nimble/UserManual/cha-progr-with-models.html at devel · nimble-dev/nimble · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
<!DOCTYPE html>
<html lang="" xml:lang="">
<head>

  <meta charset="utf-8" />
  <meta http-equiv="X-UA-Compatible" content="IE=edge" />
  <title>Chapter 16 Writing nimbleFunctions to interact with models | NimbleUserManual.knit</title>
  <meta name="description" content="This is the NIMBLE User Manual." />
  <meta name="generator" content="bookdown 0.37 and GitBook 2.6.7" />

  <meta property="og:title" content="Chapter 16 Writing nimbleFunctions to interact with models | NimbleUserManual.knit" />
  <meta property="og:type" content="book" />
  <meta property="og:image" content="/nimble-icon.png" />
  <meta property="og:description" content="This is the NIMBLE User Manual." />
  <meta name="github-repo" content="nimble-dev/nimble" />

  <meta name="twitter:card" content="summary" />
  <meta name="twitter:title" content="Chapter 16 Writing nimbleFunctions to interact with models | NimbleUserManual.knit" />

  <meta name="twitter:description" content="This is the NIMBLE User Manual." />
  <meta name="twitter:image" content="/nimble-icon.png" />


  <meta name="viewport" content="width=device-width, initial-scale=1" />
  <meta name="apple-mobile-web-app-capable" content="yes" />
  <meta name="apple-mobile-web-app-status-bar-style" content="black" />


<link rel="prev" href="cha-data-structures.html"/>
<link rel="next" href="cha-AD.html"/>
<script src="libs/jquery-3.6.0/jquery-3.6.0.min.js"></script>
<script src="https://cdn.jsdelivr.net/npm/fuse.js@6.4.6/dist/fuse.min.js"></script>
<link href="libs/gitbook-2.6.7/css/style.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-table.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-bookdown.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-highlight.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-search.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-fontsettings.css" rel="stylesheet" />
<link href="libs/gitbook-2.6.7/css/plugin-clipboard.css" rel="stylesheet" />


<link href="libs/anchor-sections-1.1.0/anchor-sections.css" rel="stylesheet" />
<link href="libs/anchor-sections-1.1.0/anchor-sections-hash.css" rel="stylesheet" />
<script src="libs/anchor-sections-1.1.0/anchor-sections.js"></script>


<style type="text/css">
pre > code.sourceCode { white-space: pre; position: relative; }
pre > code.sourceCode > span { line-height: 1.25; }
pre > code.sourceCode > span:empty { height: 1.2em; }
.sourceCode { overflow: visible; }
code.sourceCode > span { color: inherit; text-decoration: inherit; }
pre.sourceCode { margin: 0; }
@media screen {
div.sourceCode { overflow: auto; }
}
@media print {
pre > code.sourceCode { white-space: pre-wrap; }
pre > code.sourceCode > span { text-indent: -5em; padding-left: 5em; }
}
pre.numberSource code
  { counter-reset: source-line 0; }
pre.numberSource code > span
  { position: relative; left: -4em; counter-increment: source-line; }
pre.numberSource code > span > a:first-child::before
  { content: counter(source-line);
    position: relative; left: -1em; text-align: right; vertical-align: baseline;
    border: none; display: inline-block;
    -webkit-touch-callout: none; -webkit-user-select: none;
    -khtml-user-select: none; -moz-user-select: none;
    -ms-user-select: none; user-select: none;
    padding: 0 4px; width: 4em;
    color: #aaaaaa;
  }
pre.numberSource { margin-left: 3em; border-left: 1px solid #aaaaaa;  padding-left: 4px; }
div.sourceCode
  {   }
@media screen {
pre > code.sourceCode > span > a:first-child::before { text-decoration: underline; }
}
code span.al { color: #ff0000; font-weight: bold; } /* Alert */
code span.an { color: #60a0b0; font-weight: bold; font-style: italic; } /* Annotation */
code span.at { color: #7d9029; } /* Attribute */
code span.bn { color: #40a070; } /* BaseN */
code span.bu { color: #008000; } /* BuiltIn */
code span.cf { color: #007020; font-weight: bold; } /* ControlFlow */
code span.ch { color: #4070a0; } /* Char */
code span.cn { color: #880000; } /* Constant */
code span.co { color: #60a0b0; font-style: italic; } /* Comment */
code span.cv { color: #60a0b0; font-weight: bold; font-style: italic; } /* CommentVar */
code span.do { color: #ba2121; font-style: italic; } /* Documentation */
code span.dt { color: #902000; } /* DataType */
code span.dv { color: #40a070; } /* DecVal */
code span.er { color: #ff0000; font-weight: bold; } /* Error */
code span.ex { } /* Extension */
code span.fl { color: #40a070; } /* Float */
code span.fu { color: #06287e; } /* Function */
code span.im { color: #008000; font-weight: bold; } /* Import */
code span.in { color: #60a0b0; font-weight: bold; font-style: italic; } /* Information */
code span.kw { color: #007020; font-weight: bold; } /* Keyword */
code span.op { color: #666666; } /* Operator */
code span.ot { color: #007020; } /* Other */
code span.pp { color: #bc7a00; } /* Preprocessor */
code span.sc { color: #4070a0; } /* SpecialChar */
code span.ss { color: #bb6688; } /* SpecialString */
code span.st { color: #4070a0; } /* String */
code span.va { color: #19177c; } /* Variable */
code span.vs { color: #4070a0; } /* VerbatimString */
code span.wa { color: #60a0b0; font-weight: bold; font-style: italic; } /* Warning */
</style>

<style type="text/css">

  div.hanging-indent{margin-left: 1.5em; text-indent: -1.5em;}
</style>
<style type="text/css">
/* Used with Pandoc 2.11+ new --citeproc when CSL is used */
div.csl-bib-body { }
div.csl-entry {
  clear: both;
}
.hanging div.csl-entry {
  margin-left:2em;
  text-indent:-2em;
}
div.csl-left-margin {
  min-width:2em;
  float:left;
}
div.csl-right-inline {
  margin-left:2em;
  padding-left:1em;
}
div.csl-indent {
  margin-left: 2em;
}
</style>

<link rel="stylesheet" href="style.css" type="text/css" />
</head>

<body>


  <div class="book without-animation with-summary font-size-2 font-family-1" data-basepath=".">

    <div class="book-summary">
      <nav role="navigation">

<ul class="summary">
<img src="./nimble-icon.png"
     width=100>
<li><a href="./cha-welcome-nimble.html">NIMBLE User Manual, Version 1.4.2</a></li>
<li><a href="https://github.com/nimble-dev/nimble">NIMBLE Development Team</a></li>
<li><a href="https://R-nimble.org">https://R-nimble.org</a></li>

<li class="divider"></li>
<li class="part"><span><b>I Introduction</b></span></li>
<li class="chapter" data-level="1" data-path="cha-welcome-nimble.html"><a href="cha-welcome-nimble.html"><i class="fa fa-check"></i><b>1</b> Welcome to NIMBLE</a>
<ul>
<li class="chapter" data-level="1.1" data-path="cha-welcome-nimble.html"><a href="cha-welcome-nimble.html#sec:what-is-nimble"><i class="fa fa-check"></i><b>1.1</b> What does NIMBLE do?</a></li>
<li class="chapter" data-level="1.2" data-path="cha-welcome-nimble.html"><a href="cha-welcome-nimble.html#how-to-use-this-manual"><i class="fa fa-check"></i><b>1.2</b> How to use this manual</a></li>
</ul></li>
<li class="chapter" data-level="2" data-path="cha-lightning-intro.html"><a href="cha-lightning-intro.html"><i class="fa fa-check"></i><b>2</b> Lightning introduction</a>
<ul>
<li class="chapter" data-level="2.1" data-path="cha-lightning-intro.html"><a href="cha-lightning-intro.html#sec:brief-example"><i class="fa fa-check"></i><b>2.1</b> A brief example</a></li>
<li class="chapter" data-level="2.2" data-path="cha-lightning-intro.html"><a href="cha-lightning-intro.html#sec:creating-model"><i class="fa fa-check"></i><b>2.2</b> Creating a model</a></li>
<li class="chapter" data-level="2.3" data-path="cha-lightning-intro.html"><a href="cha-lightning-intro.html#sec:compiling-model"><i class="fa fa-check"></i><b>2.3</b> Compiling the model</a></li>
<li class="chapter" data-level="2.4" data-path="cha-lightning-intro.html"><a href="cha-lightning-intro.html#sec:intro-runMCMC"><i class="fa fa-check"></i><b>2.4</b> One-line invocation of MCMC</a></li>
<li class="chapter" data-level="2.5" data-path="cha-lightning-intro.html"><a href="cha-lightning-intro.html#sec:creating-mcmc"><i class="fa fa-check"></i><b>2.5</b> Creating, compiling and running a basic MCMC configuration</a></li>
<li class="chapter" data-level="2.6" data-path="cha-lightning-intro.html"><a href="cha-lightning-intro.html#sec:customizing-mcmc"><i class="fa fa-check"></i><b>2.6</b> Customizing the MCMC</a></li>
<li class="chapter" data-level="2.7" data-path="cha-lightning-intro.html"><a href="cha-lightning-intro.html#sec:running-mcem"><i class="fa fa-check"></i><b>2.7</b> Running MCEM</a></li>
<li class="chapter" data-level="2.8" data-path="cha-lightning-intro.html"><a href="cha-lightning-intro.html#sec:creating-your-own"><i class="fa fa-check"></i><b>2.8</b> Creating your own functions</a></li>
</ul></li>
<li class="chapter" data-level="3" data-path="cha-more-introduction.html"><a href="cha-more-introduction.html"><i class="fa fa-check"></i><b>3</b> More introduction</a>
<ul>
<li class="chapter" data-level="3.1" data-path="cha-more-introduction.html"><a href="cha-more-introduction.html#nimble-adopts-and-extends-the-bugs-language-for-specifying-models"><i class="fa fa-check"></i><b>3.1</b> NIMBLE adopts and extends the BUGS language for specifying models</a></li>
<li class="chapter" data-level="3.2" data-path="cha-more-introduction.html"><a href="cha-more-introduction.html#sec:nimble-lang-writ"><i class="fa fa-check"></i><b>3.2</b> nimbleFunctions for writing algorithms</a></li>
<li class="chapter" data-level="3.3" data-path="cha-more-introduction.html"><a href="cha-more-introduction.html#sec:nimble-algor-libr"><i class="fa fa-check"></i><b>3.3</b> The NIMBLE algorithm library</a></li>
</ul></li>
<li class="chapter" data-level="4" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html"><i class="fa fa-check"></i><b>4</b> Installing NIMBLE</a>
<ul>
<li class="chapter" data-level="4.1" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#sec:requ-run-nimble"><i class="fa fa-check"></i><b>4.1</b> Requirements to run NIMBLE</a></li>
<li class="chapter" data-level="4.2" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#sec:compiler"><i class="fa fa-check"></i><b>4.2</b> Installing a C++ compiler for NIMBLE to use</a>
<ul>
<li class="chapter" data-level="4.2.1" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#macos"><i class="fa fa-check"></i><b>4.2.1</b> MacOS</a></li>
<li class="chapter" data-level="4.2.2" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#linux"><i class="fa fa-check"></i><b>4.2.2</b> Linux</a></li>
<li class="chapter" data-level="4.2.3" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#windows"><i class="fa fa-check"></i><b>4.2.3</b> Windows</a></li>
</ul></li>
<li class="chapter" data-level="4.3" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#installing-the-nimble-package"><i class="fa fa-check"></i><b>4.3</b> Installing the NIMBLE package</a></li>
<li class="chapter" data-level="4.4" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#troubleshooting-installation-problems"><i class="fa fa-check"></i><b>4.4</b> Troubleshooting installation problems</a></li>
<li class="chapter" data-level="4.5" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#customizing-your-installation"><i class="fa fa-check"></i><b>4.5</b> Customizing your installation</a>
<ul>
<li class="chapter" data-level="4.5.1" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#using-your-own-copy-of-eigen"><i class="fa fa-check"></i><b>4.5.1</b> Using your own copy of Eigen</a></li>
<li class="chapter" data-level="4.5.2" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#using-libnimble"><i class="fa fa-check"></i><b>4.5.2</b> Using libnimble</a></li>
<li class="chapter" data-level="4.5.3" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#sec:blas"><i class="fa fa-check"></i><b>4.5.3</b> BLAS and LAPACK</a></li>
<li class="chapter" data-level="4.5.4" data-path="cha-installing-nimble.html"><a href="cha-installing-nimble.html#customizing-compilation-of-the-nimble-generated-c"><i class="fa fa-check"></i><b>4.5.4</b> Customizing compilation of the NIMBLE-generated C++</a></li>
</ul></li>
</ul></li>
<li class="part"><span><b>II Models in NIMBLE</b></span></li>
<li class="chapter" data-level="5" data-path="cha-writing-models.html"><a href="cha-writing-models.html"><i class="fa fa-check"></i><b>5</b> Writing models in NIMBLE’s dialect of BUGS</a>
<ul>
<li class="chapter" data-level="5.1" data-path="cha-writing-models.html"><a href="cha-writing-models.html#sec:supp-feat-bugs"><i class="fa fa-check"></i><b>5.1</b> Comparison to BUGS dialects supported by WinBUGS, OpenBUGS and JAGS</a>
<ul>
<li class="chapter" data-level="5.1.1" data-path="cha-writing-models.html"><a href="cha-writing-models.html#supported-features-of-bugs-and-jags"><i class="fa fa-check"></i><b>5.1.1</b> Supported features of BUGS and JAGS</a></li>
<li class="chapter" data-level="5.1.2" data-path="cha-writing-models.html"><a href="cha-writing-models.html#sec:extensions-bugs"><i class="fa fa-check"></i><b>5.1.2</b> NIMBLE’s Extensions to BUGS and JAGS</a></li>
<li class="chapter" data-level="5.1.3" data-path="cha-writing-models.html"><a href="cha-writing-models.html#sec:not-yet-supported"><i class="fa fa-check"></i><b>5.1.3</b> Not-supported features of BUGS and JAGS</a></li>
</ul></li>
<li class="chapter" data-level="5.2" data-path="cha-writing-models.html"><a href="cha-writing-models.html#writing-models"><i class="fa fa-check"></i><b>5.2</b> Writing models</a>
<ul>
<li class="chapter" data-level="5.2.1" data-path="cha-writing-models.html"><a href="cha-writing-models.html#declaring-stochastic-and-deterministic-nodes"><i class="fa fa-check"></i><b>5.2.1</b> Declaring stochastic and deterministic nodes</a></li>
<li class="chapter" data-level="5.2.2" data-path="cha-writing-models.html"><a href="cha-writing-models.html#sec:more-kinds-bugs"><i class="fa fa-check"></i><b>5.2.2</b> More kinds of BUGS declarations</a></li>
<li class="chapter" data-level="5.2.3" data-path="cha-writing-models.html"><a href="cha-writing-models.html#subsec:vectorized-versus-scalar-declarations"><i class="fa fa-check"></i><b>5.2.3</b> Vectorized versus scalar declarations</a></li>
<li class="chapter" data-level="5.2.4" data-path="cha-writing-models.html"><a href="cha-writing-models.html#subsec:dists-and-functions"><i class="fa fa-check"></i><b>5.2.4</b> Available distributions</a></li>
<li class="chapter" data-level="5.2.5" data-path="cha-writing-models.html"><a href="cha-writing-models.html#subsec:BUGS-lang-fxns"><i class="fa fa-check"></i><b>5.2.5</b> Available BUGS language functions</a></li>
<li class="chapter" data-level="5.2.6" data-path="cha-writing-models.html"><a href="cha-writing-models.html#subsec:BUGS-link"><i class="fa fa-check"></i><b>5.2.6</b> Available link functions</a></li>
<li class="chapter" data-level="5.2.7" data-path="cha-writing-models.html"><a href="cha-writing-models.html#subsec:trunc"><i class="fa fa-check"></i><b>5.2.7</b> Truncation, censoring, and constraints</a></li>
<li class="chapter" data-level="5.2.8" data-path="cha-writing-models.html"><a href="cha-writing-models.html#subsec:macros"><i class="fa fa-check"></i><b>5.2.8</b> Model macros</a></li>
</ul></li>
</ul></li>
<li class="chapter" data-level="6" data-path="cha-building-models.html"><a href="cha-building-models.html"><i class="fa fa-check"></i><b>6</b> Building and using models</a>
<ul>
<li class="chapter" data-level="6.1" data-path="cha-building-models.html"><a href="cha-building-models.html#creating-model-objects"><i class="fa fa-check"></i><b>6.1</b> Creating model objects</a>
<ul>
<li class="chapter" data-level="6.1.1" data-path="cha-building-models.html"><a href="cha-building-models.html#using-nimblemodel-to-create-a-model"><i class="fa fa-check"></i><b>6.1.1</b> Using <em>nimbleModel</em> to create a model</a></li>
<li class="chapter" data-level="6.1.2" data-path="cha-building-models.html"><a href="cha-building-models.html#sec:readBUGSmodel"><i class="fa fa-check"></i><b>6.1.2</b> Creating a model from standard BUGS and JAGS input files</a></li>
<li class="chapter" data-level="6.1.3" data-path="cha-building-models.html"><a href="cha-building-models.html#sub:multiple-instances"><i class="fa fa-check"></i><b>6.1.3</b> Making multiple instances from the same model definition</a></li>
</ul></li>
<li class="chapter" data-level="6.2" data-path="cha-building-models.html"><a href="cha-building-models.html#sec:nodes-and-variables"><i class="fa fa-check"></i><b>6.2</b> NIMBLE models are objects you can query and manipulate</a>
<ul>
<li class="chapter" data-level="6.2.1" data-path="cha-building-models.html"><a href="cha-building-models.html#sec:what-are-nodes-and-variables"><i class="fa fa-check"></i><b>6.2.1</b> What are variables and nodes?</a></li>
<li class="chapter" data-level="6.2.2" data-path="cha-building-models.html"><a href="cha-building-models.html#determining-the-nodes-and-variables-in-a-model"><i class="fa fa-check"></i><b>6.2.2</b> Determining the nodes and variables in a model</a></li>
<li class="chapter" data-level="6.2.3" data-path="cha-building-models.html"><a href="cha-building-models.html#sec:accessing-nodes"><i class="fa fa-check"></i><b>6.2.3</b> Accessing nodes</a></li>
<li class="chapter" data-level="6.2.4" data-path="cha-building-models.html"><a href="cha-building-models.html#sec:how-nodes-are"><i class="fa fa-check"></i><b>6.2.4</b> How nodes are named</a></li>
<li class="chapter" data-level="6.2.5" data-path="cha-building-models.html"><a href="cha-building-models.html#sec:why-use-node"><i class="fa fa-check"></i><b>6.2.5</b> Why use node names?</a></li>
<li class="chapter" data-level="6.2.6" data-path="cha-building-models.html"><a href="cha-building-models.html#sec:cdisdata"><i class="fa fa-check"></i><b>6.2.6</b> Checking if a node holds data</a></li>
</ul></li>
<li class="chapter" data-level="6.3" data-path="cha-building-models.html"><a href="cha-building-models.html#using-models-in-parallel"><i class="fa fa-check"></i><b>6.3</b> Using models in parallel</a></li>
</ul></li>
<li class="part"><span><b>III Algorithms in NIMBLE</b></span></li>
<li class="chapter" data-level="7" data-path="cha-mcmc.html"><a href="cha-mcmc.html"><i class="fa fa-check"></i><b>7</b> MCMC</a>
<ul>
<li class="chapter" data-level="7.1" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:nimbleMCMC"><i class="fa fa-check"></i><b>7.1</b> One-line invocation of MCMC: <em>nimbleMCMC</em></a></li>
<li class="chapter" data-level="7.2" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:mcmc-configuration"><i class="fa fa-check"></i><b>7.2</b> The MCMC configuration</a>
<ul>
<li class="chapter" data-level="7.2.1" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:default-mcmc-conf"><i class="fa fa-check"></i><b>7.2.1</b> Default MCMC configuration</a></li>
<li class="chapter" data-level="7.2.2" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:customizing-mcmc-conf"><i class="fa fa-check"></i><b>7.2.2</b> Customizing the MCMC configuration</a></li>
<li class="chapter" data-level="7.2.3" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:derived-quantities"><i class="fa fa-check"></i><b>7.2.3</b> Setting up derived quantities for additional quantities of interest</a></li>
</ul></li>
<li class="chapter" data-level="7.3" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:build-compile-mcmc"><i class="fa fa-check"></i><b>7.3</b> Building and compiling the MCMC</a></li>
<li class="chapter" data-level="7.4" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:initMCMC"><i class="fa fa-check"></i><b>7.4</b> Initializing MCMC</a></li>
<li class="chapter" data-level="7.5" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:runMCMC"><i class="fa fa-check"></i><b>7.5</b> User-friendly execution of MCMC algorithms: <em>runMCMC</em></a></li>
<li class="chapter" data-level="7.6" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:executing-the-mcmc-algorithm"><i class="fa fa-check"></i><b>7.6</b> Running the MCMC</a>
<ul>
<li class="chapter" data-level="7.6.1" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:mcmc-rerun"><i class="fa fa-check"></i><b>7.6.1</b> Rerunning versus restarting an MCMC</a></li>
<li class="chapter" data-level="7.6.2" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:sampler-time"><i class="fa fa-check"></i><b>7.6.2</b> Measuring sampler computation times: <em>getTimes</em></a></li>
<li class="chapter" data-level="7.6.3" data-path="cha-mcmc.html"><a href="cha-mcmc.html#assessing-the-adaption-process-of-rw-and-rw_block-samplers"><i class="fa fa-check"></i><b>7.6.3</b> Assessing the adaption process of <em>RW</em> and <em>RW_block</em> samplers</a></li>
</ul></li>
<li class="chapter" data-level="7.7" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:extracting-samples"><i class="fa fa-check"></i><b>7.7</b> Extracting MCMC samples</a></li>
<li class="chapter" data-level="7.8" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:WAIC"><i class="fa fa-check"></i><b>7.8</b> Calculating WAIC</a></li>
<li class="chapter" data-level="7.9" data-path="cha-mcmc.html"><a href="cha-mcmc.html#k-fold-cross-validation"><i class="fa fa-check"></i><b>7.9</b> k-fold cross-validation</a></li>
<li class="chapter" data-level="7.10" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:rjmcmc"><i class="fa fa-check"></i><b>7.10</b> Variable selection using Reversible Jump MCMC</a>
<ul>
<li class="chapter" data-level="7.10.1" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:rjmcmc-indicator"><i class="fa fa-check"></i><b>7.10.1</b> Using indicator variables</a></li>
<li class="chapter" data-level="7.10.2" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:rjmcmc-no-indicator"><i class="fa fa-check"></i><b>7.10.2</b> Without indicator variables</a></li>
</ul></li>
<li class="chapter" data-level="7.11" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:samplers-provided"><i class="fa fa-check"></i><b>7.11</b> Samplers provided with NIMBLE</a>
<ul>
<li class="chapter" data-level="7.11.1" data-path="cha-mcmc.html"><a href="cha-mcmc.html#conjugate-gibbs-samplers"><i class="fa fa-check"></i><b>7.11.1</b> Conjugate (‘Gibbs’) samplers</a></li>
<li class="chapter" data-level="7.11.2" data-path="cha-mcmc.html"><a href="cha-mcmc.html#subsec:HMC"><i class="fa fa-check"></i><b>7.11.2</b> Hamiltonian Monte Carlo (HMC)</a></li>
<li class="chapter" data-level="7.11.3" data-path="cha-mcmc.html"><a href="cha-mcmc.html#particle-filter-samplers"><i class="fa fa-check"></i><b>7.11.3</b> Particle filter samplers</a></li>
<li class="chapter" data-level="7.11.4" data-path="cha-mcmc.html"><a href="cha-mcmc.html#customized-log-likelihood-evaluations-rw_llfunction-sampler"><i class="fa fa-check"></i><b>7.11.4</b> Customized log-likelihood evaluations: <em>RW_llFunction sampler</em></a></li>
</ul></li>
<li class="chapter" data-level="7.12" data-path="cha-mcmc.html"><a href="cha-mcmc.html#sec:mcmc-example-litters"><i class="fa fa-check"></i><b>7.12</b> Detailed MCMC example: <em>litters</em></a></li>
<li class="chapter" data-level="7.13" data-path="cha-mcmc.html"><a href="cha-mcmc.html#mcmc-suite-compare-mcmcs"><i class="fa fa-check"></i><b>7.13</b> Comparing different MCMCs with <em>compareMCMCs</em></a></li>
<li class="chapter" data-level="7.14" data-path="cha-mcmc.html"><a href="cha-mcmc.html#running-mcmc-chains-in-parallel"><i class="fa fa-check"></i><b>7.14</b> Running MCMC chains in parallel</a></li>
</ul></li>
<li class="chapter" data-level="8" data-path="cha-algos-provided.html"><a href="cha-algos-provided.html"><i class="fa fa-check"></i><b>8</b> Particle Filters, PMCMC, and MCEM</a>
<ul>
<li class="chapter" data-level="8.1" data-path="cha-algos-provided.html"><a href="cha-algos-provided.html#particle-filters-sequential-monte-carlo-and-iterated-filtering"><i class="fa fa-check"></i><b>8.1</b> Particle filters / sequential Monte Carlo and iterated filtering</a>
<ul>
<li class="chapter" data-level="8.1.1" data-path="cha-algos-provided.html"><a href="cha-algos-provided.html#filtering-algorithms"><i class="fa fa-check"></i><b>8.1.1</b> Filtering algorithms</a></li>
<li class="chapter" data-level="8.1.2" data-path="cha-algos-provided.html"><a href="cha-algos-provided.html#sec:particle-mcmc"><i class="fa fa-check"></i><b>8.1.2</b> Particle MCMC (PMCMC)</a></li>
</ul></li>
<li class="chapter" data-level="8.2" data-path="cha-algos-provided.html"><a href="cha-algos-provided.html#monte-carlo-expectation-maximization-mcem"><i class="fa fa-check"></i><b>8.2</b> Monte Carlo Expectation Maximization (MCEM)</a>
<ul>
<li class="chapter" data-level="8.2.1" data-path="cha-algos-provided.html"><a href="cha-algos-provided.html#sec:estimate-mcem-cov"><i class="fa fa-check"></i><b>8.2.1</b> Estimating the asymptotic covariance From MCEM</a></li>
</ul></li>
</ul></li>
<li class="chapter" data-level="9" data-path="cha-laplace.html"><a href="cha-laplace.html"><i class="fa fa-check"></i><b>9</b> Laplace, AGHQ, and nested approximations</a>
<ul>
<li class="chapter" data-level="9.1" data-path="cha-laplace.html"><a href="cha-laplace.html#sec:AD-laplace"><i class="fa fa-check"></i><b>9.1</b> Laplace approximation and adaptive Gauss-Hermite quadrature (AGHQ)</a>
<ul>
<li class="chapter" data-level="9.1.1" data-path="cha-laplace.html"><a href="cha-laplace.html#glmm-example"><i class="fa fa-check"></i><b>9.1.1</b> GLMM example</a></li>
<li class="chapter" data-level="9.1.2" data-path="cha-laplace.html"><a href="cha-laplace.html#using-laplace-approximation"><i class="fa fa-check"></i><b>9.1.2</b> Using Laplace approximation</a></li>
<li class="chapter" data-level="9.1.3" data-path="cha-laplace.html"><a href="cha-laplace.html#using-the-laplace-approximation-methods-directly"><i class="fa fa-check"></i><b>9.1.3</b> Using the Laplace approximation methods directly</a></li>
<li class="chapter" data-level="9.1.4" data-path="cha-laplace.html"><a href="cha-laplace.html#changing-the-optimization-methods"><i class="fa fa-check"></i><b>9.1.4</b> Changing the optimization methods</a></li>
</ul></li>
<li class="chapter" data-level="9.2" data-path="cha-laplace.html"><a href="cha-laplace.html#sec:nested-approx"><i class="fa fa-check"></i><b>9.2</b> Nested approximation (INLA-like) methods</a>
<ul>
<li class="chapter" data-level="9.2.1" data-path="cha-laplace.html"><a href="cha-laplace.html#overview-of-the-methodology"><i class="fa fa-check"></i><b>9.2.1</b> Overview of the methodology</a></li>
<li class="chapter" data-level="9.2.2" data-path="cha-laplace.html"><a href="cha-laplace.html#using-nimbles-nested-approximation"><i class="fa fa-check"></i><b>9.2.2</b> Using NIMBLE’s nested approximation</a></li>
</ul></li>
</ul></li>
<li class="chapter" data-level="10" data-path="cha-spatial.html"><a href="cha-spatial.html"><i class="fa fa-check"></i><b>10</b> Spatial models</a>
<ul>
<li class="chapter" data-level="10.1" data-path="cha-spatial.html"><a href="cha-spatial.html#intrinsic-gaussian-car-model-dcar_normal"><i class="fa fa-check"></i><b>10.1</b> Intrinsic Gaussian CAR model: <em>dcar_normal</em></a>
<ul>
<li class="chapter" data-level="10.1.1" data-path="cha-spatial.html"><a href="cha-spatial.html#specification-and-density"><i class="fa fa-check"></i><b>10.1.1</b> Specification and density</a></li>
<li class="chapter" data-level="10.1.2" data-path="cha-spatial.html"><a href="cha-spatial.html#example"><i class="fa fa-check"></i><b>10.1.2</b> Example</a></li>
</ul></li>
<li class="chapter" data-level="10.2" data-path="cha-spatial.html"><a href="cha-spatial.html#proper-gaussian-car-model-dcar_proper"><i class="fa fa-check"></i><b>10.2</b> Proper Gaussian CAR model: <em>dcar_proper</em></a>
<ul>
<li class="chapter" data-level="10.2.1" data-path="cha-spatial.html"><a href="cha-spatial.html#specification-and-density-1"><i class="fa fa-check"></i><b>10.2.1</b> Specification and density</a></li>
<li class="chapter" data-level="10.2.2" data-path="cha-spatial.html"><a href="cha-spatial.html#example-1"><i class="fa fa-check"></i><b>10.2.2</b> Example</a></li>
</ul></li>
<li class="chapter" data-level="10.3" data-path="cha-spatial.html"><a href="cha-spatial.html#sec:spatial-mcmc-sampling-car"><i class="fa fa-check"></i><b>10.3</b> MCMC Sampling of CAR models</a>
<ul>
<li class="chapter" data-level="10.3.1" data-path="cha-spatial.html"><a href="cha-spatial.html#initial-values"><i class="fa fa-check"></i><b>10.3.1</b> Initial values</a></li>
<li class="chapter" data-level="10.3.2" data-path="cha-spatial.html"><a href="cha-spatial.html#zero-neighbor-regions"><i class="fa fa-check"></i><b>10.3.2</b> Zero-neighbor regions</a></li>
<li class="chapter" data-level="10.3.3" data-path="cha-spatial.html"><a href="cha-spatial.html#zero-mean-constraint"><i class="fa fa-check"></i><b>10.3.3</b> Zero-mean constraint</a></li>
</ul></li>
</ul></li>
<li class="chapter" data-level="11" data-path="cha-bnp.html"><a href="cha-bnp.html"><i class="fa fa-check"></i><b>11</b> Bayesian nonparametric models</a>
<ul>
<li class="chapter" data-level="11.1" data-path="cha-bnp.html"><a href="cha-bnp.html#sec:bnpmixtures"><i class="fa fa-check"></i><b>11.1</b> Bayesian nonparametric mixture models</a></li>
<li class="chapter" data-level="11.2" data-path="cha-bnp.html"><a href="cha-bnp.html#sec:crp"><i class="fa fa-check"></i><b>11.2</b> Chinese Restaurant Process model</a>
<ul>
<li class="chapter" data-level="11.2.1" data-path="cha-bnp.html"><a href="cha-bnp.html#specification-and-density-2"><i class="fa fa-check"></i><b>11.2.1</b> Specification and density</a></li>
<li class="chapter" data-level="11.2.2" data-path="cha-bnp.html"><a href="cha-bnp.html#sec:excrp"><i class="fa fa-check"></i><b>11.2.2</b> Example</a></li>
<li class="chapter" data-level="11.2.3" data-path="cha-bnp.html"><a href="cha-bnp.html#sec:extensionscrp"><i class="fa fa-check"></i><b>11.2.3</b> Extensions</a></li>
</ul></li>
<li class="chapter" data-level="11.3" data-path="cha-bnp.html"><a href="cha-bnp.html#sec:sb"><i class="fa fa-check"></i><b>11.3</b> Stick-breaking model</a>
<ul>
<li class="chapter" data-level="11.3.1" data-path="cha-bnp.html"><a href="cha-bnp.html#specification-and-function"><i class="fa fa-check"></i><b>11.3.1</b> Specification and function</a></li>
<li class="chapter" data-level="11.3.2" data-path="cha-bnp.html"><a href="cha-bnp.html#sec:exsb"><i class="fa fa-check"></i><b>11.3.2</b> Example</a></li>
</ul></li>
<li class="chapter" data-level="11.4" data-path="cha-bnp.html"><a href="cha-bnp.html#mcmc-sampling-of-bnp-models"><i class="fa fa-check"></i><b>11.4</b> MCMC sampling of BNP models</a>
<ul>
<li class="chapter" data-level="11.4.1" data-path="cha-bnp.html"><a href="cha-bnp.html#sec:mcmcdcrp"><i class="fa fa-check"></i><b>11.4.1</b> Sampling CRP models</a></li>
<li class="chapter" data-level="11.4.2" data-path="cha-bnp.html"><a href="cha-bnp.html#sec:mcmcsb"><i class="fa fa-check"></i><b>11.4.2</b> Sampling stick-breaking models</a></li>
</ul></li>
</ul></li>
<li class="part"><span><b>IV Programming with NIMBLE</b></span></li>
<li class="chapter" data-level="" data-path="overview.html"><a href="overview.html"><i class="fa fa-check"></i>Overview</a></li>
<li class="chapter" data-level="12" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html"><i class="fa fa-check"></i><b>12</b> Writing simple nimbleFunctions</a>
<ul>
<li class="chapter" data-level="12.1" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:RC-intro"><i class="fa fa-check"></i><b>12.1</b> Introduction to simple nimbleFunctions</a></li>
<li class="chapter" data-level="12.2" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:r-fiunctions-implemented"><i class="fa fa-check"></i><b>12.2</b> R functions (or variants) implemented in NIMBLE</a>
<ul>
<li class="chapter" data-level="12.2.1" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#finding-help-for-nimbles-versions-of-r-functions"><i class="fa fa-check"></i><b>12.2.1</b> Finding help for NIMBLE’s versions of R functions</a></li>
<li class="chapter" data-level="12.2.2" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#basic-operations"><i class="fa fa-check"></i><b>12.2.2</b> Basic operations</a></li>
<li class="chapter" data-level="12.2.3" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:basic-math-linear"><i class="fa fa-check"></i><b>12.2.3</b> Math and linear algebra</a></li>
<li class="chapter" data-level="12.2.4" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:nimble-dist-funs"><i class="fa fa-check"></i><b>12.2.4</b> Distribution functions</a></li>
<li class="chapter" data-level="12.2.5" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:basic-flow-control"><i class="fa fa-check"></i><b>12.2.5</b> Flow control: <em>if-then-else</em>, <em>for</em>, <em>while</em>, and <em>stop</em></a></li>
<li class="chapter" data-level="12.2.6" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:print"><i class="fa fa-check"></i><b>12.2.6</b> <em>print</em> and <em>cat</em></a></li>
<li class="chapter" data-level="12.2.7" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:check-user-interr"><i class="fa fa-check"></i><b>12.2.7</b> Checking for user interrupts: <em>checkInterrupt</em></a></li>
<li class="chapter" data-level="12.2.8" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#optimization-optim-and-nimoptim"><i class="fa fa-check"></i><b>12.2.8</b> Optimization: <em>optim</em> and <em>nimOptim</em></a></li>
<li class="chapter" data-level="12.2.9" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#integration-integrate-and-nimintegrate"><i class="fa fa-check"></i><b>12.2.9</b> Integration: <em>integrate</em> and <em>nimIntegrate</em></a></li>
<li class="chapter" data-level="12.2.10" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:altern-keyw-some"><i class="fa fa-check"></i><b>12.2.10</b> ‘nim’ synonyms for some functions</a></li>
</ul></li>
<li class="chapter" data-level="12.3" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:how-nimble-handles"><i class="fa fa-check"></i><b>12.3</b> How NIMBLE handles types of variables</a>
<ul>
<li class="chapter" data-level="12.3.1" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:nimbleList-RCFuns"><i class="fa fa-check"></i><b>12.3.1</b> nimbleList data structures</a></li>
<li class="chapter" data-level="12.3.2" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:how-types-work"><i class="fa fa-check"></i><b>12.3.2</b> How numeric types work</a></li>
</ul></li>
<li class="chapter" data-level="12.4" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:decl-argum-return"><i class="fa fa-check"></i><b>12.4</b> Declaring argument and return types</a></li>
<li class="chapter" data-level="12.5" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:comp-nimbl-pass"><i class="fa fa-check"></i><b>12.5</b> Compiled nimbleFunctions pass arguments by reference</a></li>
<li class="chapter" data-level="12.6" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:calling-external-code"><i class="fa fa-check"></i><b>12.6</b> Calling external compiled code</a></li>
<li class="chapter" data-level="12.7" data-path="cha-RCfunctions.html"><a href="cha-RCfunctions.html#sec:calling-R-code"><i class="fa fa-check"></i><b>12.7</b> Calling uncompiled R functions from compiled nimbleFunctions</a></li>
</ul></li>
<li class="chapter" data-level="13" data-path="cha-user-defined.html"><a href="cha-user-defined.html"><i class="fa fa-check"></i><b>13</b> Creating user-defined distributions and functions for models</a>
<ul>
<li class="chapter" data-level="13.1" data-path="cha-user-defined.html"><a href="cha-user-defined.html#sec:user-functions"><i class="fa fa-check"></i><b>13.1</b> User-defined functions</a></li>
<li class="chapter" data-level="13.2" data-path="cha-user-defined.html"><a href="cha-user-defined.html#sec:user-distributions"><i class="fa fa-check"></i><b>13.2</b> User-defined distributions</a>
<ul>
<li class="chapter" data-level="13.2.1" data-path="cha-user-defined.html"><a href="cha-user-defined.html#sec:registerDistributions"><i class="fa fa-check"></i><b>13.2.1</b> Using <em>registerDistributions</em> for alternative parameterizations and providing other information</a></li>
</ul></li>
<li class="chapter" data-level="13.3" data-path="cha-user-defined.html"><a href="cha-user-defined.html#sec:adv-user-def"><i class="fa fa-check"></i><b>13.3</b> Advanced user-defined functions and distributions</a></li>
<li class="chapter" data-level="13.4" data-path="cha-user-defined.html"><a href="cha-user-defined.html#sec:user-macros"><i class="fa fa-check"></i><b>13.4</b> User-defined model macros</a></li>
</ul></li>
<li class="chapter" data-level="14" data-path="cha-using-models.html"><a href="cha-using-models.html"><i class="fa fa-check"></i><b>14</b> Working with NIMBLE models</a>
<ul>
<li class="chapter" data-level="14.1" data-path="cha-using-models.html"><a href="cha-using-models.html#sec:accessing-variables"><i class="fa fa-check"></i><b>14.1</b> The variables and nodes in a NIMBLE model</a>
<ul>
<li class="chapter" data-level="14.1.1" data-path="cha-using-models.html"><a href="cha-using-models.html#sec:get-nodes"><i class="fa fa-check"></i><b>14.1.1</b> Determining the nodes in a model</a></li>
<li class="chapter" data-level="14.1.2" data-path="cha-using-models.html"><a href="cha-using-models.html#sec:introduced-nodes"><i class="fa fa-check"></i><b>14.1.2</b> Understanding lifted nodes</a></li>
<li class="chapter" data-level="14.1.3" data-path="cha-using-models.html"><a href="cha-using-models.html#sec:cdgetdependencies"><i class="fa fa-check"></i><b>14.1.3</b> Determining dependencies in a model</a></li>
</ul></li>
<li class="chapter" data-level="14.2" data-path="cha-using-models.html"><a href="cha-using-models.html#sec:nodeInfo"><i class="fa fa-check"></i><b>14.2</b> Accessing information about nodes and variables</a>
<ul>
<li class="chapter" data-level="14.2.1" data-path="cha-using-models.html"><a href="cha-using-models.html#getting-distributional-information-about-a-node"><i class="fa fa-check"></i><b>14.2.1</b> Getting distributional information about a node</a></li>
<li class="chapter" data-level="14.2.2" data-path="cha-using-models.html"><a href="cha-using-models.html#getting-information-about-a-distribution"><i class="fa fa-check"></i><b>14.2.2</b> Getting information about a distribution</a></li>
<li class="chapter" data-level="14.2.3" data-path="cha-using-models.html"><a href="cha-using-models.html#sec:getParam"><i class="fa fa-check"></i><b>14.2.3</b> Getting distribution parameter values for a node</a></li>
<li class="chapter" data-level="14.2.4" data-path="cha-using-models.html"><a href="cha-using-models.html#sec:getBound"><i class="fa fa-check"></i><b>14.2.4</b> Getting distribution bounds for a node</a></li>
</ul></li>
<li class="chapter" data-level="14.3" data-path="cha-using-models.html"><a href="cha-using-models.html#sec:cdcalc-cdsim-cdgetl"><i class="fa fa-check"></i><b>14.3</b> Carrying out model calculations</a>
<ul>
<li class="chapter" data-level="14.3.1" data-path="cha-using-models.html"><a href="cha-using-models.html#core-model-operations-calculation-and-simulation"><i class="fa fa-check"></i><b>14.3.1</b> Core model operations: calculation and simulation</a></li>
<li class="chapter" data-level="14.3.2" data-path="cha-using-models.html"><a href="cha-using-models.html#sec:cdsimn-cdcalcn-cdget"><i class="fa fa-check"></i><b>14.3.2</b> Pre-defined nimbleFunctions for operating on model nodes: <em>simNodes</em>, <em>calcNodes</em>, and <em>getLogProbNodes</em></a></li>
<li class="chapter" data-level="14.3.3" data-path="cha-using-models.html"><a href="cha-using-models.html#sec:access-log-prob"><i class="fa fa-check"></i><b>14.3.3</b> Accessing log probabilities via <em>logProb</em> variables</a></li>
</ul></li>
</ul></li>
<li class="chapter" data-level="15" data-path="cha-data-structures.html"><a href="cha-data-structures.html"><i class="fa fa-check"></i><b>15</b> Data structures in NIMBLE</a>
<ul>
<li class="chapter" data-level="15.1" data-path="cha-data-structures.html"><a href="cha-data-structures.html#sec:modelValues-struct"><i class="fa fa-check"></i><b>15.1</b> The modelValues data structure</a>
<ul>
<li class="chapter" data-level="15.1.1" data-path="cha-data-structures.html"><a href="cha-data-structures.html#creating-modelvalues-objects"><i class="fa fa-check"></i><b>15.1.1</b> Creating modelValues objects</a></li>
<li class="chapter" data-level="15.1.2" data-path="cha-data-structures.html"><a href="cha-data-structures.html#sec:access-cont-modelv"><i class="fa fa-check"></i><b>15.1.2</b> Accessing contents of modelValues</a></li>
</ul></li>
<li class="chapter" data-level="15.2" data-path="cha-data-structures.html"><a href="cha-data-structures.html#sec:nimbleLists"><i class="fa fa-check"></i><b>15.2</b> The nimbleList data structure</a>
<ul>
<li class="chapter" data-level="15.2.1" data-path="cha-data-structures.html"><a href="cha-data-structures.html#sec:predef-nimbleLists"><i class="fa fa-check"></i><b>15.2.1</b> Pre-defined nimbleList types</a></li>
<li class="chapter" data-level="15.2.2" data-path="cha-data-structures.html"><a href="cha-data-structures.html#sec:eigen-nimFunctions"><i class="fa fa-check"></i><b>15.2.2</b> Using <em>eigen</em> and <em>svd</em> in nimbleFunctions</a></li>
</ul></li>
</ul></li>
<li class="chapter" data-level="16" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html"><i class="fa fa-check"></i><b>16</b> Writing nimbleFunctions to interact with models</a>
<ul>
<li class="chapter" data-level="16.1" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:writ-nimble-funct"><i class="fa fa-check"></i><b>16.1</b> Overview</a></li>
<li class="chapter" data-level="16.2" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:using-comp-nimbl"><i class="fa fa-check"></i><b>16.2</b> Using and compiling nimbleFunctions</a></li>
<li class="chapter" data-level="16.3" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#writing-setup-code"><i class="fa fa-check"></i><b>16.3</b> Writing setup code</a>
<ul>
<li class="chapter" data-level="16.3.1" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#useful-tools-for-setup-functions"><i class="fa fa-check"></i><b>16.3.1</b> Useful tools for setup functions</a></li>
<li class="chapter" data-level="16.3.2" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:access-modify-numer"><i class="fa fa-check"></i><b>16.3.2</b> Accessing and modifying numeric values from setup</a></li>
<li class="chapter" data-level="16.3.3" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#determining-numeric-types-in-nimblefunctions"><i class="fa fa-check"></i><b>16.3.3</b> Determining numeric types in nimbleFunctions</a></li>
<li class="chapter" data-level="16.3.4" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:determ-pers-texttts"><i class="fa fa-check"></i><b>16.3.4</b> Control of setup outputs</a></li>
</ul></li>
<li class="chapter" data-level="16.4" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:nimble-lang-comp"><i class="fa fa-check"></i><b>16.4</b> Writing run code</a>
<ul>
<li class="chapter" data-level="16.4.1" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:driv-models:-calc"><i class="fa fa-check"></i><b>16.4.1</b> Driving models: <em>calculate</em>, <em>calculateDiff</em>, <em>simulate</em>, <em>getLogProb</em></a></li>
<li class="chapter" data-level="16.4.2" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#getting-and-setting-variable-and-node-values"><i class="fa fa-check"></i><b>16.4.2</b> Getting and setting variable and node values</a></li>
<li class="chapter" data-level="16.4.3" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#getting-parameter-values-and-node-bounds"><i class="fa fa-check"></i><b>16.4.3</b> Getting parameter values and node bounds</a></li>
<li class="chapter" data-level="16.4.4" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:access-model-modelv"><i class="fa fa-check"></i><b>16.4.4</b> Using modelValues objects</a></li>
<li class="chapter" data-level="16.4.5" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:using-model-variable"><i class="fa fa-check"></i><b>16.4.5</b> Using model variables and modelValues in expressions</a></li>
<li class="chapter" data-level="16.4.6" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:incl-other-meth"><i class="fa fa-check"></i><b>16.4.6</b> Including other methods in a nimbleFunction</a></li>
<li class="chapter" data-level="16.4.7" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:using-other-nimbl"><i class="fa fa-check"></i><b>16.4.7</b> Using other nimbleFunctions</a></li>
<li class="chapter" data-level="16.4.8" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:virt-nimbl-nimbl"><i class="fa fa-check"></i><b>16.4.8</b> Virtual nimbleFunctions and nimbleFunctionLists</a></li>
<li class="chapter" data-level="16.4.9" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#character-objects"><i class="fa fa-check"></i><b>16.4.9</b> Character objects</a></li>
<li class="chapter" data-level="16.4.10" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:user-defined-data"><i class="fa fa-check"></i><b>16.4.10</b> User-defined data structures</a></li>
</ul></li>
<li class="chapter" data-level="16.5" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:user-samplers"><i class="fa fa-check"></i><b>16.5</b> Example: writing user-defined samplers to extend NIMBLE’s MCMC engine</a>
<ul>
<li class="chapter" data-level="16.5.1" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#user-defined-samplers-and-posterior-predictive-nodes"><i class="fa fa-check"></i><b>16.5.1</b> User-defined samplers and posterior predictive nodes</a></li>
</ul></li>
<li class="chapter" data-level="16.6" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#copying-nimblefunctions-and-nimble-models"><i class="fa fa-check"></i><b>16.6</b> Copying nimbleFunctions (and NIMBLE models)</a></li>
<li class="chapter" data-level="16.7" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#sec:debugging"><i class="fa fa-check"></i><b>16.7</b> Debugging nimbleFunctions</a></li>
<li class="chapter" data-level="16.8" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#timing-nimblefunctions-with-run.time"><i class="fa fa-check"></i><b>16.8</b> Timing nimbleFunctions with <em>run.time</em></a></li>
<li class="chapter" data-level="16.9" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#clearing-and-unloading-compiled-objects"><i class="fa fa-check"></i><b>16.9</b> Clearing and unloading compiled objects</a></li>
<li class="chapter" data-level="16.10" data-path="cha-progr-with-models.html"><a href="cha-progr-with-models.html#reducing-memory-usage"><i class="fa fa-check"></i><b>16.10</b> Reducing memory usage</a></li>
</ul></li>
<li class="part"><span><b>V Automatic Derivatives in NIMBLE</b></span></li>
<li class="chapter" data-level="17" data-path="cha-AD.html"><a href="cha-AD.html"><i class="fa fa-check"></i><b>17</b> Automatic Derivatives</a>
<ul>
<li class="chapter" data-level="17.1" data-path="cha-AD.html"><a href="cha-AD.html#sec:use-derivs"><i class="fa fa-check"></i><b>17.1</b> How to turn on derivatives in a model</a></li>
<li class="chapter" data-level="17.2" data-path="cha-AD.html"><a href="cha-AD.html#sec:AD-user-def"><i class="fa fa-check"></i><b>17.2</b> How to support derivatives in user-defined functions and distributions</a></li>
<li class="chapter" data-level="17.3" data-path="cha-AD.html"><a href="cha-AD.html#what-operations-are-and-arent-supported-for-ad"><i class="fa fa-check"></i><b>17.3</b> What operations are and aren’t supported for AD</a></li>
<li class="chapter" data-level="17.4" data-path="cha-AD.html"><a href="cha-AD.html#basics-of-obtaining-derivatives-in-nimblefunctions"><i class="fa fa-check"></i><b>17.4</b> Basics of obtaining derivatives in <code>nimbleFunctions</code></a>
<ul>
<li class="chapter" data-level="17.4.1" data-path="cha-AD.html"><a href="cha-AD.html#checking-derivatives-with-uncompiled-execution"><i class="fa fa-check"></i><b>17.4.1</b> Checking derivatives with uncompiled execution</a></li>
<li class="chapter" data-level="17.4.2" data-path="cha-AD.html"><a href="cha-AD.html#sec:AD-holding-out"><i class="fa fa-check"></i><b>17.4.2</b> Holding some local variables out of derivative tracking</a></li>
<li class="chapter" data-level="17.4.3" data-path="cha-AD.html"><a href="cha-AD.html#sec:AD-multiple-NF"><i class="fa fa-check"></i><b>17.4.3</b> Using AD with multiple nimbleFunctions</a></li>
<li class="chapter" data-level="17.4.4" data-path="cha-AD.html"><a href="cha-AD.html#sec:understanding-more-AD"><i class="fa fa-check"></i><b>17.4.4</b> Understanding more about how AD works: <em>taping</em> of operations</a></li>
<li class="chapter" data-level="17.4.5" data-path="cha-AD.html"><a href="cha-AD.html#resetting-a-nimderivs-call"><i class="fa fa-check"></i><b>17.4.5</b> Resetting a <code>nimDerivs</code> call</a></li>
<li class="chapter" data-level="17.4.6" data-path="cha-AD.html"><a href="cha-AD.html#a-note-on-performance-benchmarking"><i class="fa fa-check"></i><b>17.4.6</b> A note on performance benchmarking</a></li>
</ul></li>
<li class="chapter" data-level="17.5" data-path="cha-AD.html"><a href="cha-AD.html#advanced-uses-double-taping"><i class="fa fa-check"></i><b>17.5</b> Advanced uses: double taping</a></li>
<li class="chapter" data-level="17.6" data-path="cha-AD.html"><a href="cha-AD.html#derivatives-involving-model-calculations"><i class="fa fa-check"></i><b>17.6</b> Derivatives involving model calculations</a>
<ul>
<li class="chapter" data-level="17.6.1" data-path="cha-AD.html"><a href="cha-AD.html#method-1-nimderivs-of-modelcalculate"><i class="fa fa-check"></i><b>17.6.1</b> Method 1: <code>nimDerivs</code> of <code>model$calculate</code></a></li>
<li class="chapter" data-level="17.6.2" data-path="cha-AD.html"><a href="cha-AD.html#method-2-nimderivs-of-a-method-that-calls-modelcalculate"><i class="fa fa-check"></i><b>17.6.2</b> Method 2: <code>nimDerivs</code> of a method that calls <code>model$calculate</code></a></li>
</ul></li>
<li class="chapter" data-level="17.7" data-path="cha-AD.html"><a href="cha-AD.html#sec:parameter-transform"><i class="fa fa-check"></i><b>17.7</b> Parameter transformations</a></li>
</ul></li>
<li class="chapter" data-level="18" data-path="example-maximum-likelihood-estimation-using-optim-with-gradients-from-nimderivs..html"><a href="example-maximum-likelihood-estimation-using-optim-with-gradients-from-nimderivs..html"><i class="fa fa-check"></i><b>18</b> Example: maximum likelihood estimation using <code>optim</code> with gradients from <code>nimDerivs</code>.</a></li>
<li class="chapter" data-level="" data-path="references.html"><a href="references.html"><i class="fa fa-check"></i>References</a></li>
</ul>

      </nav>
    </div>

    <div class="book-body">
      <div class="body-inner">
        <div class="book-header" role="navigation">
          <h1>
            <i class="fa fa-circle-o-notch fa-spin"></i><a href="./"></a>
          </h1>
        </div>

        <div class="page-wrapper" tabindex="-1" role="main">
          <div class="page-inner">

            <section class="normal" id="section-">
<div id="cha-progr-with-models" class="section level1 hasAnchor" number="16">
<h1><span class="header-section-number">Chapter 16</span> Writing nimbleFunctions to interact with models<a href="cha-progr-with-models.html#cha-progr-with-models" class="anchor-section" aria-label="Anchor link to header"></a></h1>
<div id="sec:writ-nimble-funct" class="section level2 hasAnchor" number="16.1">
<h2><span class="header-section-number">16.1</span> Overview<a href="cha-progr-with-models.html#sec:writ-nimble-funct" class="anchor-section" aria-label="Anchor link to header"></a></h2>
<p>When you write an R function, you say what the input arguments are,
you provide the code for execution, and in that code you give the
value to be returned<a href="#fn29" class="footnote-ref" id="fnref29"><sup>29</sup></a>. Using the
<code>function</code> keyword in R triggers the operation of
creating an object that is the function.</p>
<p>Creating nimbleFunctions is similar, but there
are two kinds of code and two steps of execution:</p>
<ol style="list-style-type: decimal">
<li><p><em>Setup</em> code is provided as a regular R function, but the
programmer does not control what it returns. Typically the inputs to
<em>setup</em> code are objects like a model, a vector of nodes,
a modelValues object or a modelValues configuration, or another nimbleFunction. The setup code,
as its name implies, sets up information for run-time code. It is
executed in R, so it can use any aspect of R.</p></li>
<li><p><em>Run</em> code is provided in the NIMBLE language, which was
introduced in Chapter <a href="cha-RCfunctions.html#cha-RCfunctions">12</a>. This is
similar to a narrow subset of R, but it is important to remember
that it is different – defined by what can be compiled – and much more limited. <em>Run</em> code can
use the objects created by the <em>setup</em> code. In addition,
some information on variable types must be provided for input
arguments, the return value, and in some circumstances for local
variables. There are two
kinds of <em>run</em> code:</p>
<ol style="list-style-type: lower-alpha">
<li>There is always a primary function, given as the argument <code>run</code><a href="#fn30" class="footnote-ref" id="fnref30"><sup>30</sup></a>.</li>
<li>There can optionally be other functions, or ‘methods’ in the
language of object-oriented programming, that share the same
objects created by the <em>setup</em> function.</li>
</ol></li>
</ol>
<p>Here is a small example to fix ideas:</p>
<div class="sourceCode" id="cb539"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb539-1"><a href="cha-progr-with-models.html#cb539-1" tabindex="-1"></a>logProbCalcPlus <span class="ot">&lt;-</span> <span class="fu">nimbleFunction</span>(</span>
<span id="cb539-2"><a href="cha-progr-with-models.html#cb539-2" tabindex="-1"></a>    <span class="at">setup =</span> <span class="cf">function</span>(model, node) {</span>
<span id="cb539-3"><a href="cha-progr-with-models.html#cb539-3" tabindex="-1"></a>        dependentNodes <span class="ot">&lt;-</span> model<span class="sc">$</span><span class="fu">getDependencies</span>(node)</span>
<span id="cb539-4"><a href="cha-progr-with-models.html#cb539-4" tabindex="-1"></a>        valueToAdd <span class="ot">&lt;-</span> <span class="dv">1</span></span>
<span id="cb539-5"><a href="cha-progr-with-models.html#cb539-5" tabindex="-1"></a>    },</span>
<span id="cb539-6"><a href="cha-progr-with-models.html#cb539-6" tabindex="-1"></a>    <span class="at">run =</span> <span class="cf">function</span>(<span class="at">P =</span> <span class="fu">double</span>(<span class="dv">0</span>)) {</span>
<span id="cb539-7"><a href="cha-progr-with-models.html#cb539-7" tabindex="-1"></a>        model[[node]] <span class="ot">&lt;&lt;-</span> P <span class="sc">+</span> valueToAdd</span>
<span id="cb539-8"><a href="cha-progr-with-models.html#cb539-8" tabindex="-1"></a>        <span class="fu">return</span>(model<span class="sc">$</span><span class="fu">calculate</span>(dependentNodes))</span>
<span id="cb539-9"><a href="cha-progr-with-models.html#cb539-9" tabindex="-1"></a>        <span class="fu">returnType</span>(<span class="fu">double</span>(<span class="dv">0</span>))</span>
<span id="cb539-10"><a href="cha-progr-with-models.html#cb539-10" tabindex="-1"></a>    })</span>
<span id="cb539-11"><a href="cha-progr-with-models.html#cb539-11" tabindex="-1"></a></span>
<span id="cb539-12"><a href="cha-progr-with-models.html#cb539-12" tabindex="-1"></a>code <span class="ot">&lt;-</span> <span class="fu">nimbleCode</span>({</span>
<span id="cb539-13"><a href="cha-progr-with-models.html#cb539-13" tabindex="-1"></a>    a <span class="sc">~</span> <span class="fu">dnorm</span>(<span class="dv">0</span>, <span class="dv">1</span>)</span>
<span id="cb539-14"><a href="cha-progr-with-models.html#cb539-14" tabindex="-1"></a>    b <span class="sc">~</span> <span class="fu">dnorm</span>(a, <span class="dv">1</span>)</span>
<span id="cb539-15"><a href="cha-progr-with-models.html#cb539-15" tabindex="-1"></a>})</span>
<span id="cb539-16"><a href="cha-progr-with-models.html#cb539-16" tabindex="-1"></a>testModel <span class="ot">&lt;-</span> <span class="fu">nimbleModel</span>(code, <span class="at">check =</span> <span class="cn">FALSE</span>)</span>
<span id="cb539-17"><a href="cha-progr-with-models.html#cb539-17" tabindex="-1"></a>logProbCalcPlusA <span class="ot">&lt;-</span> <span class="fu">logProbCalcPlus</span>(testModel, <span class="st">&quot;a&quot;</span>)</span>
<span id="cb539-18"><a href="cha-progr-with-models.html#cb539-18" tabindex="-1"></a>testModel<span class="sc">$</span>b <span class="ot">&lt;-</span> <span class="fl">1.5</span></span>
<span id="cb539-19"><a href="cha-progr-with-models.html#cb539-19" tabindex="-1"></a>logProbCalcPlusA<span class="sc">$</span><span class="fu">run</span>(<span class="fl">0.25</span>) </span></code></pre></div>
<pre><code>## [1] -2.650377</code></pre>
<div class="sourceCode" id="cb541"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb541-1"><a href="cha-progr-with-models.html#cb541-1" tabindex="-1"></a><span class="fu">dnorm</span>(<span class="fl">1.25</span>,<span class="dv">0</span>,<span class="dv">1</span>,<span class="cn">TRUE</span>)<span class="sc">+</span><span class="fu">dnorm</span>(<span class="fl">1.5</span>,<span class="fl">1.25</span>,<span class="dv">1</span>,<span class="cn">TRUE</span>) <span class="co"># direct validation</span></span></code></pre></div>
<pre><code>## [1] -2.650377</code></pre>
<div class="sourceCode" id="cb543"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb543-1"><a href="cha-progr-with-models.html#cb543-1" tabindex="-1"></a>testModel<span class="sc">$</span>a  <span class="co"># &quot;a&quot; was set to 0.5 + valueToAdd</span></span></code></pre></div>
<pre><code>## [1] 1.25</code></pre>
<p>The call to the R function called <code>nimbleFunction</code> returns a
function, similarly to defining a function in R. That function,
<code>logProbCalcPlus</code>, takes arguments for its <code>setup</code> function,
executes it, and returns an object, <code>logProbCalcPlusA</code>, that has a
<em>run</em> member function (method) accessed by <code>$run</code>. In this case, the
<code>setup</code> function obtains the stochastic dependencies of the
<code>node</code> using the <code>getDependencies</code> member function of the model
(see Section <a href="cha-using-models.html#sec:cdgetdependencies">14.1.3</a>) and stores them in
<code>dependentNodes</code>. In this way, <code>logProbCalcPlus</code> can adapt to any
model. It also creates a variable, <code>valueToAdd</code>, that can be used by the nimbleFunction.</p>
<p>The object <code>logProbCalcPlusA</code>, returned by <code>logProbCalcPlus</code>,
is permanently bound to the results of the processed <code>setup</code>
function. In this case, <code>logProbCalcPlusA$run</code> takes a scalar input value, <code>P</code>,
assigns <code>P + valueToAdd</code> to
the given node in the model, and returns the sum of the log
probabilities of that node and its stochastic
dependencies<a href="#fn31" class="footnote-ref" id="fnref31"><sup>31</sup></a>. We say <code>logProbCalcPlusA</code> is an ‘instance’ of
<code>logProbCalcPlus</code> that is
‘specialized’ or ‘bound’ to <code>a</code> and <code>testModel</code>. Usually, the
<code>setup</code> code will be where information about the model
structure is determined, and then the <code>run</code> code can use that
information without repeatedly, redundantly recomputing it. A
nimbleFunction can
be called repeatedly (one can think of it as a generator), each time returning a specialized
nimbleFunction.</p>
<p>Readers familiar with object-oriented programming may find it useful
to think in terms of class definitions and objects. <code>nimbleFunction</code>
creates a class definition. Each specialized nimbleFunction is one object
in the class. The setup arguments are used to define member data in
the object.</p>
</div>
<div id="sec:using-comp-nimbl" class="section level2 hasAnchor" number="16.2">
<h2><span class="header-section-number">16.2</span> Using and compiling nimbleFunctions<a href="cha-progr-with-models.html#sec:using-comp-nimbl" class="anchor-section" aria-label="Anchor link to header"></a></h2>
<p>To compile the nimbleFunction, together with its model, we use <code>compileNimble</code>:</p>
<div class="sourceCode" id="cb545"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb545-1"><a href="cha-progr-with-models.html#cb545-1" tabindex="-1"></a>CnfDemo <span class="ot">&lt;-</span> <span class="fu">compileNimble</span>(testModel, logProbCalcPlusA)</span>
<span id="cb545-2"><a href="cha-progr-with-models.html#cb545-2" tabindex="-1"></a>CtestModel <span class="ot">&lt;-</span> CnfDemo<span class="sc">$</span>testModel</span>
<span id="cb545-3"><a href="cha-progr-with-models.html#cb545-3" tabindex="-1"></a>ClogProbCalcPlusA <span class="ot">&lt;-</span> CnfDemo<span class="sc">$</span>logProbCalcPlusA</span></code></pre></div>
<p>These have been initialized with the values from their uncompiled
versions and can be used in the same way:</p>
<div class="sourceCode" id="cb546"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb546-1"><a href="cha-progr-with-models.html#cb546-1" tabindex="-1"></a>CtestModel<span class="sc">$</span>a      <span class="co"># values were initialized from testModel</span></span></code></pre></div>
<pre><code>## [1] 1.25</code></pre>
<div class="sourceCode" id="cb548"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb548-1"><a href="cha-progr-with-models.html#cb548-1" tabindex="-1"></a>CtestModel<span class="sc">$</span>b</span></code></pre></div>
<pre><code>## [1] 1.5</code></pre>
<div class="sourceCode" id="cb550"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb550-1"><a href="cha-progr-with-models.html#cb550-1" tabindex="-1"></a>lpA <span class="ot">&lt;-</span> ClogProbCalcPlusA<span class="sc">$</span><span class="fu">run</span>(<span class="fl">1.5</span>)</span>
<span id="cb550-2"><a href="cha-progr-with-models.html#cb550-2" tabindex="-1"></a>lpA</span></code></pre></div>
<pre><code>## [1] -5.462877</code></pre>
<div class="sourceCode" id="cb552"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb552-1"><a href="cha-progr-with-models.html#cb552-1" tabindex="-1"></a><span class="co"># verify the answer:</span></span>
<span id="cb552-2"><a href="cha-progr-with-models.html#cb552-2" tabindex="-1"></a><span class="fu">dnorm</span>(CtestModel<span class="sc">$</span>b, CtestModel<span class="sc">$</span>a, <span class="dv">1</span>, <span class="at">log =</span> <span class="cn">TRUE</span>) <span class="sc">+</span> </span>
<span id="cb552-3"><a href="cha-progr-with-models.html#cb552-3" tabindex="-1"></a>    <span class="fu">dnorm</span>(CtestModel<span class="sc">$</span>a, <span class="dv">0</span>, <span class="dv">1</span>, <span class="at">log =</span> <span class="cn">TRUE</span>) </span></code></pre></div>
<pre><code>## [1] -5.462877</code></pre>
<div class="sourceCode" id="cb554"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb554-1"><a href="cha-progr-with-models.html#cb554-1" tabindex="-1"></a>CtestModel<span class="sc">$</span>a       <span class="co"># a was modified in the compiled model</span></span></code></pre></div>
<pre><code>## [1] 2.5</code></pre>
<div class="sourceCode" id="cb556"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb556-1"><a href="cha-progr-with-models.html#cb556-1" tabindex="-1"></a>testModel<span class="sc">$</span>a        <span class="co"># the uncompiled model was not modified</span></span></code></pre></div>
<pre><code>## [1] 1.25</code></pre>
</div>
<div id="writing-setup-code" class="section level2 hasAnchor" number="16.3">
<h2><span class="header-section-number">16.3</span> Writing setup code<a href="cha-progr-with-models.html#writing-setup-code" class="anchor-section" aria-label="Anchor link to header"></a></h2>
<div id="useful-tools-for-setup-functions" class="section level3 hasAnchor" number="16.3.1">
<h3><span class="header-section-number">16.3.1</span> Useful tools for setup functions<a href="cha-progr-with-models.html#useful-tools-for-setup-functions" class="anchor-section" aria-label="Anchor link to header"></a></h3>
<p>The setup function is typically used to determine information on nodes
in a model, set up modelValues or nimbleList objects, set up (nested) nimbleFunctions
or nimbleFunctionLists, and set up any persistent numeric objects.
For example, the setup code of an MCMC nimbleFunction creates the
nimbleFunctionList of sampler nimbleFunctions. The values of numeric
objects created in setup code can be modified by run code and will
persist across calls.</p>
<p>Some of the useful tools and objects to create in setup functions include:</p>
<ul>
<li><strong>vectors of node names, often from a model</strong> Often these are obtained from the
<code>getNodeNames</code>, <code>getDependencies</code>, and other methods of a model,
described in Sections <a href="cha-using-models.html#sec:accessing-variables">14.1</a>-<a href="cha-using-models.html#sec:nodeInfo">14.2</a>.</li>
<li><strong>modelValues objects</strong> These are discussed in Sections <a href="cha-data-structures.html#sec:modelValues-struct">15.1</a> and <a href="cha-progr-with-models.html#sec:access-model-modelv">16.4.4</a>.</li>
<li><strong>nimbleList objects</strong> New instances of <code>nimbleList</code>s can
be created from a nimbleList definition in either setup or run code. See Section <a href="cha-data-structures.html#sec:nimbleLists">15.2</a> for more information.
<!---  Comment: It is actually sketchy that we allow a nimbleList
  % definition to be created in setup code, because it would  have to
 be the same definition always used by all specializations.  So
  % I've put the emphasis here on creating nimbleList objects. -PdV --></li>
<li><strong>specializations of other nimbleFunctions</strong> A useful NIMBLE
programming technique is to have one nimbleFunction contain other
nimbleFunctions, which it can use in its run-time code (Section <a href="cha-progr-with-models.html#sec:using-other-nimbl">16.4.7</a>).</li>
<li><strong>lists of other nimbleFunctions</strong> In addition to containing single
other nimbleFunctions, a nimbleFunction can contain a list of other
nimbleFunctions (Section <a href="cha-progr-with-models.html#sec:virt-nimbl-nimbl">16.4.8</a>).</li>
</ul>
<p>If one wants a nimbleFunction that does get specialized but has
empty setup code, use <code>setup = function() {}</code> or <code>setup = TRUE</code>.</p>
</div>
<div id="sec:access-modify-numer" class="section level3 hasAnchor" number="16.3.2">
<h3><span class="header-section-number">16.3.2</span> Accessing and modifying numeric values from setup<a href="cha-progr-with-models.html#sec:access-modify-numer" class="anchor-section" aria-label="Anchor link to header"></a></h3>
<p>While models and nodes created during setup cannot be
modified<a href="#fn32" class="footnote-ref" id="fnref32"><sup>32</sup></a>,
numeric values and modelValues can be, as illustrated by extending the example from above.</p>
<div class="sourceCode" id="cb558"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb558-1"><a href="cha-progr-with-models.html#cb558-1" tabindex="-1"></a>logProbCalcPlusA<span class="sc">$</span>valueToAdd  <span class="co"># in the uncompiled version</span></span></code></pre></div>
<pre><code>## [1] 1</code></pre>
<div class="sourceCode" id="cb560"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb560-1"><a href="cha-progr-with-models.html#cb560-1" tabindex="-1"></a>logProbCalcPlusA<span class="sc">$</span>valueToAdd <span class="ot">&lt;-</span> <span class="dv">2</span></span>
<span id="cb560-2"><a href="cha-progr-with-models.html#cb560-2" tabindex="-1"></a>ClogProbCalcPlusA<span class="sc">$</span>valueToAdd  <span class="co"># or in the compiled version</span></span></code></pre></div>
<pre><code>## [1] 1</code></pre>
<div class="sourceCode" id="cb562"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb562-1"><a href="cha-progr-with-models.html#cb562-1" tabindex="-1"></a>ClogProbCalcPlusA<span class="sc">$</span>valueToAdd <span class="ot">&lt;-</span> <span class="dv">3</span></span>
<span id="cb562-2"><a href="cha-progr-with-models.html#cb562-2" tabindex="-1"></a>ClogProbCalcPlusA<span class="sc">$</span><span class="fu">run</span>(<span class="fl">1.5</span>)</span></code></pre></div>
<pre><code>## [1] -16.46288</code></pre>
<div class="sourceCode" id="cb564"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb564-1"><a href="cha-progr-with-models.html#cb564-1" tabindex="-1"></a>CtestModel<span class="sc">$</span>a     <span class="co"># a == 1.5 + 3</span></span></code></pre></div>
<pre><code>## [1] 4.5</code></pre>
</div>
<div id="determining-numeric-types-in-nimblefunctions" class="section level3 hasAnchor" number="16.3.3">
<h3><span class="header-section-number">16.3.3</span> Determining numeric types in nimbleFunctions<a href="cha-progr-with-models.html#determining-numeric-types-in-nimblefunctions" class="anchor-section" aria-label="Anchor link to header"></a></h3>
<p>For numeric variables from the <code>setup</code> function that
appear in the <code>run</code> function or other member functions (or
are declared in <code>setupOutputs</code>), the
type is determined from the values created by the setup
code. The types created by setup code must be
consistent across all specializations of the nimbleFunction. For
example if <code>X</code> is created as a matrix (two-dimensional double) in one
specialization but as a vector (one-dimensional double) in another, there
will be a problem during compilation. The sizes may differ in each specialization.</p>
<p>Treatment of vectors of length one presents special challenges because
they could be treated as scalars or vectors. Currently they are
treated as scalars. If you want a vector, ensure that the length is
greater than one in the setup code and then use <code>setSize</code> in the
run-time code.</p>
</div>
<div id="sec:determ-pers-texttts" class="section level3 hasAnchor" number="16.3.4">
<h3><span class="header-section-number">16.3.4</span> Control of setup outputs<a href="cha-progr-with-models.html#sec:determ-pers-texttts" class="anchor-section" aria-label="Anchor link to header"></a></h3>
<p>Sometimes setup code may create variables that are not used in
run code. By default, NIMBLE inspects run code and omits
variables from setup that do not appear in run code from
compilation. However, sometimes a programmer may want to force a
numeric or character variable to be included in compilation, even if it
is not used directly in run code. As shown below, such variables
can be directly accessed in one nimbleFunction from another, which
provides a way of using nimbleFunctions as general data structures.
To force NIMBLE to include variables during compilation, for
example <code>X</code> and <code>Y</code>, simply include</p>
<div class="sourceCode" id="cb566"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb566-1"><a href="cha-progr-with-models.html#cb566-1" tabindex="-1"></a><span class="fu">setupOutputs</span>(X, Y)</span></code></pre></div>
<p>anywhere in the setup code.</p>
</div>
</div>
<div id="sec:nimble-lang-comp" class="section level2 hasAnchor" number="16.4">
<h2><span class="header-section-number">16.4</span> Writing run code<a href="cha-progr-with-models.html#sec:nimble-lang-comp" class="anchor-section" aria-label="Anchor link to header"></a></h2>
<p>In Chapter <a href="cha-RCfunctions.html#cha-RCfunctions">12</a> we described the functionality of the NIMBLE language that could be used in run code without setup code (typically in cases where no models or modelValues are needed). Next we explain the additional features that allow use of models and modelValues in the run code.</p>
<div id="sec:driv-models:-calc" class="section level3 hasAnchor" number="16.4.1">
<h3><span class="header-section-number">16.4.1</span> Driving models: <em>calculate</em>, <em>calculateDiff</em>, <em>simulate</em>, <em>getLogProb</em><a href="cha-progr-with-models.html#sec:driv-models:-calc" class="anchor-section" aria-label="Anchor link to header"></a></h3>
<p>These four functions are the primary ways to operate a model. Their
syntax was explained in Section <a href="cha-using-models.html#sec:cdcalc-cdsim-cdgetl">14.3</a>. Except
for <code>getLogProb</code>, it is usually important for the <code>nodes</code>
vector to be sorted in
topological order. Model member functions such as <code>getDependencies</code> and
<code>expandNodeNames</code> will
always return topoligically sorted node names.</p>
<p>Most R-like indexing of a node vector is allowed within the
argument to <code>calculate</code>, <code>calculateDiff</code>, <code>simulate</code>, and <code>getLogProb</code>. For example, all of the following are allowed:</p>
<div class="sourceCode" id="cb567"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb567-1"><a href="cha-progr-with-models.html#cb567-1" tabindex="-1"></a>myModel<span class="sc">$</span><span class="fu">calculate</span>(nodes)</span>
<span id="cb567-2"><a href="cha-progr-with-models.html#cb567-2" tabindex="-1"></a>myModel<span class="sc">$</span><span class="fu">calculate</span>(nodes[i])</span>
<span id="cb567-3"><a href="cha-progr-with-models.html#cb567-3" tabindex="-1"></a>myModel<span class="sc">$</span><span class="fu">calculate</span>(nodes[<span class="dv">1</span><span class="sc">:</span><span class="dv">3</span>])</span>
<span id="cb567-4"><a href="cha-progr-with-models.html#cb567-4" tabindex="-1"></a>myModel<span class="sc">$</span><span class="fu">calculate</span>(nodes[<span class="fu">c</span>(<span class="dv">1</span>,<span class="dv">3</span>)])</span>
<span id="cb567-5"><a href="cha-progr-with-models.html#cb567-5" tabindex="-1"></a>myModel<span class="sc">$</span><span class="fu">calculate</span>(nodes[<span class="dv">2</span><span class="sc">:</span>i])</span>
<span id="cb567-6"><a href="cha-progr-with-models.html#cb567-6" tabindex="-1"></a>myModel<span class="sc">$</span><span class="fu">calculate</span>(nodes[ <span class="fu">values</span>(model, nodes) <span class="sc">+</span> <span class="fl">0.1</span> <span class="sc">&lt;</span> x ])</span></code></pre></div>
<p>Note that in the last line of code, one must have that the length of <code>nodes</code> is
equal to that of <code>values(model, nodes)</code>, which means that all the nodes in <code>nodes</code>
must be scalar nodes.</p>
<p>Also note that one cannot create new vectors of nodes in run code. They
can only be indexed within a call to <code>calculate</code>, <code>calculateDiff</code>,
<code>simulate</code> or <code>getLogProb</code>.</p>
</div>
<div id="getting-and-setting-variable-and-node-values" class="section level3 hasAnchor" number="16.4.2">
<h3><span class="header-section-number">16.4.2</span> Getting and setting variable and node values<a href="cha-progr-with-models.html#getting-and-setting-variable-and-node-values" class="anchor-section" aria-label="Anchor link to header"></a></h3>
<div id="using-indexing-with-nodes" class="section level4 hasAnchor" number="16.4.2.1">
<h4><span class="header-section-number">16.4.2.1</span> Using indexing with nodes<a href="cha-progr-with-models.html#using-indexing-with-nodes" class="anchor-section" aria-label="Anchor link to header"></a></h4>
<p>Here is an example that illustrates getting and setting of nodes,
subsets of nodes, or variables. Note the following:</p>
<ul>
<li>In <code>model[[v]]</code>, <code>v</code> can only be a single node or variable name, not a vector of multiple nodes nor an element of such a vector (<code>model[[ nodes[i] ]]</code> does not work). The node itself may be a vector, matrix or array node.</li>
<li>In fact, <code>v</code> can be a node-name-like character string, even if it is not actually a node in the model. See example 4 in the code below.</li>
<li>One can also use <code>model$varName</code>, with the caveat that <code>varName</code> must be a variable name. This usage would only make sense for a nimbleFunction written for models known to have a specific variable. (Note that if <code>a</code> is a scalar node in <code>model</code>, then <code>model[['a']]</code> will be a scalar but <code>model$a</code> will be a vector of length 1.</li>
<li>one should use the <code>&lt;&lt;-</code> global assignment operator to assign into model nodes.</li>
</ul>
<p>Note that NIMBLE does not allow variables to change dimensions. Model nodes are the same, and indeed are more restricted because they can’t change sizes. In addition, NIMBLE distinguishes between scalars and vectors of length 1. These rules, and ways to handle them correctly, are illustrated in the following code as well as in Section <a href="cha-RCfunctions.html#sec:how-nimble-handles">12.3</a>.</p>
<div class="sourceCode" id="cb568"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb568-1"><a href="cha-progr-with-models.html#cb568-1" tabindex="-1"></a>code <span class="ot">&lt;-</span> <span class="fu">nimbleCode</span>({</span>
<span id="cb568-2"><a href="cha-progr-with-models.html#cb568-2" tabindex="-1"></a>    z <span class="sc">~</span> <span class="fu">dnorm</span>(<span class="dv">0</span>, <span class="at">sd =</span> sigma)</span>
<span id="cb568-3"><a href="cha-progr-with-models.html#cb568-3" tabindex="-1"></a>    sigma <span class="sc">~</span> <span class="fu">dunif</span>(<span class="dv">0</span>, <span class="dv">10</span>)</span>
<span id="cb568-4"><a href="cha-progr-with-models.html#cb568-4" tabindex="-1"></a>    y[<span class="dv">1</span><span class="sc">:</span>n] <span class="sc">~</span> <span class="fu">dmnorm</span>(zeroes[<span class="dv">1</span><span class="sc">:</span>n], <span class="at">cov =</span> C[<span class="dv">1</span><span class="sc">:</span><span class="dv">5</span>, <span class="dv">1</span><span class="sc">:</span><span class="dv">5</span>])</span>
<span id="cb568-5"><a href="cha-progr-with-models.html#cb568-5" tabindex="-1"></a>})</span>
<span id="cb568-6"><a href="cha-progr-with-models.html#cb568-6" tabindex="-1"></a>n <span class="ot">&lt;-</span> <span class="dv">5</span></span>
<span id="cb568-7"><a href="cha-progr-with-models.html#cb568-7" tabindex="-1"></a>m <span class="ot">&lt;-</span> <span class="fu">nimbleModel</span>(code, <span class="at">constants =</span> <span class="fu">list</span>(<span class="at">n =</span> n, <span class="at">zeroes =</span> <span class="fu">rep</span>(<span class="dv">0</span>, n),</span>
<span id="cb568-8"><a href="cha-progr-with-models.html#cb568-8" tabindex="-1"></a>                                        <span class="at">C =</span> <span class="fu">diag</span>(n)))</span>
<span id="cb568-9"><a href="cha-progr-with-models.html#cb568-9" tabindex="-1"></a>cm <span class="ot">&lt;-</span> <span class="fu">compileNimble</span>(m)</span>
<span id="cb568-10"><a href="cha-progr-with-models.html#cb568-10" tabindex="-1"></a></span>
<span id="cb568-11"><a href="cha-progr-with-models.html#cb568-11" tabindex="-1"></a>nfGen <span class="ot">&lt;-</span> <span class="fu">nimbleFunction</span>(</span>
<span id="cb568-12"><a href="cha-progr-with-models.html#cb568-12" tabindex="-1"></a>    <span class="at">setup =</span> <span class="cf">function</span>(model) {</span>
<span id="cb568-13"><a href="cha-progr-with-models.html#cb568-13" tabindex="-1"></a>    <span class="co"># node1 and node2 would typically be setup arguments, so they could</span></span>
<span id="cb568-14"><a href="cha-progr-with-models.html#cb568-14" tabindex="-1"></a>    <span class="co"># have different values for different models.  We are assigning values</span></span>
<span id="cb568-15"><a href="cha-progr-with-models.html#cb568-15" tabindex="-1"></a>    <span class="co"># here so the example is clearer.</span></span>
<span id="cb568-16"><a href="cha-progr-with-models.html#cb568-16" tabindex="-1"></a>        node1 <span class="ot">&lt;-</span> <span class="st">&#39;sigma&#39;</span>            <span class="co"># a scalar node </span></span>
<span id="cb568-17"><a href="cha-progr-with-models.html#cb568-17" tabindex="-1"></a>        node2 <span class="ot">&lt;-</span> <span class="st">&#39;y[1:5]&#39;</span>           <span class="co"># a vector node</span></span>
<span id="cb568-18"><a href="cha-progr-with-models.html#cb568-18" tabindex="-1"></a>        notReallyANode <span class="ot">&lt;-</span> <span class="st">&#39;y[2:4]&#39;</span>  <span class="co"># y[2:4] allowed even though not a node!</span></span>
<span id="cb568-19"><a href="cha-progr-with-models.html#cb568-19" tabindex="-1"></a>    },</span>
<span id="cb568-20"><a href="cha-progr-with-models.html#cb568-20" tabindex="-1"></a>    <span class="at">run =</span> <span class="cf">function</span>(<span class="at">vals =</span> <span class="fu">double</span>(<span class="dv">1</span>)) {</span>
<span id="cb568-21"><a href="cha-progr-with-models.html#cb568-21" tabindex="-1"></a>        tmp0 <span class="ot">&lt;-</span> model[[node1]]           <span class="co"># 1. tmp0 will be a scalar</span></span>
<span id="cb568-22"><a href="cha-progr-with-models.html#cb568-22" tabindex="-1"></a>        tmp1 <span class="ot">&lt;-</span> model[[node2]]           <span class="co"># 2. tmp1 will be a vector</span></span>
<span id="cb568-23"><a href="cha-progr-with-models.html#cb568-23" tabindex="-1"></a>        tmp2 <span class="ot">&lt;-</span> model[[node2]][<span class="dv">1</span>]        <span class="co"># 3. tmp2 will be a scalar</span></span>
<span id="cb568-24"><a href="cha-progr-with-models.html#cb568-24" tabindex="-1"></a>        tmp3 <span class="ot">&lt;-</span> model[[notReallyANode]]  <span class="co"># 4. tmp3 will be a vector</span></span>
<span id="cb568-25"><a href="cha-progr-with-models.html#cb568-25" tabindex="-1"></a>        tmp4 <span class="ot">&lt;-</span> model<span class="sc">$</span>y[<span class="dv">3</span><span class="sc">:</span><span class="dv">4</span>]      <span class="co"># 5. hard-coded access to a model variable</span></span>
<span id="cb568-26"><a href="cha-progr-with-models.html#cb568-26" tabindex="-1"></a>        <span class="co"># 6. node1 is scalar so can be assigned a scalar:</span></span>
<span id="cb568-27"><a href="cha-progr-with-models.html#cb568-27" tabindex="-1"></a>        model[[node1]] <span class="ot">&lt;&lt;-</span> <span class="fu">runif</span>(<span class="dv">1</span>) </span>
<span id="cb568-28"><a href="cha-progr-with-models.html#cb568-28" tabindex="-1"></a>        model[[node2]][<span class="dv">1</span>] <span class="ot">&lt;&lt;-</span> <span class="fu">runif</span>(<span class="dv">1</span>) </span>
<span id="cb568-29"><a href="cha-progr-with-models.html#cb568-29" tabindex="-1"></a>                 <span class="co"># 7. an element of node2 can be assigned a scalar </span></span>
<span id="cb568-30"><a href="cha-progr-with-models.html#cb568-30" tabindex="-1"></a>        model[[node2]] <span class="ot">&lt;&lt;-</span> <span class="fu">runif</span>(<span class="fu">length</span>(model[[node2]]))</span>
<span id="cb568-31"><a href="cha-progr-with-models.html#cb568-31" tabindex="-1"></a>                 <span class="co"># 8. a vector can be assigned to the vector node2</span></span>
<span id="cb568-32"><a href="cha-progr-with-models.html#cb568-32" tabindex="-1"></a>        model[[node2]][<span class="dv">1</span><span class="sc">:</span><span class="dv">3</span>] <span class="ot">&lt;&lt;-</span> vals[<span class="dv">1</span><span class="sc">:</span><span class="dv">3</span>] </span>
<span id="cb568-33"><a href="cha-progr-with-models.html#cb568-33" tabindex="-1"></a>             <span class="co"># elements of node2 can be indexed as needed</span></span>
<span id="cb568-34"><a href="cha-progr-with-models.html#cb568-34" tabindex="-1"></a>        <span class="fu">returnType</span>(<span class="fu">double</span>(<span class="dv">1</span>))</span>
<span id="cb568-35"><a href="cha-progr-with-models.html#cb568-35" tabindex="-1"></a>        out <span class="ot">&lt;-</span> model[[node2]] <span class="co"># we can return a vector   </span></span>
<span id="cb568-36"><a href="cha-progr-with-models.html#cb568-36" tabindex="-1"></a>        <span class="fu">return</span>(out)</span>
<span id="cb568-37"><a href="cha-progr-with-models.html#cb568-37" tabindex="-1"></a>    }</span>
<span id="cb568-38"><a href="cha-progr-with-models.html#cb568-38" tabindex="-1"></a>    )</span>
<span id="cb568-39"><a href="cha-progr-with-models.html#cb568-39" tabindex="-1"></a>    </span>
<span id="cb568-40"><a href="cha-progr-with-models.html#cb568-40" tabindex="-1"></a>Rnf <span class="ot">&lt;-</span> <span class="fu">nfGen</span>(m)</span>
<span id="cb568-41"><a href="cha-progr-with-models.html#cb568-41" tabindex="-1"></a>Cnf <span class="ot">&lt;-</span> <span class="fu">compileNimble</span>(Rnf)</span>
<span id="cb568-42"><a href="cha-progr-with-models.html#cb568-42" tabindex="-1"></a>Cnf<span class="sc">$</span><span class="fu">run</span>(<span class="fu">rnorm</span>(<span class="dv">10</span>))</span></code></pre></div>
<pre><code>## [1] -1.4887622 -1.1607519  1.4577382  0.4855725  0.6932443</code></pre>
<p>Use of <code>[[ ]]</code> allows one to programmatically access a node based on a character variable containing the node name; this character variable would generally be set in setup code. In contrast, use of <code>$</code> hard codes the variable name and would not generally be suitable for nimbleFunctions intended for use with arbitrary models.</p>
</div>
<div id="sec:getting-setting-more" class="section level4 hasAnchor" number="16.4.2.2">
<h4><span class="header-section-number">16.4.2.2</span> Getting and setting more than one model node or variable at a time using <code>values</code><a href="cha-progr-with-models.html#sec:getting-setting-more" class="anchor-section" aria-label="Anchor link to header"></a></h4>
<p>Sometimes it is useful to set a collection of nodes or variables at
one time. For example, one might want a nimbleFunction that will
serve as the objective function for an optimizer. The input to the
nimbleFunction would be a vector, which should be used to fill a
collection of nodes in the model before calculating their log
probabilities. This can be done using <code>values</code>:</p>
<div class="sourceCode" id="cb570"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb570-1"><a href="cha-progr-with-models.html#cb570-1" tabindex="-1"></a><span class="co"># get values from a set of model nodes into a vector</span></span>
<span id="cb570-2"><a href="cha-progr-with-models.html#cb570-2" tabindex="-1"></a>P <span class="ot">&lt;-</span> <span class="fu">values</span>(model, nodes)</span>
<span id="cb570-3"><a href="cha-progr-with-models.html#cb570-3" tabindex="-1"></a><span class="co"># or put values from a vector into a set of model nodes</span></span>
<span id="cb570-4"><a href="cha-progr-with-models.html#cb570-4" tabindex="-1"></a><span class="fu">values</span>(model, nodes) <span class="ot">&lt;-</span> P</span></code></pre></div>
<p>where the first line would assign the collection of values from <code>nodes</code>
into <code>P</code>, and the second would do the inverse. In both cases, values
from nodes with two or more dimensions are flattened into a vector in
column-wise order.</p>
<p><code>values(model, nodes)</code> may be used as a
vector in other expressions, e.g.,</p>
<div class="sourceCode" id="cb571"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb571-1"><a href="cha-progr-with-models.html#cb571-1" tabindex="-1"></a>Y <span class="ot">&lt;-</span> A <span class="sc">%*%</span> <span class="fu">values</span>(model, nodes) <span class="sc">+</span> b</span></code></pre></div>
<p>One can also index elements of nodes in the argument to values, in the same manner as discussed for <code>calculate</code> and related functions in Section <a href="cha-progr-with-models.html#sec:driv-models:-calc">16.4.1</a>.</p>
<p>Note again the potential for confusion between scalars and vectors of
length 1. <code>values</code> returns a vector and expects a vector when used
on the left-hand side of an assignment. If only a single value is
being assigned, it must be a vector of length 1, not a scalar. This
can be achieved by wrapping a scalar in <code>c()</code> when necessary.
For example:</p>
<div class="sourceCode" id="cb572"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb572-1"><a href="cha-progr-with-models.html#cb572-1" tabindex="-1"></a><span class="co"># c(rnorm(1)) creates vector of length one:</span></span>
<span id="cb572-2"><a href="cha-progr-with-models.html#cb572-2" tabindex="-1"></a><span class="fu">values</span>(model, nodes[<span class="dv">1</span>]) <span class="ot">&lt;-</span> <span class="fu">c</span>(<span class="fu">rnorm</span>(<span class="dv">1</span>)) </span>
<span id="cb572-3"><a href="cha-progr-with-models.html#cb572-3" tabindex="-1"></a><span class="co"># won&#39;t compile because rnorm(1) is a scalar</span></span>
<span id="cb572-4"><a href="cha-progr-with-models.html#cb572-4" tabindex="-1"></a><span class="co"># values(model, nodes[1]) &lt;- rnorm(1)  </span></span>
<span id="cb572-5"><a href="cha-progr-with-models.html#cb572-5" tabindex="-1"></a></span>
<span id="cb572-6"><a href="cha-progr-with-models.html#cb572-6" tabindex="-1"></a>out <span class="ot">&lt;-</span> <span class="fu">values</span>(model, nodes[<span class="dv">1</span>]) <span class="co"># out is a vector</span></span>
<span id="cb572-7"><a href="cha-progr-with-models.html#cb572-7" tabindex="-1"></a>out2 <span class="ot">&lt;-</span> <span class="fu">values</span>(model, nodes[<span class="dv">1</span>])[<span class="dv">1</span>] <span class="co"># out2 is a scalar</span></span></code></pre></div>
</div>
</div>
<div id="getting-parameter-values-and-node-bounds" class="section level3 hasAnchor" number="16.4.3">
<h3><span class="header-section-number">16.4.3</span> Getting parameter values and node bounds<a href="cha-progr-with-models.html#getting-parameter-values-and-node-bounds" class="anchor-section" aria-label="Anchor link to header"></a></h3>
<p>Sections <a href="cha-using-models.html#sec:getParam">14.2.3</a>-<a href="cha-using-models.html#sec:getBound">14.2.4</a> describe how to get the parameter values for a node and the range (bounds) of possible values for the node using <code>getParam</code> and <code>getBound</code>. Both of these can be used in run code.</p>
</div>
<div id="sec:access-model-modelv" class="section level3 hasAnchor" number="16.4.4">
<h3><span class="header-section-number">16.4.4</span> Using modelValues objects<a href="cha-progr-with-models.html#sec:access-model-modelv" class="anchor-section" aria-label="Anchor link to header"></a></h3>
<p>The <code>modelValues</code> structure was introduced in Section
<a href="cha-data-structures.html#sec:modelValues-struct">15.1</a>. Inside nimbleFunctions, modelValues are
designed to easily save values from a model object during the running
of a nimbleFunction. A modelValues object used in run code
must always exist in the setup code, either by passing it in as a
setup argument or creating it in the setup code.</p>
<p>To illustrate this, we will create a nimbleFunction for computing
importance weights for importance sampling. This function will use two
modelValues objects. <code>propModelValues</code> will contain a set of
values simulated from the importance sampling distribution and a field <code>propLL</code>
for their log
probabilities (densities). <code>savedWeights</code> will contain the
difference in log probability (density) between the model and the
<code>propLL</code> value provided for each set of values.</p>
<div class="sourceCode" id="cb573"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb573-1"><a href="cha-progr-with-models.html#cb573-1" tabindex="-1"></a><span class="co"># Accepting modelValues as a setup argument</span></span>
<span id="cb573-2"><a href="cha-progr-with-models.html#cb573-2" tabindex="-1"></a>swConf <span class="ot">&lt;-</span> <span class="fu">modelValuesConf</span>(<span class="at">vars =</span> <span class="st">&quot;w&quot;</span>,</span>
<span id="cb573-3"><a href="cha-progr-with-models.html#cb573-3" tabindex="-1"></a>                                    <span class="at">types =</span> <span class="st">&quot;double&quot;</span>,</span>
<span id="cb573-4"><a href="cha-progr-with-models.html#cb573-4" tabindex="-1"></a>                                    <span class="at">sizes =</span> <span class="dv">1</span>)</span>
<span id="cb573-5"><a href="cha-progr-with-models.html#cb573-5" tabindex="-1"></a>setup <span class="ot">=</span> <span class="cf">function</span>(propModelValues, model, savedWeightsConf){</span>
<span id="cb573-6"><a href="cha-progr-with-models.html#cb573-6" tabindex="-1"></a>    <span class="co"># Building a modelValues in the setup function </span></span>
<span id="cb573-7"><a href="cha-progr-with-models.html#cb573-7" tabindex="-1"></a>    savedWeights <span class="ot">&lt;-</span> <span class="fu">modelValues</span>(<span class="at">conf =</span> savedWeightsConf)</span>
<span id="cb573-8"><a href="cha-progr-with-models.html#cb573-8" tabindex="-1"></a>    <span class="co"># List of nodes to be used in run function</span></span>
<span id="cb573-9"><a href="cha-progr-with-models.html#cb573-9" tabindex="-1"></a>    modelNodes <span class="ot">&lt;-</span> model<span class="sc">$</span><span class="fu">getNodeNames</span>(<span class="at">stochOnly =</span> <span class="cn">TRUE</span>,</span>
<span id="cb573-10"><a href="cha-progr-with-models.html#cb573-10" tabindex="-1"></a>                                     <span class="at">includeData =</span> <span class="cn">FALSE</span>)</span>
<span id="cb573-11"><a href="cha-progr-with-models.html#cb573-11" tabindex="-1"></a>}</span></code></pre></div>
<p>The simplest way to pass values back and forth between models and
modelValues inside of a nimbleFunction is with <code>copy</code>, which
has the synonym <code>nimCopy</code>. See <code>help(nimCopy)</code> for argument details.
<!--- % This takes arguments
  %%
 %   1.[`from`, `to`] which can either be models or modelValues
  %%   1.[`row`, `rowTo`] which refers to the rows of a
 %   modelValues object, if either `from` or `to` is a
  %%   modelValues. If `rowTo` is omitted, it is assumed to be equal to `row` if necessary.
 %     1.[`nodes`, `nodesTo`] which is a vector of the names of the nodes
  %%     to be copied.  The node names will be expanded when variable names are provided.  If
 %     `nodesTo` is omitted it will be set equal to `nodes`.
  %%
   TODO: CHECK THESE USAGES --></p>
<p>Alternatively, the values may be accessed via indexing of individual
rows, using the notation <code>mv[var, i]</code>, where <code>mv</code> is a
modelValues object, <code>var</code> is a variable name (not a node name),
and <code>i</code> is a row number. Likewise, the <code>getsize</code> and
<code>resize</code> functions can be used as discussed in Section <a href="cha-data-structures.html#sec:modelValues-struct">15.1</a>. However the function
<code>as.matrix</code> does not work in run code.</p>
<p>Here is a run function to use these modelValues:</p>
<div class="sourceCode" id="cb574"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb574-1"><a href="cha-progr-with-models.html#cb574-1" tabindex="-1"></a>run <span class="ot">=</span> <span class="cf">function</span>(){</span>
<span id="cb574-2"><a href="cha-progr-with-models.html#cb574-2" tabindex="-1"></a>    <span class="co"># gets the number of rows of propSamples</span></span>
<span id="cb574-3"><a href="cha-progr-with-models.html#cb574-3" tabindex="-1"></a>    m <span class="ot">&lt;-</span> <span class="fu">getsize</span>(propModelValues)</span>
<span id="cb574-4"><a href="cha-progr-with-models.html#cb574-4" tabindex="-1"></a></span>
<span id="cb574-5"><a href="cha-progr-with-models.html#cb574-5" tabindex="-1"></a>    <span class="co"># resized savedWeights to have the proper rows</span></span>
<span id="cb574-6"><a href="cha-progr-with-models.html#cb574-6" tabindex="-1"></a>    <span class="fu">resize</span>(savedWeights, m)</span>
<span id="cb574-7"><a href="cha-progr-with-models.html#cb574-7" tabindex="-1"></a>    <span class="cf">for</span>(i <span class="cf">in</span> <span class="dv">1</span><span class="sc">:</span>m){</span>
<span id="cb574-8"><a href="cha-progr-with-models.html#cb574-8" tabindex="-1"></a>        <span class="co"># Copying from propSamples to model. </span></span>
<span id="cb574-9"><a href="cha-progr-with-models.html#cb574-9" tabindex="-1"></a>        <span class="co"># Node names of propSamples and model must match!</span></span>
<span id="cb574-10"><a href="cha-progr-with-models.html#cb574-10" tabindex="-1"></a>        <span class="fu">nimCopy</span>(<span class="at">from =</span> propModelValues, <span class="at">to =</span> model, <span class="at">row =</span> i,</span>
<span id="cb574-11"><a href="cha-progr-with-models.html#cb574-11" tabindex="-1"></a>                <span class="at">nodes =</span> modelNodes, <span class="at">logProb =</span> <span class="cn">FALSE</span>)</span>
<span id="cb574-12"><a href="cha-progr-with-models.html#cb574-12" tabindex="-1"></a>        <span class="co"># calculates the log likelihood of the model</span></span>
<span id="cb574-13"><a href="cha-progr-with-models.html#cb574-13" tabindex="-1"></a>        targLL <span class="ot">&lt;-</span> model<span class="sc">$</span><span class="fu">calculate</span>()</span>
<span id="cb574-14"><a href="cha-progr-with-models.html#cb574-14" tabindex="-1"></a>        <span class="co"># retrieves the saved log likelihood from the proposed model</span></span>
<span id="cb574-15"><a href="cha-progr-with-models.html#cb574-15" tabindex="-1"></a>        propLL <span class="ot">&lt;-</span> propModelValues[<span class="st">&quot;propLL&quot;</span>,i][<span class="dv">1</span>]</span>
<span id="cb574-16"><a href="cha-progr-with-models.html#cb574-16" tabindex="-1"></a>        <span class="co"># saves the importance weight for the i-th sample </span></span>
<span id="cb574-17"><a href="cha-progr-with-models.html#cb574-17" tabindex="-1"></a>        savedWeights[<span class="st">&quot;w&quot;</span>, i][<span class="dv">1</span>] <span class="ot">&lt;&lt;-</span> <span class="fu">exp</span>(targLL <span class="sc">-</span> propLL)</span>
<span id="cb574-18"><a href="cha-progr-with-models.html#cb574-18" tabindex="-1"></a>    }</span>
<span id="cb574-19"><a href="cha-progr-with-models.html#cb574-19" tabindex="-1"></a>    <span class="co"># does not return anything</span></span>
<span id="cb574-20"><a href="cha-progr-with-models.html#cb574-20" tabindex="-1"></a>}</span></code></pre></div>
<p>Once the nimbleFunction is built, the modelValues object can be accessed
using <code>$</code>, which is shown in more detail below. In
fact, since modelValues, like most NIMBLE objects, are reference class
objects, one can get a reference to them before the function is
executed and then use that reference afterwards.</p>
<div class="sourceCode" id="cb575"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb575-1"><a href="cha-progr-with-models.html#cb575-1" tabindex="-1"></a><span class="co"># simple model and modelValues for example use with code above</span></span>
<span id="cb575-2"><a href="cha-progr-with-models.html#cb575-2" tabindex="-1"></a>targetModelCode <span class="ot">&lt;-</span> <span class="fu">nimbleCode</span>({</span>
<span id="cb575-3"><a href="cha-progr-with-models.html#cb575-3" tabindex="-1"></a>    x <span class="sc">~</span> <span class="fu">dnorm</span>(<span class="dv">0</span>,<span class="dv">1</span>)</span>
<span id="cb575-4"><a href="cha-progr-with-models.html#cb575-4" tabindex="-1"></a>    <span class="cf">for</span>(i <span class="cf">in</span> <span class="dv">1</span><span class="sc">:</span><span class="dv">4</span>)</span>
<span id="cb575-5"><a href="cha-progr-with-models.html#cb575-5" tabindex="-1"></a>        y[i] <span class="sc">~</span> <span class="fu">dnorm</span>(<span class="dv">0</span>,<span class="dv">1</span>)</span>
<span id="cb575-6"><a href="cha-progr-with-models.html#cb575-6" tabindex="-1"></a>})</span>
<span id="cb575-7"><a href="cha-progr-with-models.html#cb575-7" tabindex="-1"></a></span>
<span id="cb575-8"><a href="cha-progr-with-models.html#cb575-8" tabindex="-1"></a><span class="co"># code for proposal model</span></span>
<span id="cb575-9"><a href="cha-progr-with-models.html#cb575-9" tabindex="-1"></a>propModelCode <span class="ot">&lt;-</span> <span class="fu">nimbleCode</span>({</span>
<span id="cb575-10"><a href="cha-progr-with-models.html#cb575-10" tabindex="-1"></a>    x <span class="sc">~</span> <span class="fu">dnorm</span>(<span class="dv">0</span>,<span class="dv">2</span>)</span>
<span id="cb575-11"><a href="cha-progr-with-models.html#cb575-11" tabindex="-1"></a>    <span class="cf">for</span>(i <span class="cf">in</span> <span class="dv">1</span><span class="sc">:</span><span class="dv">4</span>)</span>
<span id="cb575-12"><a href="cha-progr-with-models.html#cb575-12" tabindex="-1"></a>        y[i] <span class="sc">~</span> <span class="fu">dnorm</span>(<span class="dv">0</span>,<span class="dv">2</span>)</span>
<span id="cb575-13"><a href="cha-progr-with-models.html#cb575-13" tabindex="-1"></a>})</span>
<span id="cb575-14"><a href="cha-progr-with-models.html#cb575-14" tabindex="-1"></a></span>
<span id="cb575-15"><a href="cha-progr-with-models.html#cb575-15" tabindex="-1"></a><span class="co"># creating the models</span></span>
<span id="cb575-16"><a href="cha-progr-with-models.html#cb575-16" tabindex="-1"></a>targetModel <span class="ot">=</span> <span class="fu">nimbleModel</span>(targetModelCode, <span class="at">check =</span> <span class="cn">FALSE</span>)</span>
<span id="cb575-17"><a href="cha-progr-with-models.html#cb575-17" tabindex="-1"></a>propModel <span class="ot">=</span> <span class="fu">nimbleModel</span>(propModelCode, <span class="at">check =</span> <span class="cn">FALSE</span>)</span>
<span id="cb575-18"><a href="cha-progr-with-models.html#cb575-18" tabindex="-1"></a>cTargetModel <span class="ot">=</span> <span class="fu">compileNimble</span>(targetModel)</span>
<span id="cb575-19"><a href="cha-progr-with-models.html#cb575-19" tabindex="-1"></a>cPropModel <span class="ot">=</span> <span class="fu">compileNimble</span>(propModel)</span>
<span id="cb575-20"><a href="cha-progr-with-models.html#cb575-20" tabindex="-1"></a></span>
<span id="cb575-21"><a href="cha-progr-with-models.html#cb575-21" tabindex="-1"></a></span>
<span id="cb575-22"><a href="cha-progr-with-models.html#cb575-22" tabindex="-1"></a>sampleMVConf <span class="ot">=</span> <span class="fu">modelValuesConf</span>(<span class="at">vars =</span> <span class="fu">c</span>(<span class="st">&quot;x&quot;</span>, <span class="st">&quot;y&quot;</span>, <span class="st">&quot;propLL&quot;</span>), </span>
<span id="cb575-23"><a href="cha-progr-with-models.html#cb575-23" tabindex="-1"></a>    <span class="at">types =</span> <span class="fu">c</span>(<span class="st">&quot;double&quot;</span>, <span class="st">&quot;double&quot;</span>, <span class="st">&quot;double&quot;</span>), </span>
<span id="cb575-24"><a href="cha-progr-with-models.html#cb575-24" tabindex="-1"></a>    <span class="at">sizes =</span> <span class="fu">list</span>(<span class="at">x =</span> <span class="dv">1</span>, <span class="at">y =</span> <span class="dv">4</span>, <span class="at">propLL =</span> <span class="dv">1</span>) )</span>
<span id="cb575-25"><a href="cha-progr-with-models.html#cb575-25" tabindex="-1"></a></span>
<span id="cb575-26"><a href="cha-progr-with-models.html#cb575-26" tabindex="-1"></a>sampleMV <span class="ot">&lt;-</span> <span class="fu">modelValues</span>(sampleMVConf)</span>
<span id="cb575-27"><a href="cha-progr-with-models.html#cb575-27" tabindex="-1"></a></span>
<span id="cb575-28"><a href="cha-progr-with-models.html#cb575-28" tabindex="-1"></a><span class="co">#   nimbleFunction for generating proposal sample</span></span>
<span id="cb575-29"><a href="cha-progr-with-models.html#cb575-29" tabindex="-1"></a>PropSamp_Gen <span class="ot">&lt;-</span> <span class="fu">nimbleFunction</span>(</span>
<span id="cb575-30"><a href="cha-progr-with-models.html#cb575-30" tabindex="-1"></a>    <span class="at">setup =</span> <span class="cf">function</span>(mv, propModel){</span>
<span id="cb575-31"><a href="cha-progr-with-models.html#cb575-31" tabindex="-1"></a>        nodeNames <span class="ot">&lt;-</span> propModel<span class="sc">$</span><span class="fu">getNodeNames</span>()</span>
<span id="cb575-32"><a href="cha-progr-with-models.html#cb575-32" tabindex="-1"></a>    },</span>
<span id="cb575-33"><a href="cha-progr-with-models.html#cb575-33" tabindex="-1"></a>    <span class="at">run =</span> <span class="cf">function</span>(<span class="at">m =</span> <span class="fu">integer</span>() ){</span>
<span id="cb575-34"><a href="cha-progr-with-models.html#cb575-34" tabindex="-1"></a>        <span class="fu">resize</span>(mv, m)</span>
<span id="cb575-35"><a href="cha-progr-with-models.html#cb575-35" tabindex="-1"></a>        <span class="cf">for</span>(i <span class="cf">in</span> <span class="dv">1</span><span class="sc">:</span>m){</span>
<span id="cb575-36"><a href="cha-progr-with-models.html#cb575-36" tabindex="-1"></a>            propModel<span class="sc">$</span><span class="fu">simulate</span>()</span>
<span id="cb575-37"><a href="cha-progr-with-models.html#cb575-37" tabindex="-1"></a>            <span class="fu">nimCopy</span>(<span class="at">from =</span> propModel, <span class="at">to =</span> mv, <span class="at">nodes =</span> nodeNames, <span class="at">row =</span> i)</span>
<span id="cb575-38"><a href="cha-progr-with-models.html#cb575-38" tabindex="-1"></a>            mv[<span class="st">&quot;propLL&quot;</span>, i][<span class="dv">1</span>] <span class="ot">&lt;&lt;-</span> propModel<span class="sc">$</span><span class="fu">calculate</span>()</span>
<span id="cb575-39"><a href="cha-progr-with-models.html#cb575-39" tabindex="-1"></a>        }</span>
<span id="cb575-40"><a href="cha-progr-with-models.html#cb575-40" tabindex="-1"></a>    }</span>
<span id="cb575-41"><a href="cha-progr-with-models.html#cb575-41" tabindex="-1"></a>    )</span>
<span id="cb575-42"><a href="cha-progr-with-models.html#cb575-42" tabindex="-1"></a></span>
<span id="cb575-43"><a href="cha-progr-with-models.html#cb575-43" tabindex="-1"></a><span class="co"># nimbleFunction for calculating importance weights</span></span>
<span id="cb575-44"><a href="cha-progr-with-models.html#cb575-44" tabindex="-1"></a><span class="co"># uses setup and run functions as defined in previous code chunk</span></span>
<span id="cb575-45"><a href="cha-progr-with-models.html#cb575-45" tabindex="-1"></a>impWeights_Gen <span class="ot">&lt;-</span> <span class="fu">nimbleFunction</span>(<span class="at">setup =</span> setup,</span>
<span id="cb575-46"><a href="cha-progr-with-models.html#cb575-46" tabindex="-1"></a>                                 <span class="at">run =</span> run)</span>
<span id="cb575-47"><a href="cha-progr-with-models.html#cb575-47" tabindex="-1"></a>      </span>
<span id="cb575-48"><a href="cha-progr-with-models.html#cb575-48" tabindex="-1"></a></span>
<span id="cb575-49"><a href="cha-progr-with-models.html#cb575-49" tabindex="-1"></a><span class="co"># making instances of nimbleFunctions</span></span>
<span id="cb575-50"><a href="cha-progr-with-models.html#cb575-50" tabindex="-1"></a><span class="co"># note that both functions share the same modelValues object</span></span>
<span id="cb575-51"><a href="cha-progr-with-models.html#cb575-51" tabindex="-1"></a>RPropSamp <span class="ot">&lt;-</span> <span class="fu">PropSamp_Gen</span>(sampleMV, propModel)</span>
<span id="cb575-52"><a href="cha-progr-with-models.html#cb575-52" tabindex="-1"></a>RImpWeights <span class="ot">&lt;-</span> <span class="fu">impWeights_Gen</span>(sampleMV, targetModel, swConf)</span>
<span id="cb575-53"><a href="cha-progr-with-models.html#cb575-53" tabindex="-1"></a></span>
<span id="cb575-54"><a href="cha-progr-with-models.html#cb575-54" tabindex="-1"></a><span class="co"># compiling </span></span>
<span id="cb575-55"><a href="cha-progr-with-models.html#cb575-55" tabindex="-1"></a>CPropSamp <span class="ot">&lt;-</span> <span class="fu">compileNimble</span>(RPropSamp, <span class="at">project =</span> propModel)</span>
<span id="cb575-56"><a href="cha-progr-with-models.html#cb575-56" tabindex="-1"></a>CImpWeights <span class="ot">&lt;-</span> <span class="fu">compileNimble</span>(RImpWeights, <span class="at">project =</span> targetModel)</span>
<span id="cb575-57"><a href="cha-progr-with-models.html#cb575-57" tabindex="-1"></a></span>
<span id="cb575-58"><a href="cha-progr-with-models.html#cb575-58" tabindex="-1"></a><span class="co"># generating and saving proposal sample of size 10</span></span>
<span id="cb575-59"><a href="cha-progr-with-models.html#cb575-59" tabindex="-1"></a>CPropSamp<span class="sc">$</span><span class="fu">run</span>(<span class="dv">10</span>)</span>
<span id="cb575-60"><a href="cha-progr-with-models.html#cb575-60" tabindex="-1"></a></span>
<span id="cb575-61"><a href="cha-progr-with-models.html#cb575-61" tabindex="-1"></a><span class="co"># calculating the importance weights and saving to mv</span></span>
<span id="cb575-62"><a href="cha-progr-with-models.html#cb575-62" tabindex="-1"></a>CImpWeights<span class="sc">$</span><span class="fu">run</span>()</span>
<span id="cb575-63"><a href="cha-progr-with-models.html#cb575-63" tabindex="-1"></a></span>
<span id="cb575-64"><a href="cha-progr-with-models.html#cb575-64" tabindex="-1"></a><span class="co"># retrieving the modelValues objects</span></span>
<span id="cb575-65"><a href="cha-progr-with-models.html#cb575-65" tabindex="-1"></a><span class="co"># extracted objects are C-based modelValues objects</span></span>
<span id="cb575-66"><a href="cha-progr-with-models.html#cb575-66" tabindex="-1"></a></span>
<span id="cb575-67"><a href="cha-progr-with-models.html#cb575-67" tabindex="-1"></a>savedPropSamp_1 <span class="ot">=</span> CImpWeights<span class="sc">$</span>propModelValues</span>
<span id="cb575-68"><a href="cha-progr-with-models.html#cb575-68" tabindex="-1"></a>savedPropSamp_2 <span class="ot">=</span> CPropSamp<span class="sc">$</span>mv</span>
<span id="cb575-69"><a href="cha-progr-with-models.html#cb575-69" tabindex="-1"></a></span>
<span id="cb575-70"><a href="cha-progr-with-models.html#cb575-70" tabindex="-1"></a><span class="co"># Subtle note: savedPropSamp_1 and savedPropSamp_2</span></span>
<span id="cb575-71"><a href="cha-progr-with-models.html#cb575-71" tabindex="-1"></a><span class="co"># both provide interface to the same compiled modelValues objects!</span></span>
<span id="cb575-72"><a href="cha-progr-with-models.html#cb575-72" tabindex="-1"></a><span class="co"># This is because they were both built from sampleMV.</span></span>
<span id="cb575-73"><a href="cha-progr-with-models.html#cb575-73" tabindex="-1"></a></span>
<span id="cb575-74"><a href="cha-progr-with-models.html#cb575-74" tabindex="-1"></a>savedPropSamp_1[<span class="st">&quot;x&quot;</span>,<span class="dv">1</span>]</span></code></pre></div>
<pre><code>## [1] -0.3919851</code></pre>
<div class="sourceCode" id="cb577"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb577-1"><a href="cha-progr-with-models.html#cb577-1" tabindex="-1"></a>savedPropSamp_2[<span class="st">&quot;x&quot;</span>,<span class="dv">1</span>]</span></code></pre></div>
<pre><code>## [1] -0.3919851</code></pre>
<div class="sourceCode" id="cb579"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb579-1"><a href="cha-progr-with-models.html#cb579-1" tabindex="-1"></a>savedPropSamp_1[<span class="st">&quot;x&quot;</span>,<span class="dv">1</span>] <span class="ot">&lt;-</span> <span class="dv">0</span> <span class="co"># example of directly setting a value</span></span>
<span id="cb579-2"><a href="cha-progr-with-models.html#cb579-2" tabindex="-1"></a>savedPropSamp_2[<span class="st">&quot;x&quot;</span>,<span class="dv">1</span>]</span></code></pre></div>
<pre><code>## [1] 0</code></pre>
<div class="sourceCode" id="cb581"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb581-1"><a href="cha-progr-with-models.html#cb581-1" tabindex="-1"></a><span class="co"># viewing the saved importance weights</span></span>
<span id="cb581-2"><a href="cha-progr-with-models.html#cb581-2" tabindex="-1"></a>savedWeights <span class="ot">&lt;-</span> CImpWeights<span class="sc">$</span>savedWeights</span>
<span id="cb581-3"><a href="cha-progr-with-models.html#cb581-3" tabindex="-1"></a><span class="fu">unlist</span>(savedWeights[[<span class="st">&quot;w&quot;</span>]])</span></code></pre></div>
<pre><code>##  [1] 0.5024644 1.7765357 3.5155732 0.3893390 0.5530012 0.5085069 2.3890876
##  [8] 0.6548453 0.5940573 0.4943950</code></pre>
<div class="sourceCode" id="cb583"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb583-1"><a href="cha-progr-with-models.html#cb583-1" tabindex="-1"></a><span class="co"># viewing first 3 rows -- note that savedPropSsamp_1[&quot;x&quot;, 1] was altered </span></span>
<span id="cb583-2"><a href="cha-progr-with-models.html#cb583-2" tabindex="-1"></a><span class="fu">as.matrix</span>(savedPropSamp_1)[<span class="dv">1</span><span class="sc">:</span><span class="dv">3</span>, ]</span></code></pre></div>
<pre><code>##      propLL[1]        x[1]       y[1]       y[2]       y[3]       y[4]
## [1,] -4.951100  0.00000000 -0.3951785 -1.0040021 -0.6420246  0.5993671
## [2,] -7.476891  0.35317374 -0.4510107  1.7080201  1.0471602 -0.5225390
## [3,] -8.841966 -0.06913039 -2.2884703  0.6767462  0.3955090 -0.3519276</code></pre>
<p>Importance sampling could also be written using simple vectors for the
weights, but we illustrated putting them in a modelValues object along
with model variables.</p>
</div>
<div id="sec:using-model-variable" class="section level3 hasAnchor" number="16.4.5">
<h3><span class="header-section-number">16.4.5</span> Using model variables and modelValues in expressions<a href="cha-progr-with-models.html#sec:using-model-variable" class="anchor-section" aria-label="Anchor link to header"></a></h3>
<p>Each way of accessing a variable, node, or modelValues can be used amidst mathematical
expressions, including with indexing, or passed to another
nimbleFunction as an argument. For example, the following two
statements would be valid:</p>
<div class="sourceCode" id="cb585"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb585-1"><a href="cha-progr-with-models.html#cb585-1" tabindex="-1"></a>model[[<span class="st">&quot;x[2:8, ]&quot;</span>]][<span class="dv">2</span><span class="sc">:</span><span class="dv">4</span>, <span class="dv">1</span><span class="sc">:</span><span class="dv">3</span>] <span class="sc">%*%</span> Z</span></code></pre></div>
<p>if Z is a vector or matrix, and</p>
<div class="sourceCode" id="cb586"><pre class="sourceCode r"><code class="sourceCode r"><span id="cb586-1"><a href="cha-progr-with-models.html#cb586-1" tabindex="-1"></a>C[<span class="dv">6</span><span class="sc">:</span><span class="dv">10</span>] <span class="ot">&lt;-</span> mv[v, i][<span class="dv">1</span><span class="sc">:</span><span class="dv">5</span>, k] <span class="sc">+</span> B</span></code></pre></div>
<p>if B is a vector or matrix.</p>
<p>The NIMBLE language allows scalars, but models defined from BUGS code