[ADMB Users] new version of R2admb on R-forge

Thu Aug 19 10:27:36 PDT 2010

PS -- I confirm that the following test works cleanly on my system --
see quoted message below for platform, version of ADMB, etc. (I used the
defaults, clean=TRUE, safe=TRUE, and verbose=FALSE for this test ...).
So cleaning stuff up will require some kind of mucking around with
different versions of ADMB (I am happy to update to a newer version, and
require it for R2admb, if someone will tell me how to do so ...) -- in
the worst case scenario, which I would ***really*** rather not deal
with, testing for different platforms/versions and using workarounds ...

  cheers
    Ben

---------------------
# from Motoki Wu
library(R2admb)
setup_admb()

file.copy(file.path(Sys.getenv("ADMB_HOME"),
                    "examples","admb","simple","simple.tpl"),"./simple.tpl")
x <- rnorm(20, 5, 2)
y <- rnorm(20, 2, 9)
(d <- do_admb("simple",
        data = list(nobs = length(x), Y = y, x = x),
        params = list(a = 2, b = 1),
        checkparam = "ignore", checkdata = "ignore"))
summary(d)
unlink(c("simple","simple.tpl"))
--------------------

Ben Bolker wrote:
>   You can use safe=FALSE (the default is TRUE) and extra.args="-o", I
> believe.
> 
>   Is there a simple way (something like admb --version) to retrieve ADMB
> version info?  I believe I'm using revision 76 of the public SVN, which
> I *think* is identical to "admb-9.1-linux-gcc4.2-32bit.zip  ( ADMB 9.1
> Linux for gcc 4.2 (32 bit); Dec 2009)";
> 
> "svn log | head"
> 
>  gives
> 
> ------------------------------------------------------------------------
> r74 | jnancheta | 2009-12-31 20:28:45 -0500 (Thu, 31 Dec 2009) | 1 line
> 
> [Issue 1] Release of full ADMB source code.  Milestone Dec.2009 completed.
> 
>   (this makes me a little bit puzzled where r75 and r76 went, but never
> mind ...)
> 
>   I'll take a look at the other issue a little later today.
> 
>   Ben
> 
> John Sibert wrote:
>> The warning about the -s is caused by a change to the admb script making 
>> the "safe" option the default. So you do not need to specify -s to 
>> invoke the safe library. You need to use admb -o to build your 
>> application with the "optimized" (meaning no subscript checking mostly) 
>> library. I have no idea how to change the option in Ben's package.
>>
>> John
>>
>> On 08/18/2010 06:51 PM, Motoki Wu wrote:
>>> Hi Ben, thanks a lot for this.
>>>
>>> I am checking the code for the simple.tpl example in the original ADMB installation, and I am getting an error when I do do_admb(). When I follow the error message, it leads to the following portion of the code:
>>>
>>>      cred<- coutfile[!substr(coutfile, 1, 85) %in% c("cat: xxalloc4.tmp: No such file or directory",
>>>          "cat: xxalloc5.tmp: No such file or directory", "Error executing command cat xxglobal.tmp   xxhtop.tmp   header.tmp   xxalloc1.tmp   x")]
>>>      if (length(cred)>  0)
>>>          stop("errors detected in compilation: run with verbose=TRUE to view")
>>>
>>> For the variable cred, I get a warning:
>>>
>>>     ld: warning: option -s is obsolete and being ignored
>>>
>>> Getting rid of the -s argument should work, but also, shouldn't the negation (!) be removed to actually detect the error?
>>>
>>> Thanks
>>>
>>> - Motoki
>>>
>>> ps: my session (admb-9.1.556-macos10.6-xcode3.2-intel-32bit.dmg)
>>>
>>> x<- rnorm(20, 5, 2)
>>> y<- rnorm(20, 2, 9)
>>> do_admb("simple", data = list(nobs = length(x), Y = y, x = x), params = list(a = 2, b = 1), verbose = T, checkparam = "ignore", checkdata = "ignore", objfunname = "f", clean = F)
>>>
>>>    
>>>> R.Version()[c(1,13)]
>>>>      
>>> $platform
>>> [1] "x86_64-apple-darwin9.8.0"
>>>
>>> $version.string
>>> [1] "R version 2.11.1 (2010-05-31)"
>>>
>>>
>>> On Aug 18, 2010, at 6:14 PM, Ben Bolker wrote:
>>>
>>>    
>>>>   For any R users out there trying to get up the learning curve with
>>>> ADMB: I've put a new version of "R2admb" on r-forge (hence you should be
>>>> able to install from within R via
>>>> install.packages("R2admb",repos="http://r-forge.r-project.org") . It's
>>>> supposed to handle as much as possible of the work of passing R objects
>>>> to AD Model Builder, compiling/linking/running, getting the results back
>>>> into R, and analyzing them.  You still need to write the PROCEDURE
>>>> section of the TPL file, but R2admb can write most of the PARAMETER and
>>>> DATA sections automatically, on the basis of the data and parameter
>>>> lists you pass it.  The package includes a 'vignette' (R-speak for a PDF
>>>> document with examples); after installing and loading
>>>> [library("R2admb")] the package, you can access it via vignette("R2admb").
>>>>
>>>>   Bug reports, wish lists, comments are all welcome.
>>>>
>>>>   Ben Bolker
>>>>
>>>> _______________________________________________
>>>> Users mailing list
>>>> Users at admb-project.org
>>>> http://lists.admb-project.org/mailman/listinfo/users
>>>>      
>>> _______________________________________________
>>> Users mailing list
>>> Users at admb-project.org
>>> http://lists.admb-project.org/mailman/listinfo/users
>>>
>>>    
> 
> 

-- 
Ben Bolker
bbolker at gmail.com , bolker at mcmaster.ca
http://www.math.mcmaster.ca/~bolker
GPG key: http://www.math.mcmaster.ca/~bolker/benbolker-publickey.asc

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 261 bytes
Desc: OpenPGP digital signature
URL: <http://lists.admb-project.org/pipermail/users/attachments/20100819/510ecae6/attachment.pgp>