In theory, the TMB and R output of gadget3 should be
identically-behaving functions, so instead of trying to decipher what
the TMB model is doing wrong, you can use R function instead in a more
familiar environment, for example using standard tools such as
options(error=recover)
.
You can also use edit()
to edit the R function directly
and re-run the model:
# Model setup will look something like this
ling_model <- g3_to_r(...)
ling_model <- edit(ling_model) ; ling_model(ling_param)
…this is useful if you want to add trace print()
statements around a particular part of the model that’s failing, e.g. or
insert a breakpoint by adding a recover()
line.
It’s possible to have a working R model that doesn’t compile using TMB, due to the relative strictness of the Eigen array library in comparison to R arrays, for example. In which case you’ll need to debug the TMB version.
If the model crashes whilst forming the TMB ADFun object, then it
takes your R session with it. To prevent this, wrap
g3_tmb_adfun()
with TMB::gdbsource()
as
follows:
# Model setup will look something like this
tmb_ling <- g3_to_tmb(...)
tmb_param <- attr(tmb_ling, 'parameter_template')
writeLines(TMB::gdbsource(g3_tmb_adfun(
tmb_ling,
tmb_param,
compile_flags = "-g",
output_script = TRUE)))
output_script = TRUE
tells g3_tmb_adfun()
to, after compilation, write a temporary R script that will build the
TMB ADFun object (and presumably crash in the process).
TMB::gdbsource()
in turn runs a provided R script in a
fresh R session wrapped in gdb. By default it will print a stacktrace
and quit, which should show you where the crash occured.
As with the R model you can edit the raw C++ source before building:
tmb_ling <- edit(tmb_ling)
writeLines(TMB::gdbsource(g3_tmb_adfun(
tmb_ling,
tmb_param,
compile_flags = "-g",
output_script = TRUE)))
Through this you can…
std::cout << ling__Linf << std::endl;
().cols()
and ().rows()
to get the
size of an array expressionling_imm__consratio.print()
In theory you can use interactive = TRUE
with
TMB::gdbsource()
, however as this eats error messages it’s
better to do this by hand:
> g3_tmb_adfun(tmb_ling, tmb_param, compile_flags = "-g", output_script = TRUE)
[1] "/tmp/RtmpysTVvW/file3da4a6f13a80c.R"
R -d gdb
(gdb) run --vanilla < /tmp/RtmpysTVvW/file3da4a6f13a80c.R
. . . Compilation, crash at some point . . .
(gdb) up
(gdb) up
(gdb) up
(gdb) call ling_imm__consratio.print()
Array dim: 35 1 8
Array val: -nan -nan -nan -nan
(gdb) call ling_imm__num.print()
Array dim: 35 1 8
Array val: -nan -nan -nan -nan -nan
(gdb) print cur_time
$5 = 0
Note that for the .print()
method to be available for
arrays, it has to be referenced at least once in the model source,
otherwise it won’t be compiled in. Use edit(tmb_ling)
to
add it somewhere first.
Some notes on debugging errors with random effects models.
Control arguments for the inner TMB:::newton()
model can
be provided to g3_tmb_adfun()
, e.g. to add tracing:
m
The error:
Error in if (m < 0) { : missing value where TRUE/FALSE needed
…comes from TMB’s newton optimiser, and essentially says there is
NaN
in the hessian matrix. m
in this case is
equivalent to:
Error in if (norm(par - parold) < step.tol) { :
missing value where TRUE/FALSE needed
The par
list of parameters to optimise became NaN, for
various reasons, but likely solveCholesky()
failed.
Generally, TMB:::newton()
is called from
obj.fn$env$ff
. You can extract the arguments provided by
editing this function:
Around line 16, add code to extract the arguments provided, e.g:
assign("newt.args", c(list(par = eval(random.start),
fn = f0, gr = function(x) f0(x, order = 1), he = H0,
env = env), inner.control), env = globalenv())
Then you can perform a single run to extract arguments:
…or modify the newton function: