
|
|
|
|
|
=item bad |
2638
|
|
|
|
|
|
|
|
2639
|
|
|
|
|
|
|
contains the fill value used for 'truncate' boundary. (default 0) |
2640
|
|
|
|
|
|
|
|
2641
|
|
|
|
|
|
|
=item fft |
2642
|
|
|
|
|
|
|
|
2643
|
|
|
|
|
|
|
An array ref whose associated list is used to stash the FFT of the source |
2644
|
|
|
|
|
|
|
data, for the FFT method. |
2645
|
|
|
|
|
|
|
|
2646
|
|
|
|
|
|
|
=back |
2647
|
|
|
|
|
|
|
|
2648
|
|
|
|
|
|
|
=cut |
2649
|
|
|
|
|
|
|
|
2650
|
|
|
|
|
|
|
*interpND = *PDLA::interpND; |
2651
|
|
|
|
|
|
|
sub PDLA::interpND { |
2652
|
1
|
|
|
1
|
0
|
105
|
my $source = shift; |
2653
|
1
|
|
|
|
|
2
|
my $index = shift; |
2654
|
1
|
|
|
|
|
2
|
my $options = shift; |
2655
|
|
|
|
|
|
|
|
2656
|
1
|
50
|
33
|
|
|
6
|
barf 'Usage: interp_nd($source,$index,[{%options}])\n' |
2657
|
|
|
|
|
|
|
if(defined $options and ref $options ne 'HASH'); |
2658
|
|
|
|
|
|
|
|
2659
|
1
|
50
|
|
|
|
5
|
my($opt) = (defined $options) ? $options : {}; |
2660
|
|
|
|
|
|
|
|
2661
|
1
|
|
33
|
|
|
14
|
my($method) = $opt->{m} || $opt->{meth} || $opt->{method} || $opt->{Method}; |
2662
|
1
|
50
|
|
|
|
5
|
if(!defined $method) { |
2663
|
1
|
50
|
|
|
|
4
|
$method = ($source->type <= zeroes(long,1)->type) ? |
2664
|
|
|
|
|
|
|
'sample' : |
2665
|
|
|
|
|
|
|
'linear'; |
2666
|
|
|
|
|
|
|
} |
2667
|
|
|
|
|
|
|
|
2668
|
1
|
|
50
|
|
|
28
|
my($boundary) = $opt->{b} || $opt->{boundary} || $opt->{Boundary} || $opt->{bound} || $opt->{Bound} || 'extend'; |
2669
|
1
|
|
50
|
|
|
7
|
my($bad) = $opt->{bad} || $opt->{Bad} || 0.0; |
2670
|
|
|
|
|
|
|
|
2671
|
1
|
50
|
33
|
|
|
14
|
if($method =~ m/^s(am(p(le)?)?)?/i) { |
|
|
50
|
0
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
2672
|
0
|
|
|
|
|
0
|
return $source->range(PDLA::Math::floor($index+0.5),0,$boundary); |
2673
|
|
|
|
|
|
|
} |
2674
|
|
|
|
|
|
|
|
2675
|
|
|
|
|
|
|
elsif (($method eq 1) || $method =~ m/^l(in(ear)?)?/i) { |
2676
|
|
|
|
|
|
|
## key: (ith = index thread; cth = cube thread; sth = source thread) |
2677
|
1
|
|
|
|
|
7
|
my $d = $index->dim(0); |
2678
|
1
|
|
|
|
|
5
|
my $di = $index->ndims - 1; |
2679
|
|
|
|
|
|
|
|
2680
|
|
|
|
|
|
|
# Grab a 2-on-a-side n-cube around each desired pixel |
2681
|
1
|
|
|
|
|
77
|
my $samp = $source->range($index->floor,2,$boundary); # (ith, cth, sth) |
2682
|
|
|
|
|
|
|
|
2683
|
|
|
|
|
|
|
# Reorder to put the cube dimensions in front and convert to a list |
2684
|
1
|
|
|
|
|
31
|
$samp = $samp->reorder( $di .. $di+$d-1, |
2685
|
|
|
|
|
|
|
0 .. $di-1, |
2686
|
|
|
|
|
|
|
$di+$d .. $samp->ndims-1) # (cth, ith, sth) |
2687
|
|
|
|
|
|
|
->clump($d); # (clst, ith, sth) |
2688
|
|
|
|
|
|
|
|
2689
|
|
|
|
|
|
|
# Enumerate the corners of an n-cube and convert to a list |
2690
|
|
|
|
|
|
|
# (the 'x' is the normal perl repeat operator) |
2691
|
1
|
|
|
|
|
9
|
my $crnr = PDLA::Basic::ndcoords( (2) x $index->dim(0) ) # (index,cth) |
2692
|
|
|
|
|
|
|
->mv(0,-1)->clump($index->dim(0))->mv(-1,0); # (index, clst) |
2693
|
|
|
|
|
|
|
|
2694
|
|
|
|
|
|
|
# a & b are the weighting coefficients. |
2695
|
1
|
|
|
|
|
6
|
my($x,$y); |
2696
|
1
|
|
|
|
|
0
|
my($indexwhere); |
2697
|
1
|
|
|
|
|
19
|
($indexwhere = $index->where( 0 * $index )) .= -10; # Change NaN to invalid |
2698
|
|
|
|
|
|
|
{ |
2699
|
1
|
|
|
|
|
6
|
my $bb = PDLA::Math::floor($index); |
|
1
|
|
|
|
|
14
|
|
2700
|
1
|
|
|
|
|
24
|
$x = ($index - $bb) -> dummy(1,$crnr->dim(1)); # index, clst, ith |
2701
|
1
|
|
|
|
|
41
|
$y = ($bb + 1 - $index) -> dummy(1,$crnr->dim(1)); # index, clst, ith |
2702
|
|
|
|
|
|
|
} |
2703
|
|
|
|
|
|
|
|
2704
|
|
|
|
|
|
|
# Use 1/0 corners to select which multiplier happens, multiply |
2705
|
|
|
|
|
|
|
# 'em all together to get sample weights, and sum to get the answer. |
2706
|
1
|
|
|
|
|
109
|
my $out0 = ( ($x * ($crnr==1) + $y * ($crnr==0)) #index, clst, ith |
2707
|
|
|
|
|
|
|
-> prodover #clst, ith |
2708
|
|
|
|
|
|
|
); |
2709
|
|
|
|
|
|
|
|
2710
|
1
|
|
|
|
|
68
|
my $out = ($out0 * $samp)->sumover; # ith, sth |
2711
|
|
|
|
|
|
|
|
2712
|
|
|
|
|
|
|
# Work around BAD-not-being-contagious bug in PDLA <= 2.6 bad handling code --CED 3-April-2013 |
2713
|
1
|
50
|
33
|
|
|
13
|
if($PDLA::Bad::Status and $source->badflag) { |
2714
|
0
|
|
|
|
|
0
|
my $baddies = $samp->isbad->orover; |
2715
|
0
|
|
|
|
|
0
|
$out = $out->setbadif($baddies); |
2716
|
|
|
|
|
|
|
} |
2717
|
|
|
|
|
|
|
|
2718
|
1
|
|
|
|
|
41
|
return $out; |
2719
|
|
|
|
|
|
|
|
2720
|
|
|
|
|
|
|
} elsif(($method eq 3) || $method =~ m/^c(u(b(e|ic)?)?)?/i) { |
2721
|
|
|
|
|
|
|
|
2722
|
0
|
|
|
|
|
0
|
my ($d,@di) = $index->dims; |
2723
|
0
|
|
|
|
|
0
|
my $di = $index->ndims - 1; |
2724
|
|
|
|
|
|
|
|
2725
|
|
|
|
|
|
|
# Grab a 4-on-a-side n-cube around each desired pixel |
2726
|
0
|
|
|
|
|
0
|
my $samp = $source->range($index->floor - 1,4,$boundary) #ith, cth, sth |
2727
|
|
|
|
|
|
|
->reorder( $di .. $di+$d-1, 0..$di-1, $di+$d .. $source->ndims-1 ); |
2728
|
|
|
|
|
|
|
# (cth, ith, sth) |
2729
|
|
|
|
|
|
|
|
2730
|
|
|
|
|
|
|
# Make a cube of the subpixel offsets, and expand its dims to |
2731
|
|
|
|
|
|
|
# a 4-on-a-side N-1 cube, to match the slices of $samp (used below). |
2732
|
0
|
|
|
|
|
0
|
my $y = $index - $index->floor; |
2733
|
0
|
|
|
|
|
0
|
for my $i(1..$d-1) { |
2734
|
0
|
|
|
|
|
0
|
$y = $y->dummy($i,4); |
2735
|
|
|
|
|
|
|
} |
2736
|
|
|
|
|
|
|
|
2737
|
|
|
|
|
|
|
# Collapse by interpolation, one dimension at a time... |
2738
|
0
|
|
|
|
|
0
|
for my $i(0..$d-1) { |
2739
|
0
|
|
|
|
|
0
|
my $a0 = $samp->slice("(1)"); # Just-under-sample |
2740
|
0
|
|
|
|
|
0
|
my $a1 = $samp->slice("(2)"); # Just-over-sample |
2741
|
0
|
|
|
|
|
0
|
my $a1a0 = $a1 - $a0; |
2742
|
|
|
|
|
|
|
|
2743
|
0
|
|
|
|
|
0
|
my $gradient = 0.5 * ($samp->slice("2:3")-$samp->slice("0:1")); |
2744
|
0
|
|
|
|
|
0
|
my $s0 = $gradient->slice("(0)"); # Just-under-gradient |
2745
|
0
|
|
|
|
|
0
|
my $s1 = $gradient->slice("(1)"); # Just-over-gradient |
2746
|
|
|
|
|
|
|
|
2747
|
0
|
|
|
|
|
0
|
$bb = $y->slice("($i)"); |
2748
|
|
|
|
|
|
|
|
2749
|
|
|
|
|
|
|
# Collapse the sample... |
2750
|
0
|
|
|
|
|
0
|
$samp = ( $a0 + |
2751
|
|
|
|
|
|
|
$bb * ( |
2752
|
|
|
|
|
|
|
$s0 + |
2753
|
|
|
|
|
|
|
$bb * ( (3 * $a1a0 - 2*$s0 - $s1) + |
2754
|
|
|
|
|
|
|
$bb * ( $s1 + $s0 - 2*$a1a0 ) |
2755
|
|
|
|
|
|
|
) |
2756
|
|
|
|
|
|
|
) |
2757
|
|
|
|
|
|
|
); |
2758
|
|
|
|
|
|
|
|
2759
|
|
|
|
|
|
|
# "Collapse" the subpixel offset... |
2760
|
0
|
|
|
|
|
0
|
$y = $y->slice(":,($i)"); |
2761
|
|
|
|
|
|
|
} |
2762
|
|
|
|
|
|
|
|
2763
|
0
|
|
|
|
|
0
|
return $samp; |
2764
|
|
|
|
|
|
|
|
2765
|
|
|
|
|
|
|
} elsif($method =~ m/^f(ft|ourier)?/i) { |
2766
|
|
|
|
|
|
|
|
2767
|
0
|
|
|
|
|
0
|
eval "use PDLA::FFT;"; |
2768
|
0
|
|
|
|
|
0
|
my $fftref = $opt->{fft}; |
2769
|
0
|
0
|
|
|
|
0
|
$fftref = [] unless(ref $fftref eq 'ARRAY'); |
2770
|
0
|
0
|
|
|
|
0
|
if(@$fftref != 2) { |
2771
|
0
|
|
|
|
|
0
|
my $x = $source->copy; |
2772
|
0
|
|
|
|
|
0
|
my $y = zeroes($source); |
2773
|
0
|
|
|
|
|
0
|
fftnd($x,$y); |
2774
|
0
|
|
|
|
|
0
|
$fftref->[0] = sqrt($x*$x+$y*$y) / $x->nelem; |
2775
|
0
|
|
|
|
|
0
|
$fftref->[1] = - atan2($y,$x); |
2776
|
|
|
|
|
|
|
} |
2777
|
|
|
|
|
|
|
|
2778
|
0
|
|
|
|
|
0
|
my $i; |
2779
|
0
|
|
|
|
|
0
|
my $c = PDLA::Basic::ndcoords($source); # (dim, source-dims) |
2780
|
0
|
|
|
|
|
0
|
for $i(1..$index->ndims-1) { |
2781
|
0
|
|
|
|
|
0
|
$c = $c->dummy($i,$index->dim($i)) |
2782
|
|
|
|
|
|
|
} |
2783
|
0
|
|
|
|
|
0
|
my $id = $index->ndims-1; |
2784
|
0
|
|
|
|
|
0
|
my $phase = (($c * $index * 3.14159 * 2 / pdl($source->dims)) |
2785
|
|
|
|
|
|
|
->sumover) # (index-dims, source-dims) |
2786
|
|
|
|
|
|
|
->reorder($id..$id+$source->ndims-1,0..$id-1); # (src, index) |
2787
|
|
|
|
|
|
|
|
2788
|
0
|
|
|
|
|
0
|
my $phref = $fftref->[1]->copy; # (source-dims) |
2789
|
0
|
|
|
|
|
0
|
my $mag = $fftref->[0]->copy; # (source-dims) |
2790
|
|
|
|
|
|
|
|
2791
|
0
|
|
|
|
|
0
|
for $i(1..$index->ndims-1) { |
2792
|
0
|
|
|
|
|
0
|
$phref = $phref->dummy(-1,$index->dim($i)); |
2793
|
0
|
|
|
|
|
0
|
$mag = $mag->dummy(-1,$index->dim($i)); |
2794
|
|
|
|
|
|
|
} |
2795
|
0
|
|
|
|
|
0
|
my $out = cos($phase + $phref ) * $mag; |
2796
|
0
|
|
|
|
|
0
|
$out = $out->clump($source->ndims)->sumover; |
2797
|
|
|
|
|
|
|
|
2798
|
0
|
|
|
|
|
0
|
return $out; |
2799
|
|
|
|
|
|
|
} else { |
2800
|
0
|
|
|
|
|
0
|
barf("interpND: unknown method '$method'; valid ones are 'linear' and 'sample'.\n"); |
2801
|
|
|
|
|
|
|
} |
2802
|
|
|
|
|
|
|
} |
2803
|
|
|
|
|
|
|
|
2804
|
|
|
|
|
|
|
|
2805
|
|
|
|
|
|
|
|
2806
|
|
|
|
|
|
|
|
2807
|
|
|
|
|
|
|
=head2 one2nd |
2808
|
|
|
|
|
|
|
|
2809
|
|
|
|
|
|
|
=for ref |
2810
|
|
|
|
|
|
|
|
2811
|
|
|
|
|
|
|
Converts a one dimensional index piddle to a set of ND coordinates |
2812
|
|
|
|
|
|
|
|
2813
|
|
|
|
|
|
|
=for usage |
2814
|
|
|
|
|
|
|
|
2815
|
|
|
|
|
|
|
@coords=one2nd($x, $indices) |
2816
|
|
|
|
|
|
|
|
2817
|
|
|
|
|
|
|
returns an array of piddles containing the ND indexes corresponding to |
2818
|
|
|
|
|
|
|
the one dimensional list indices. The indices are assumed to |
2819
|
|
|
|
|
|
|
correspond to array C<$x> clumped using C. This routine is |
2820
|
|
|
|
|
|
|
used in the old vector form of L, but is useful on |
2821
|
|
|
|
|
|
|
its own occasionally. |
2822
|
|
|
|
|
|
|
|
2823
|
|
|
|
|
|
|
Returned piddles have the L datatype. C<$indices> can have |
2824
|
|
|
|
|
|
|
values larger than C<< $x->nelem >> but negative values in C<$indices> |
2825
|
|
|
|
|
|
|
will not give the answer you expect. |
2826
|
|
|
|
|
|
|
|
2827
|
|
|
|
|
|
|
=for example |
2828
|
|
|
|
|
|
|
|
2829
|
|
|
|
|
|
|
pdla> $x=pdl [[[1,2],[-1,1]], [[0,-3],[3,2]]]; $c=$x->clump(-1) |
2830
|
|
|
|
|
|
|
pdla> $maxind=maximum_ind($c); p $maxind; |
2831
|
|
|
|
|
|
|
6 |
2832
|
|
|
|
|
|
|
pdla> print one2nd($x, maximum_ind($c)) |
2833
|
|
|
|
|
|
|
0 1 1 |
2834
|
|
|
|
|
|
|
pdla> p $x->at(0,1,1) |
2835
|
|
|
|
|
|
|
3 |
2836
|
|
|
|
|
|
|
|
2837
|
|
|
|
|
|
|
=cut |
2838
|
|
|
|
|
|
|
|
2839
|
|
|
|
|
|
|
*one2nd = \&PDLA::one2nd; |
2840
|
|
|
|
|
|
|
sub PDLA::one2nd { |
2841
|
1
|
50
|
|
1
|
0
|
8
|
barf "Usage: one2nd \$array \$indices\n" if $#_ != 1; |
2842
|
1
|
|
|
|
|
2
|
my ($x, $ind)=@_; |
2843
|
1
|
|
|
|
|
4
|
my @dimension=$x->dims; |
2844
|
1
|
|
|
|
|
4
|
$ind = indx($ind); |
2845
|
1
|
|
|
|
|
3
|
my(@index); |
2846
|
1
|
|
|
|
|
3
|
my $count=0; |
2847
|
1
|
|
|
|
|
3
|
foreach (@dimension) { |
2848
|
3
|
|
|
|
|
40
|
$index[$count++]=$ind % $_; |
2849
|
3
|
|
|
|
|
20
|
$ind /= $_; |
2850
|
|
|
|
|
|
|
} |
2851
|
1
|
|
|
|
|
9
|
return @index; |
2852
|
|
|
|
|
|
|
} |
2853
|
|
|
|
|
|
|
|
2854
|
|
|
|
|
|
|
|
2855
|
|
|
|
|
|
|
|
2856
|
|
|
|
|
|
|
|
2857
|
|
|
|
|
|
|
|
2858
|
|
|
|
|
|
|
=head2 which |
2859
|
|
|
|
|
|
|
|
2860
|
|
|
|
|
|
|
=for sig |
2861
|
|
|
|
|
|
|
|
2862
|
|
|
|
|
|
|
Signature: (mask(n); indx [o] inds(m)) |
2863
|
|
|
|
|
|
|
|
2864
|
|
|
|
|
|
|
|
2865
|
|
|
|
|
|
|
=for ref |
2866
|
|
|
|
|
|
|
|
2867
|
|
|
|
|
|
|
Returns indices of non-zero values from a 1-D PDLA |
2868
|
|
|
|
|
|
|
|
2869
|
|
|
|
|
|
|
=for usage |
2870
|
|
|
|
|
|
|
|
2871
|
|
|
|
|
|
|
$i = which($mask); |
2872
|
|
|
|
|
|
|
|
2873
|
|
|
|
|
|
|
returns a pdl with indices for all those elements that are nonzero in |
2874
|
|
|
|
|
|
|
the mask. Note that the returned indices will be 1D. If you feed in a |
2875
|
|
|
|
|
|
|
multidimensional mask, it will be flattened before the indices are |
2876
|
|
|
|
|
|
|
calculated. See also L for multidimensional masks. |
2877
|
|
|
|
|
|
|
|
2878
|
|
|
|
|
|
|
If you want to index into the original mask or a similar piddle |
2879
|
|
|
|
|
|
|
with output from C, remember to flatten it before calling index: |
2880
|
|
|
|
|
|
|
|
2881
|
|
|
|
|
|
|
$data = random 5, 5; |
2882
|
|
|
|
|
|
|
$idx = which $data > 0.5; # $idx is now 1D |
2883
|
|
|
|
|
|
|
$bigsum = $data->flat->index($idx)->sum; # flatten before indexing |
2884
|
|
|
|
|
|
|
|
2885
|
|
|
|
|
|
|
Compare also L for similar functionality. |
2886
|
|
|
|
|
|
|
|
2887
|
|
|
|
|
|
|
SEE ALSO: |
2888
|
|
|
|
|
|
|
|
2889
|
|
|
|
|
|
|
L returns separately the indices of both |
2890
|
|
|
|
|
|
|
zero and nonzero values in the mask. |
2891
|
|
|
|
|
|
|
|
2892
|
|
|
|
|
|
|
L returns associated values from a data PDLA, rather than |
2893
|
|
|
|
|
|
|
indices into the mask PDLA. |
2894
|
|
|
|
|
|
|
|
2895
|
|
|
|
|
|
|
L returns N-D indices into a multidimensional PDLA. |
2896
|
|
|
|
|
|
|
|
2897
|
|
|
|
|
|
|
=for example |
2898
|
|
|
|
|
|
|
|
2899
|
|
|
|
|
|
|
pdla> $x = sequence(10); p $x |
2900
|
|
|
|
|
|
|
[0 1 2 3 4 5 6 7 8 9] |
2901
|
|
|
|
|
|
|
pdla> $indx = which($x>6); p $indx |
2902
|
|
|
|
|
|
|
[7 8 9] |
2903
|
|
|
|
|
|
|
|
2904
|
|
|
|
|
|
|
|
2905
|
|
|
|
|
|
|
|
2906
|
|
|
|
|
|
|
=for bad |
2907
|
|
|
|
|
|
|
|
2908
|
|
|
|
|
|
|
which processes bad values. |
2909
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
2910
|
|
|
|
|
|
|
|
2911
|
|
|
|
|
|
|
|
2912
|
|
|
|
|
|
|
=cut |
2913
|
|
|
|
|
|
|
|
2914
|
|
|
|
|
|
|
|
2915
|
|
|
|
|
|
|
|
2916
|
|
|
|
|
|
|
|
2917
|
167
|
|
|
167
|
1
|
741
|
sub which { my ($this,$out) = @_; |
2918
|
167
|
|
|
|
|
526
|
$this = $this->flat; |
2919
|
167
|
50
|
|
|
|
580
|
$out = $this->nullcreate unless defined $out; |
2920
|
167
|
|
|
|
|
3005
|
PDLA::_which_int($this,$out); |
2921
|
167
|
|
|
|
|
1752
|
return $out; |
2922
|
|
|
|
|
|
|
} |
2923
|
|
|
|
|
|
|
*PDLA::which = \&which; |
2924
|
|
|
|
|
|
|
|
2925
|
|
|
|
|
|
|
|
2926
|
|
|
|
|
|
|
*which = \&PDLA::which; |
2927
|
|
|
|
|
|
|
|
2928
|
|
|
|
|
|
|
|
2929
|
|
|
|
|
|
|
|
2930
|
|
|
|
|
|
|
|
2931
|
|
|
|
|
|
|
|
2932
|
|
|
|
|
|
|
=head2 which_both |
2933
|
|
|
|
|
|
|
|
2934
|
|
|
|
|
|
|
=for sig |
2935
|
|
|
|
|
|
|
|
2936
|
|
|
|
|
|
|
Signature: (mask(n); indx [o] inds(m); indx [o]notinds(q)) |
2937
|
|
|
|
|
|
|
|
2938
|
|
|
|
|
|
|
|
2939
|
|
|
|
|
|
|
=for ref |
2940
|
|
|
|
|
|
|
|
2941
|
|
|
|
|
|
|
Returns indices of zero and nonzero values in a mask PDLA |
2942
|
|
|
|
|
|
|
|
2943
|
|
|
|
|
|
|
=for usage |
2944
|
|
|
|
|
|
|
|
2945
|
|
|
|
|
|
|
($i, $c_i) = which_both($mask); |
2946
|
|
|
|
|
|
|
|
2947
|
|
|
|
|
|
|
This works just as L, but the complement of C<$i> will be in |
2948
|
|
|
|
|
|
|
C<$c_i>. |
2949
|
|
|
|
|
|
|
|
2950
|
|
|
|
|
|
|
=for example |
2951
|
|
|
|
|
|
|
|
2952
|
|
|
|
|
|
|
pdla> $x = sequence(10); p $x |
2953
|
|
|
|
|
|
|
[0 1 2 3 4 5 6 7 8 9] |
2954
|
|
|
|
|
|
|
pdla> ($small, $big) = which_both ($x >= 5); p "$small\n $big" |
2955
|
|
|
|
|
|
|
[5 6 7 8 9] |
2956
|
|
|
|
|
|
|
[0 1 2 3 4] |
2957
|
|
|
|
|
|
|
|
2958
|
|
|
|
|
|
|
|
2959
|
|
|
|
|
|
|
|
2960
|
|
|
|
|
|
|
=for bad |
2961
|
|
|
|
|
|
|
|
2962
|
|
|
|
|
|
|
which_both processes bad values. |
2963
|
|
|
|
|
|
|
It will set the bad-value flag of all output piddles if the flag is set for any of the input piddles. |
2964
|
|
|
|
|
|
|
|
2965
|
|
|
|
|
|
|
|
2966
|
|
|
|
|
|
|
=cut |
2967
|
|
|
|
|
|
|
|
2968
|
|
|
|
|
|
|
|
2969
|
|
|
|
|
|
|
|
2970
|
|
|
|
|
|
|
|
2971
|
1
|
|
|
1
|
1
|
7
|
sub which_both { my ($this,$outi,$outni) = @_; |
2972
|
1
|
|
|
|
|
4
|
$this = $this->flat; |
2973
|
1
|
50
|
|
|
|
6
|
$outi = $this->nullcreate unless defined $outi; |
2974
|
1
|
50
|
|
|
|
6
|
$outni = $this->nullcreate unless defined $outni; |
2975
|
1
|
|
|
|
|
45
|
PDLA::_which_both_int($this,$outi,$outni); |
2976
|
1
|
50
|
|
|
|
8
|
return wantarray ? ($outi,$outni) : $outi; |
2977
|
|
|
|
|
|
|
} |
2978
|
|
|
|
|
|
|
*PDLA::which_both = \&which_both; |
2979
|
|
|
|
|
|
|
|
2980
|
|
|
|
|
|
|
|
2981
|
|
|
|
|
|
|
*which_both = \&PDLA::which_both; |
2982
|
|
|
|
|
|
|
|
2983
|
|
|
|
|
|
|
|
2984
|
|
|
|
|
|
|
|
2985
|
|
|
|
|
|
|
|
2986
|
|
|
|
|
|
|
=head2 where |
2987
|
|
|
|
|
|
|
|
2988
|
|
|
|
|
|
|
=for ref |
2989
|
|
|
|
|
|
|
|
2990
|
|
|
|
|
|
|
Use a mask to select values from one or more data PDLAs |
2991
|
|
|
|
|
|
|
|
2992
|
|
|
|
|
|
|
C accepts one or more data piddles and a mask piddle. It |
2993
|
|
|
|
|
|
|
returns a list of output piddles, corresponding to the input data |
2994
|
|
|
|
|
|
|
piddles. Each output piddle is a 1-dimensional list of values in its |
2995
|
|
|
|
|
|
|
corresponding data piddle. The values are drawn from locations where |
2996
|
|
|
|
|
|
|
the mask is nonzero. |
2997
|
|
|
|
|
|
|
|
2998
|
|
|
|
|
|
|
The output PDLAs are still connected to the original data PDLAs, for the |
2999
|
|
|
|
|
|
|
purpose of dataflow. |
3000
|
|
|
|
|
|
|
|
3001
|
|
|
|
|
|
|
C combines the functionality of L and L |
3002
|
|
|
|
|
|
|
into a single operation. |
3003
|
|
|
|
|
|
|
|
3004
|
|
|
|
|
|
|
BUGS: |
3005
|
|
|
|
|
|
|
|
3006
|
|
|
|
|
|
|
While C works OK for most N-dimensional cases, it does not |
3007
|
|
|
|
|
|
|
thread properly over (for example) the (N+1)th dimension in data |
3008
|
|
|
|
|
|
|
that is compared to an N-dimensional mask. Use C for that. |
3009
|
|
|
|
|
|
|
|
3010
|
|
|
|
|
|
|
=for usage |
3011
|
|
|
|
|
|
|
|
3012
|
|
|
|
|
|
|
$i = $x->where($x+5 > 0); # $i contains those elements of $x |
3013
|
|
|
|
|
|
|
# where mask ($x+5 > 0) is 1 |
3014
|
|
|
|
|
|
|
$i .= -5; # Set those elements (of $x) to -5. Together, these |
3015
|
|
|
|
|
|
|
# commands clamp $x to a maximum of -5. |
3016
|
|
|
|
|
|
|
|
3017
|
|
|
|
|
|
|
It is also possible to use the same mask for several piddles with |
3018
|
|
|
|
|
|
|
the same call: |
3019
|
|
|
|
|
|
|
|
3020
|
|
|
|
|
|
|
($i,$j,$k) = where($x,$y,$z, $x+5>0); |
3021
|
|
|
|
|
|
|
|
3022
|
|
|
|
|
|
|
Note: C<$i> is always 1-D, even if C<$x> is E1-D. |
3023
|
|
|
|
|
|
|
|
3024
|
|
|
|
|
|
|
WARNING: The first argument |
3025
|
|
|
|
|
|
|
(the values) and the second argument (the mask) currently have to have |
3026
|
|
|
|
|
|
|
the exact same dimensions (or horrible things happen). You *cannot* |
3027
|
|
|
|
|
|
|
thread over a smaller mask, for example. |
3028
|
|
|
|
|
|
|
|
3029
|
|
|
|
|
|
|
|
3030
|
|
|
|
|
|
|
=cut |
3031
|
|
|
|
|
|
|
|
3032
|
|
|
|
|
|
|
sub PDLA::where { |
3033
|
105
|
50
|
|
105
|
0
|
1343
|
barf "Usage: where( \$pdl1, ..., \$pdlN, \$mask )\n" if $#_ == 0; |
3034
|
|
|
|
|
|
|
|
3035
|
105
|
100
|
|
|
|
257
|
if($#_ == 1) { |
3036
|
104
|
|
|
|
|
194
|
my($data,$mask) = @_; |
3037
|
104
|
100
|
|
|
|
405
|
$data = $_[0]->clump(-1) if $_[0]->getndims>1; |
3038
|
104
|
100
|
|
|
|
309
|
$mask = $_[1]->clump(-1) if $_[0]->getndims>1; |
3039
|
104
|
|
|
|
|
244
|
return $data->index($mask->which()); |
3040
|
|
|
|
|
|
|
} else { |
3041
|
1
|
50
|
|
|
|
7
|
if($_[-1]->getndims > 1) { |
3042
|
0
|
|
|
|
|
0
|
my $mask = $_[-1]->clump(-1)->which; |
3043
|
0
|
|
|
|
|
0
|
return map {$_->clump(-1)->index($mask)} @_[0..$#_-1]; |
|
0
|
|
|
|
|
0
|
|
3044
|
|
|
|
|
|
|
} else { |
3045
|
1
|
|
|
|
|
4
|
my $mask = $_[-1]->which; |
3046
|
1
|
|
|
|
|
6
|
return map {$_->index($mask)} @_[0..$#_-1]; |
|
2
|
|
|
|
|
38
|
|
3047
|
|
|
|
|
|
|
} |
3048
|
|
|
|
|
|
|
} |
3049
|
|
|
|
|
|
|
} |
3050
|
|
|
|
|
|
|
*where = \&PDLA::where; |
3051
|
|
|
|
|
|
|
|
3052
|
|
|
|
|
|
|
|
3053
|
|
|
|
|
|
|
|
3054
|
|
|
|
|
|
|
|
3055
|
|
|
|
|
|
|
=head2 whereND |
3056
|
|
|
|
|
|
|
|
3057
|
|
|
|
|
|
|
=for ref |
3058
|
|
|
|
|
|
|
|
3059
|
|
|
|
|
|
|
C with support for ND masks and threading |
3060
|
|
|
|
|
|
|
|
3061
|
|
|
|
|
|
|
C accepts one or more data piddles and a |
3062
|
|
|
|
|
|
|
mask piddle. It returns a list of output piddles, |
3063
|
|
|
|
|
|
|
corresponding to the input data piddles. The values |
3064
|
|
|
|
|
|
|
are drawn from locations where the mask is nonzero. |
3065
|
|
|
|
|
|
|
|
3066
|
|
|
|
|
|
|
C differs from C in that the mask |
3067
|
|
|
|
|
|
|
dimensionality is preserved which allows for |
3068
|
|
|
|
|
|
|
proper threading of the selection operation over |
3069
|
|
|
|
|
|
|
higher dimensions. |
3070
|
|
|
|
|
|
|
|
3071
|
|
|
|
|
|
|
As with C the output PDLAs are still connected |
3072
|
|
|
|
|
|
|
to the original data PDLAs, for the purpose of dataflow. |
3073
|
|
|
|
|
|
|
|
3074
|
|
|
|
|
|
|
=for usage |
3075
|
|
|
|
|
|
|
|
3076
|
|
|
|
|
|
|
$sdata = whereND $data, $mask |
3077
|
|
|
|
|
|
|
($s1, $s2, ..., $sn) = whereND $d1, $d2, ..., $dn, $mask |
3078
|
|
|
|
|
|
|
|
3079
|
|
|
|
|
|
|
where |
3080
|
|
|
|
|
|
|
|
3081
|
|
|
|
|
|
|
$data is M dimensional |
3082
|
|
|
|
|
|
|
$mask is N < M dimensional |
3083
|
|
|
|
|
|
|
dims($data) 1..N == dims($mask) 1..N |
3084
|
|
|
|
|
|
|
with threading over N+1 to M dimensions |
3085
|
|
|
|
|
|
|
|
3086
|
|
|
|
|
|
|
=for example |
3087
|
|
|
|
|
|
|
|
3088
|
|
|
|
|
|
|
$data = sequence(4,3,2); # example data array |
3089
|
|
|
|
|
|
|
$mask4 = (random(4)>0.5); # example 1-D mask array, has $n4 true values |
3090
|
|
|
|
|
|
|
$mask43 = (random(4,3)>0.5); # example 2-D mask array, has $n43 true values |
3091
|
|
|
|
|
|
|
$sdat4 = whereND $data, $mask4; # $sdat4 is a [$n4,3,2] pdl |
3092
|
|
|
|
|
|
|
$sdat43 = whereND $data, $mask43; # $sdat43 is a [$n43,2] pdl |
3093
|
|
|
|
|
|
|
|
3094
|
|
|
|
|
|
|
Just as with C, you can use the returned value in an |
3095
|
|
|
|
|
|
|
assignment. That means that both of these examples are valid: |
3096
|
|
|
|
|
|
|
|
3097
|
|
|
|
|
|
|
# Used to create a new slice stored in $sdat4: |
3098
|
|
|
|
|
|
|
$sdat4 = $data->whereND($mask4); |
3099
|
|
|
|
|
|
|
$sdat4 .= 0; |
3100
|
|
|
|
|
|
|
# Used in lvalue context: |
3101
|
|
|
|
|
|
|
$data->whereND($mask4) .= 0; |
3102
|
|
|
|
|
|
|
|
3103
|
|
|
|
|
|
|
=cut |
3104
|
|
|
|
|
|
|
|
3105
|
|
|
|
|
|
|
sub PDLA::whereND :lvalue { |
3106
|
5
|
50
|
|
5
|
0
|
149
|
barf "Usage: whereND( \$pdl1, ..., \$pdlN, \$mask )\n" if $#_ == 0; |
3107
|
|
|
|
|
|
|
|
3108
|
5
|
|
|
|
|
10
|
my $mask = pop @_; # $mask has 0==false, 1==true |
3109
|
5
|
|
|
|
|
6
|
my @to_return; |
3110
|
|
|
|
|
|
|
|
3111
|
5
|
|
|
|
|
13
|
my $n = PDLA::sum($mask); |
3112
|
|
|
|
|
|
|
|
3113
|
5
|
|
|
|
|
14
|
foreach my $arr (@_) { |
3114
|
|
|
|
|
|
|
|
3115
|
5
|
|
|
|
|
13
|
my $sub_i = $mask * ones($arr); |
3116
|
5
|
|
|
|
|
22
|
my $where_sub_i = PDLA::where($arr, $sub_i); |
3117
|
|
|
|
|
|
|
|
3118
|
|
|
|
|
|
|
# count the number of dims in $mask and $arr |
3119
|
|
|
|
|
|
|
# $mask = a b c d e f..... |
3120
|
5
|
|
|
|
|
15
|
my @idims = dims($arr); |
3121
|
|
|
|
|
|
|
|
3122
|
|
|
|
|
|
|
# ...and pop off the number of dims in $mask |
3123
|
5
|
|
|
|
|
12
|
foreach ( dims($mask) ) { shift(@idims) }; |
|
8
|
|
|
|
|
14
|
|
3124
|
|
|
|
|
|
|
|
3125
|
5
|
|
|
|
|
9
|
my $ndim = 0; |
3126
|
5
|
|
|
|
|
12
|
foreach my $id ($n, @idims[0..($#idims-1)]) { |
3127
|
7
|
100
|
|
|
|
43
|
$where_sub_i = $where_sub_i->splitdim($ndim++,$id) if $n>0; |
3128
|
|
|
|
|
|
|
} |
3129
|
|
|
|
|
|
|
|
3130
|
5
|
|
|
|
|
19
|
push @to_return, $where_sub_i; |
3131
|
|
|
|
|
|
|
} |
3132
|
|
|
|
|
|
|
|
3133
|
5
|
50
|
|
|
|
22
|
return (@to_return == 1) ? $to_return[0] : @to_return; |
3134
|
|
|
|
|
|
|
} |
3135
|
|
|
|
|
|
|
*whereND = \&PDLA::whereND; |
3136
|
|
|
|
|
|
|
|
3137
|
|
|
|
|
|
|
|
3138
|
|
|
|
|
|
|
|
3139
|
|
|
|
|
|
|
|
3140
|
|
|
|
|
|
|
=head2 whichND |
3141
|
|
|
|
|
|
|
|
3142
|
|
|
|
|
|
|
=for ref |
3143
|
|
|
|
|
|
|
|
3144
|
|
|
|
|
|
|
Return the coordinates of non-zero values in a mask. |
3145
|
|
|
|
|
|
|
|
3146
|
|
|
|
|
|
|
=for usage |
3147
|
|
|
|
|
|
|
|
3148
|
|
|
|
|
|
|
WhichND returns the N-dimensional coordinates of each nonzero value in |
3149
|
|
|
|
|
|
|
a mask PDLA with any number of dimensions. The returned values arrive |
3150
|
|
|
|
|
|
|
as an array-of-vectors suitable for use in |
3151
|
|
|
|
|
|
|
L or L. |
3152
|
|
|
|
|
|
|
|
3153
|
|
|
|
|
|
|
$coords = whichND($mask); |
3154
|
|
|
|
|
|
|
|
3155
|
|
|
|
|
|
|
returns a PDLA containing the coordinates of the elements that are non-zero |
3156
|
|
|
|
|
|
|
in C<$mask>, suitable for use in indexND. The 0th dimension contains the |
3157
|
|
|
|
|
|
|
full coordinate listing of each point; the 1st dimension lists all the points. |
3158
|
|
|
|
|
|
|
For example, if $mask has rank 4 and 100 matching elements, then $coords has |
3159
|
|
|
|
|
|
|
dimension 4x100. |
3160
|
|
|
|
|
|
|
|
3161
|
|
|
|
|
|
|
If no such elements exist, then whichND returns a structured empty PDLA: |
3162
|
|
|
|
|
|
|
an Nx0 PDLA that contains no values (but matches, threading-wise, with |
3163
|
|
|
|
|
|
|
the vectors that would be produced if such elements existed). |
3164
|
|
|
|
|
|
|
|
3165
|
|
|
|
|
|
|
DEPRECATED BEHAVIOR IN LIST CONTEXT: |
3166
|
|
|
|
|
|
|
|
3167
|
|
|
|
|
|
|
whichND once delivered different values in list context than in scalar |
3168
|
|
|
|
|
|
|
context, for historical reasons. In list context, it returned the |
3169
|
|
|
|
|
|
|
coordinates transposed, as a collection of 1-PDLAs (one per dimension) |
3170
|
|
|
|
|
|
|
in a list. This usage is deprecated in PDLA 2.4.10, and will cause a |
3171
|
|
|
|
|
|
|
warning to be issued every time it is encountered. To avoid the |
3172
|
|
|
|
|
|
|
warning, you can set the global variable "$PDLA::whichND" to 's' to |
3173
|
|
|
|
|
|
|
get scalar behavior in all contexts, or to 'l' to get list behavior in |
3174
|
|
|
|
|
|
|
list context. |
3175
|
|
|
|
|
|
|
|
3176
|
|
|
|
|
|
|
In later versions of PDLA, the deprecated behavior will disappear. Deprecated |
3177
|
|
|
|
|
|
|
list context whichND expressions can be replaced with: |
3178
|
|
|
|
|
|
|
|
3179
|
|
|
|
|
|
|
@list = $x->whichND->mv(0,-1)->dog; |
3180
|
|
|
|
|
|
|
|
3181
|
|
|
|
|
|
|
|
3182
|
|
|
|
|
|
|
SEE ALSO: |
3183
|
|
|
|
|
|
|
|
3184
|
|
|
|
|
|
|
L finds coordinates of nonzero values in a 1-D mask. |
3185
|
|
|
|
|
|
|
|
3186
|
|
|
|
|
|
|
L extracts values from a data PDLA that are associated |
3187
|
|
|
|
|
|
|
with nonzero values in a mask PDLA. |
3188
|
|
|
|
|
|
|
|
3189
|
|
|
|
|
|
|
=for example |
3190
|
|
|
|
|
|
|
|
3191
|
|
|
|
|
|
|
pdla> $s=sequence(10,10,3,4) |
3192
|
|
|
|
|
|
|
pdla> ($x, $y, $z, $w)=whichND($s == 203); p $x, $y, $z, $w |
3193
|
|
|
|
|
|
|
[3] [0] [2] [0] |
3194
|
|
|
|
|
|
|
pdla> print $s->at(list(cat($x,$y,$z,$w))) |
3195
|
|
|
|
|
|
|
203 |
3196
|
|
|
|
|
|
|
|
3197
|
|
|
|
|
|
|
=cut |
3198
|
|
|
|
|
|
|
|
3199
|
|
|
|
|
|
|
*whichND = \&PDLA::whichND; |
3200
|
|
|
|
|
|
|
sub PDLA::whichND { |
3201
|
11
|
|
|
11
|
0
|
269
|
my $mask = shift; |
3202
|
11
|
50
|
|
|
|
43
|
$mask = PDLA::pdl('PDLA',$mask) unless(UNIVERSAL::isa($mask,'PDLA')); |
3203
|
|
|
|
|
|
|
|
3204
|
|
|
|
|
|
|
# List context: generate a perl list by dimension |
3205
|
11
|
50
|
|
|
|
37
|
if(wantarray) { |
3206
|
0
|
0
|
|
|
|
0
|
if(!defined($PDLA::whichND)) { |
|
|
0
|
|
|
|
|
|
3207
|
0
|
|
|
|
|
0
|
printf STDERR "whichND: WARNING - list context deprecated. Set \$PDLA::whichND. Details in pod."; |
3208
|
|
|
|
|
|
|
} elsif($PDLA::whichND =~ m/l/i) { |
3209
|
|
|
|
|
|
|
# old list context enabled by setting $PDLA::whichND to 'l' |
3210
|
0
|
|
|
|
|
0
|
my $ind=($mask->clump(-1))->which; |
3211
|
0
|
|
|
|
|
0
|
return $mask->one2nd($ind); |
3212
|
|
|
|
|
|
|
} |
3213
|
|
|
|
|
|
|
# if $PDLA::whichND does not contain 'l' or 'L', fall through to scalar context |
3214
|
|
|
|
|
|
|
} |
3215
|
|
|
|
|
|
|
|
3216
|
|
|
|
|
|
|
# Scalar context: generate an N-D index piddle |
3217
|
|
|
|
|
|
|
|
3218
|
11
|
100
|
|
|
|
57
|
unless($mask->nelem) { |
3219
|
2
|
|
|
|
|
6
|
return PDLA::new_from_specification('PDLA',indx,$mask->ndims,0); |
3220
|
|
|
|
|
|
|
} |
3221
|
|
|
|
|
|
|
|
3222
|
9
|
100
|
|
|
|
33
|
unless($mask->getndims) { |
3223
|
2
|
100
|
|
|
|
7
|
return $mask ? pdl(indx,0) : PDLA::new_from_specification('PDLA',indx,0); |
3224
|
|
|
|
|
|
|
} |
3225
|
|
|
|
|
|
|
|
3226
|
7
|
|
|
|
|
25
|
$ind = $mask->flat->which->dummy(0,$mask->getndims)->make_physical; |
3227
|
7
|
100
|
|
|
|
134
|
if($ind->nelem==0) { |
3228
|
|
|
|
|
|
|
# In the empty case, explicitly return the correct type of structured empty |
3229
|
1
|
|
|
|
|
5
|
return PDLA::new_from_specification('PDLA',indx,$mask->ndims, 0); |
3230
|
|
|
|
|
|
|
} |
3231
|
|
|
|
|
|
|
|
3232
|
6
|
|
|
|
|
36
|
my $mult = ones($mask->getndims)->long; |
3233
|
6
|
|
|
|
|
49
|
my @mdims = $mask->dims; |
3234
|
6
|
|
|
|
|
15
|
my $i; |
3235
|
|
|
|
|
|
|
|
3236
|
6
|
|
|
|
|
17
|
for $i(0..$#mdims-1) { |
3237
|
|
|
|
|
|
|
# use $tmp for 5.005_03 compatibility |
3238
|
10
|
|
|
|
|
500
|
(my $tmp = $mult->index($i+1)) .= $mult->index($i)*$mdims[$i]; |
3239
|
|
|
|
|
|
|
} |
3240
|
|
|
|
|
|
|
|
3241
|
6
|
|
|
|
|
26
|
for $i(0..$#mdims) { |
3242
|
16
|
|
|
|
|
118
|
my($s) = $ind->index($i); |
3243
|
16
|
|
|
|
|
163
|
$s /= $mult->index($i); |
3244
|
16
|
|
|
|
|
105
|
$s %= $mdims[$i]; |
3245
|
|
|
|
|
|
|
} |
3246
|
|
|
|
|
|
|
|
3247
|
6
|
|
|
|
|
52
|
return $ind; |
3248
|
|
|
|
|
|
|
} |
3249
|
|
|
|
|
|
|
|
3250
|
|
|
|
|
|
|
|
3251
|
|
|
|
|
|
|
|
3252
|
|
|
|
|
|
|
|
3253
|
|
|
|
|
|
|
=head2 setops |
3254
|
|
|
|
|
|
|
|
3255
|
|
|
|
|
|
|
=for ref |
3256
|
|
|
|
|
|
|
|
3257
|
|
|
|
|
|
|
Implements simple set operations like union and intersection |
3258
|
|
|
|
|
|
|
|
3259
|
|
|
|
|
|
|
=for usage |
3260
|
|
|
|
|
|
|
|
3261
|
|
|
|
|
|
|
Usage: $set = setops($x, , $y); |
3262
|
|
|
|
|
|
|
|
3263
|
|
|
|
|
|
|
The operator can be C, C or C. This is then applied |
3264
|
|
|
|
|
|
|
to C<$x> viewed as a set and C<$y> viewed as a set. Set theory says |
3265
|
|
|
|
|
|
|
that a set may not have two or more identical elements, but setops |
3266
|
|
|
|
|
|
|
takes care of this for you, so C<$x=pdl(1,1,2)> is OK. The functioning |
3267
|
|
|
|
|
|
|
is as follows: |
3268
|
|
|
|
|
|
|
|
3269
|
|
|
|
|
|
|
=over |
3270
|
|
|
|
|
|
|
|
3271
|
|
|
|
|
|
|
=item C |
3272
|
|
|
|
|
|
|
|
3273
|
|
|
|
|
|
|
The resulting vector will contain the elements that are either in C<$x> |
3274
|
|
|
|
|
|
|
I in C<$y> or both. This is the union in set operation terms |
3275
|
|
|
|
|
|
|
|
3276
|
|
|
|
|
|
|
=item C |
3277
|
|
|
|
|
|
|
|
3278
|
|
|
|
|
|
|
The resulting vector will contain the elements that are either in C<$x> |
3279
|
|
|
|
|
|
|
or C<$y>, but not in both. This is |
3280
|
|
|
|
|
|
|
|
3281
|
|
|
|
|
|
|
Union($x, $y) - Intersection($x, $y) |
3282
|
|
|
|
|
|
|
|
3283
|
|
|
|
|
|
|
in set operation terms. |
3284
|
|
|
|
|
|
|
|
3285
|
|
|
|
|
|
|
=item C |
3286
|
|
|
|
|
|
|
|
3287
|
|
|
|
|
|
|
The resulting vector will contain the intersection of C<$x> and C<$y>, so |
3288
|
|
|
|
|
|
|
the elements that are in both C<$x> and C<$y>. Note that for convenience |
3289
|
|
|
|
|
|
|
this operation is also aliased to L. |
3290
|
|
|
|
|
|
|
|
3291
|
|
|
|
|
|
|
=back |
3292
|
|
|
|
|
|
|
|
3293
|
|
|
|
|
|
|
It should be emphasized that these routines are used when one or both of |
3294
|
|
|
|
|
|
|
the sets C<$x>, C<$y> are hard to calculate or that you get from a separate |
3295
|
|
|
|
|
|
|
subroutine. |
3296
|
|
|
|
|
|
|
|
3297
|
|
|
|
|
|
|
Finally IDL users might be familiar with Craig Markwardt's C |
3298
|
|
|
|
|
|
|
routine which has inspired this routine although it was written independently |
3299
|
|
|
|
|
|
|
However the present routine has a few less options (but see the examples) |
3300
|
|
|
|
|
|
|
|
3301
|
|
|
|
|
|
|
=for example |
3302
|
|
|
|
|
|
|
|
3303
|
|
|
|
|
|
|
You will very often use these functions on an index vector, so that is |
3304
|
|
|
|
|
|
|
what we will show here. We will in fact something slightly silly. First |
3305
|
|
|
|
|
|
|
we will find all squares that are also cubes below 10000. |
3306
|
|
|
|
|
|
|
|
3307
|
|
|
|
|
|
|
Create a sequence vector: |
3308
|
|
|
|
|
|
|
|
3309
|
|
|
|
|
|
|
pdla> $x = sequence(10000) |
3310
|
|
|
|
|
|
|
|
3311
|
|
|
|
|
|
|
Find all odd and even elements: |
3312
|
|
|
|
|
|
|
|
3313
|
|
|
|
|
|
|
pdla> ($even, $odd) = which_both( ($x % 2) == 0) |
3314
|
|
|
|
|
|
|
|
3315
|
|
|
|
|
|
|
Find all squares |
3316
|
|
|
|
|
|
|
|
3317
|
|
|
|
|
|
|
pdla> $squares= which(ceil(sqrt($x)) == floor(sqrt($x))) |
3318
|
|
|
|
|
|
|
|
3319
|
|
|
|
|
|
|
Find all cubes (being careful with roundoff error!) |
3320
|
|
|
|
|
|
|
|
3321
|
|
|
|
|
|
|
pdla> $cubes= which(ceil($x**(1.0/3.0)) == floor($x**(1.0/3.0)+1e-6)) |
3322
|
|
|
|
|
|
|
|
3323
|
|
|
|
|
|
|
Then find all squares that are cubes: |
3324
|
|
|
|
|
|
|
|
3325
|
|
|
|
|
|
|
pdla> $both = setops($squares, 'AND', $cubes) |
3326
|
|
|
|
|
|
|
|
3327
|
|
|
|
|
|
|
And print these (assumes that C is loaded!) |
3328
|
|
|
|
|
|
|
|
3329
|
|
|
|
|
|
|
pdla> p $x($both) |
3330
|
|
|
|
|
|
|
[0 1 64 729 4096] |
3331
|
|
|
|
|
|
|
|
3332
|
|
|
|
|
|
|
Then find all numbers that are either cubes or squares, but not both: |
3333
|
|
|
|
|
|
|
|
3334
|
|
|
|
|
|
|
pdla> $cube_xor_square = setops($squares, 'XOR', $cubes) |
3335
|
|
|
|
|
|
|
|
3336
|
|
|
|
|
|
|
pdla> p $cube_xor_square->nelem() |
3337
|
|
|
|
|
|
|
112 |
3338
|
|
|
|
|
|
|
|
3339
|
|
|
|
|
|
|
So there are a total of 112 of these! |
3340
|
|
|
|
|
|
|
|
3341
|
|
|
|
|
|
|
Finally find all odd squares: |
3342
|
|
|
|
|
|
|
|
3343
|
|
|
|
|
|
|
pdla> $odd_squares = setops($squares, 'AND', $odd) |
3344
|
|
|
|
|
|
|
|
3345
|
|
|
|
|
|
|
|
3346
|
|
|
|
|
|
|
Another common occurrence is to want to get all objects that are |
3347
|
|
|
|
|
|
|
in C<$x> and in the complement of C<$y>. But it is almost always best |
3348
|
|
|
|
|
|
|
to create the complement explicitly since the universe that both are |
3349
|
|
|
|
|
|
|
taken from is not known. Thus use L if possible |
3350
|
|
|
|
|
|
|
to keep track of complements. |
3351
|
|
|
|
|
|
|
|
3352
|
|
|
|
|
|
|
If this is impossible the best approach is to make a temporary: |
3353
|
|
|
|
|
|
|
|
3354
|
|
|
|
|
|
|
This creates an index vector the size of the universe of the sets and |
3355
|
|
|
|
|
|
|
set all elements in C<$y> to 0 |
3356
|
|
|
|
|
|
|
|
3357
|
|
|
|
|
|
|
pdla> $tmp = ones($n_universe); $tmp($y) .= 0; |
3358
|
|
|
|
|
|
|
|
3359
|
|
|
|
|
|
|
This then finds the complement of C<$y> |
3360
|
|
|
|
|
|
|
|
3361
|
|
|
|
|
|
|
pdla> $C_b = which($tmp == 1); |
3362
|
|
|
|
|
|
|
|
3363
|
|
|
|
|
|
|
and this does the final selection: |
3364
|
|
|
|
|
|
|
|
3365
|
|
|
|
|
|
|
pdla> $set = setops($x, 'AND', $C_b) |
3366
|
|
|
|
|
|
|
|
3367
|
|
|
|
|
|
|
=cut |
3368
|
|
|
|
|
|
|
|
3369
|
|
|
|
|
|
|
*setops = \&PDLA::setops; |
3370
|
|
|
|
|
|
|
|
3371
|
|
|
|
|
|
|
sub PDLA::setops { |
3372
|
|
|
|
|
|
|
|
3373
|
5
|
|
|
5
|
0
|
744
|
my ($x, $op, $y)=@_; |
3374
|
|
|
|
|
|
|
|
3375
|
|
|
|
|
|
|
# Check that $x and $y are 1D. |
3376
|
5
|
50
|
33
|
|
|
35
|
if ($x->ndims() > 1 || $y->ndims() > 1) { |
3377
|
0
|
|
|
|
|
0
|
warn 'setops: $x and $y must be 1D - flattening them!'."\n"; |
3378
|
0
|
|
|
|
|
0
|
$x = $x->flat; |
3379
|
0
|
|
|
|
|
0
|
$y = $y->flat; |
3380
|
|
|
|
|
|
|
} |
3381
|
|
|
|
|
|
|
|
3382
|
|
|
|
|
|
|
#Make sure there are no duplicate elements. |
3383
|
5
|
|
|
|
|
15
|
$x=$x->uniq; |
3384
|
5
|
|
|
|
|
11
|
$y=$y->uniq; |
3385
|
|
|
|
|
|
|
|
3386
|
5
|
|
|
|
|
10
|
my $result; |
3387
|
|
|
|
|
|
|
|
3388
|
5
|
100
|
|
|
|
17
|
if ($op eq 'OR') { |
|
|
100
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
3389
|
|
|
|
|
|
|
# Easy... |
3390
|
1
|
|
|
|
|
10
|
$result = uniq(append($x, $y)); |
3391
|
|
|
|
|
|
|
} elsif ($op eq 'XOR') { |
3392
|
|
|
|
|
|
|
# Make ordered list of set union. |
3393
|
1
|
|
|
|
|
17
|
my $union = append($x, $y)->qsort; |
3394
|
|
|
|
|
|
|
# Index lists. |
3395
|
1
|
|
|
|
|
8
|
my $s1=zeroes(byte, $union->nelem()); |
3396
|
1
|
|
|
|
|
5
|
my $s2=zeroes(byte, $union->nelem()); |
3397
|
|
|
|
|
|
|
|
3398
|
|
|
|
|
|
|
# Find indices which are duplicated - these are to be excluded |
3399
|
|
|
|
|
|
|
# |
3400
|
|
|
|
|
|
|
# We do this by comparing x with x shifted each way. |
3401
|
1
|
|
|
|
|
21
|
my $i1 = which($union != rotate($union, 1)); |
3402
|
1
|
|
|
|
|
23
|
my $i2 = which($union != rotate($union, -1)); |
3403
|
|
|
|
|
|
|
# |
3404
|
|
|
|
|
|
|
# We then mark/mask these in the s1 and s2 arrays to indicate which ones |
3405
|
|
|
|
|
|
|
# are not equal to their neighbours. |
3406
|
|
|
|
|
|
|
# |
3407
|
1
|
|
|
|
|
8
|
my $ts; |
3408
|
1
|
50
|
|
|
|
12
|
($ts = $s1->index($i1)) .= 1 if $i1->nelem() > 0; |
3409
|
1
|
50
|
|
|
|
12
|
($ts = $s2->index($i2)) .= 1 if $i2->nelem() > 0; |
3410
|
|
|
|
|
|
|
|
3411
|
1
|
|
|
|
|
12
|
my $inds=which($s1 == $s2); |
3412
|
|
|
|
|
|
|
|
3413
|
1
|
50
|
|
|
|
8
|
if ($inds->nelem() > 0) { |
3414
|
1
|
|
|
|
|
15
|
return $union->index($inds); |
3415
|
|
|
|
|
|
|
} else { |
3416
|
0
|
|
|
|
|
0
|
return $inds; |
3417
|
|
|
|
|
|
|
} |
3418
|
|
|
|
|
|
|
|
3419
|
|
|
|
|
|
|
} elsif ($op eq 'AND') { |
3420
|
|
|
|
|
|
|
# The intersection of the arrays. |
3421
|
|
|
|
|
|
|
|
3422
|
|
|
|
|
|
|
# Make ordered list of set union. |
3423
|
3
|
|
|
|
|
44
|
my $union = append($x, $y)->qsort; |
3424
|
|
|
|
|
|
|
|
3425
|
3
|
|
|
|
|
59
|
return $union->where($union == rotate($union, -1)); |
3426
|
|
|
|
|
|
|
} else { |
3427
|
0
|
|
|
|
|
0
|
print "The operation $op is not known!"; |
3428
|
0
|
|
|
|
|
0
|
return -1; |
3429
|
|
|
|
|
|
|
} |
3430
|
|
|
|
|
|
|
|
3431
|
|
|
|
|
|
|
} |
3432
|
|
|
|
|
|
|
|
3433
|
|
|
|
|
|
|
|
3434
|
|
|
|
|
|
|
|
3435
|
|
|
|
|
|
|
=head2 intersect |
3436
|
|
|
|
|
|
|
|
3437
|
|
|
|
|
|
|
=for ref |
3438
|
|
|
|
|
|
|
|
3439
|
|
|
|
|
|
|
Calculate the intersection of two piddles |
3440
|
|
|
|
|
|
|
|
3441
|
|
|
|
|
|
|
=for usage |
3442
|
|
|
|
|
|
|
|
3443
|
|
|
|
|
|
|
Usage: $set = intersect($x, $y); |
3444
|
|
|
|
|
|
|
|
3445
|
|
|
|
|
|
|
This routine is merely a simple interface to L. See |
3446
|
|
|
|
|
|
|
that for more information |
3447
|
|
|
|
|
|
|
|
3448
|
|
|
|
|
|
|
=for example |
3449
|
|
|
|
|
|
|
|
3450
|
|
|
|
|
|
|
Find all numbers less that 100 that are of the form 2*y and 3*x |
3451
|
|
|
|
|
|
|
|
3452
|
|
|
|
|
|
|
pdla> $x=sequence(100) |
3453
|
|
|
|
|
|
|
pdla> $factor2 = which( ($x % 2) == 0) |
3454
|
|
|
|
|
|
|
pdla> $factor3 = which( ($x % 3) == 0) |
3455
|
|
|
|
|
|
|
pdla> $ii=intersect($factor2, $factor3) |
3456
|
|
|
|
|
|
|
pdla> p $x($ii) |
3457
|
|
|
|
|
|
|
[0 6 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96] |
3458
|
|
|
|
|
|
|
|
3459
|
|
|
|
|
|
|
=cut |
3460
|
|
|
|
|
|
|
|
3461
|
|
|
|
|
|
|
*intersect = \&PDLA::intersect; |
3462
|
|
|
|
|
|
|
|
3463
|
|
|
|
|
|
|
sub PDLA::intersect { |
3464
|
|
|
|
|
|
|
|
3465
|
2
|
|
|
2
|
0
|
376
|
return setops($_[0], 'AND', $_[1]); |
3466
|
|
|
|
|
|
|
|
3467
|
|
|
|
|
|
|
} |
3468
|
|
|
|
|
|
|
|
3469
|
|
|
|
|
|
|
|
3470
|
|
|
|
|
|
|
|
3471
|
|
|
|
|
|
|
; |
3472
|
|
|
|
|
|
|
|
3473
|
|
|
|
|
|
|
|
3474
|
|
|
|
|
|
|
=head1 AUTHOR |
3475
|
|
|
|
|
|
|
|
3476
|
|
|
|
|
|
|
Copyright (C) Tuomas J. Lukka 1997 (lukka@husc.harvard.edu). Contributions |
3477
|
|
|
|
|
|
|
by Christian Soeller (c.soeller@auckland.ac.nz), Karl Glazebrook |
3478
|
|
|
|
|
|
|
(kgb@aaoepp.aao.gov.au), Craig DeForest (deforest@boulder.swri.edu) |
3479
|
|
|
|
|
|
|
and Jarle Brinchmann (jarle@astro.up.pt) |
3480
|
|
|
|
|
|
|
All rights reserved. There is no warranty. You are allowed |
3481
|
|
|
|
|
|
|
to redistribute this software / documentation under certain |
3482
|
|
|
|
|
|
|
conditions. For details, see the file COPYING in the PDLA |
3483
|
|
|
|
|
|
|
distribution. If this file is separated from the PDLA distribution, |
3484
|
|
|
|
|
|
|
the copyright notice should be included in the file. |
3485
|
|
|
|
|
|
|
|
3486
|
|
|
|
|
|
|
Updated for CPAN viewing compatibility by David Mertens. |
3487
|
|
|
|
|
|
|
|
3488
|
|
|
|
|
|
|
=cut |
3489
|
|
|
|
|
|
|
|
3490
|
|
|
|
|
|
|
|
3491
|
|
|
|
|
|
|
|
3492
|
|
|
|
|
|
|
|
3493
|
|
|
|
|
|
|
|
3494
|
|
|
|
|
|
|
# Exit with OK status |
3495
|
|
|
|
|
|
|
|
3496
|
|
|
|
|
|
|
1; |
3497
|
|
|
|
|
|
|
|
3498
|
|
|
|
|
|
|
|