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
|
#
# Copyright 2007-2020 by the individuals mentioned in the source code history
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
cycleDetection <- function(flatModel) {
dependencies <- new("MxDirectedGraph")
if (length(flatModel@fitfunctions) > 0) {
for(i in 1:length(flatModel@fitfunctions)) {
dependencies <- addFitFunctionDetection(flatModel@fitfunctions[[i]], flatModel, dependencies)
# if (!is(dependencies, "MxDirectedGraph")) stop("Lost dependencies")
}
}
if (length(flatModel@expectations) > 0) {
for(i in 1:length(flatModel@expectations)) {
dependencies <- addExpectationDetection(flatModel@expectations[[i]], dependencies)
# if (!is(dependencies, "MxDirectedGraph")) stop("Lost dependencies")
}
}
if (length(flatModel@algebras) > 0) {
for(i in 1:length(flatModel@algebras)) {
dependencies <- addAlgebraDetection(flatModel@algebras[[i]], dependencies)
}
}
if (length(flatModel@matrices) > 0) {
for(i in 1:length(flatModel@matrices)) {
dependencies <- addMatrixDetection(flatModel@matrices[[i]], dependencies)
}
}
containsCycle(dependencies, flatModel)
return(dependencies)
}
containsCycle <- function(graph, flatModel) {
nodes <- graph@nodes
if (length(nodes) > 0) {
colors <- character()
backedges <- character()
colors[nodes] <- 'white'
info <- list(colors, backedges)
for(i in 1:length(nodes)) {
node <- nodes[[i]]
if (colors[[node]] == 'white') {
info <- cycleVisitor(graph, node, info, flatModel@name)
colors <- info[[1]]
}
}
}
}
cycleVisitor <- function(graph, vertex, info, modelname) {
colors <- info[[1]]
backedges <- info[[2]]
colors[[vertex]] <- 'grey'
edges <- graph@edges[[vertex]]
if (!is.null(edges) && length(edges) > 0) {
for(i in 1:length(edges)) {
destination <- edges[[i]]
backedges[[destination]] <- vertex
if (colors[[destination]] == 'grey') {
reportCycle(backedges, destination, modelname)
} else if (colors[[destination]] == 'white') {
info <- list(colors, backedges)
info <- cycleVisitor(graph, destination, info, modelname)
colors <- info[[1]]
backedges <- info[[2]]
}
}
}
colors[[vertex]] <- 'black'
info <- list(colors, backedges)
return(info)
}
reportCycle <- function(backedges, destination, modelname) {
target <- backedges[[destination]]
cycle <- union(destination, target)
while(target != destination) {
target <- backedges[[target]]
cycle <- union(cycle, target)
}
cycle <- sapply(cycle, simplifyName, modelname)
report <- cycle[!sapply(cycle, hasSquareBrackets)]
stop(paste("A cycle has been detected",
"in model", omxQuotes(modelname),
". It involved the following elements:",
omxQuotes(report)),
"\nA common trigger for this error is not providing a ",
"name string as the first parameter to mxModel."
, call. = FALSE)
}
addFitFunctionDetection <- function(fitfunction, flatModel, dependencies) {
dependencies <- genericFitDependencies(fitfunction, flatModel, dependencies)
return(dependencies)
}
addExpectationDetection <- function(expectation, dependencies) {
dependencies <- genericExpDependencies(expectation, dependencies)
return(dependencies)
}
addMatrixDetection <- function(matrix, dependencies) {
labels <- matrix@labels
select <- matrix@.squareBrackets
subs <- labels[select]
if (length(subs) > 0) {
for(i in 1:length(subs)) {
components <- splitSubstitution(subs[[i]])
name <- components[[1]]
dependencies <- imxAddDependency(name, matrix@name, dependencies)
}
}
return(dependencies)
}
addAlgebraDetection <- function(algebra, dependencies) {
if (algebra@fixed) return(dependencies)
sink <- algebra@name
formula <- algebra@formula
dependencies <- addFormulaDetection(formula, sink, dependencies)
return(dependencies)
}
addFormulaDetection <- function(formula, sink, dependencies) {
if (length(formula) == 1) {
dependencies <- imxAddDependency(as.character(formula), sink, dependencies)
} else {
for (i in 2:length(formula)) {
dependencies <- addFormulaDetection(formula[[i]], sink, dependencies)
}
}
return(dependencies)
}
##' Add a dependency
##'
##' The dependency tracking system ensures that algebra and
##' fitfunctions are not recomputed if their inputs have not changed.
##' Dependency information is computed prior to handing the model off
##' to the optimizer to reduce overhead during optimization.
##'
##' Each free parameter keeps track of all the objects that store that
##' free parameter and the transitive closure of all algebras and fit
##' functions that depend on that free parameter. Similarly, each
##' definition variable keeps track of all the objects that store that
##' free parameter and the transitive closure of all the algebras and
##' fit functions that depend on that free parameter. At each
##' iteration of the optimization, when the free parameter values are
##' updated, all of the dependencies of that free parameter are marked
##' as dirty (see \code{omxFitFunction.repopulateFun}). After an
##' algebra or fit function is computed, \code{omxMarkClean()} is
##' called to to indicate that the algebra or fit function is updated.
##' Similarly, when definition variables are populated in FIML, all of
##' the dependencies of the definition variables are marked as dirty.
##' Particularly for FIML, the fact that non-definition-variable
##' dependencies remain clean is a big performance gain.
##'
##' @param source a character vector of the names of the computation sources (inputs)
##' @param sink the name of the computation sink (output)
##' @param dependencies the dependency graph
imxAddDependency <- function(source, sink, dependencies) {
if (length(source) == 0) return(dependencies)
dependencies <- addNode(source, dependencies)
dependencies <- addNode(sink, dependencies)
for (s1 in sink) {
dependencies <- addEdge(source, s1, dependencies)
}
dependencies
}
|
