Hi Shane,
All that you write makes sense, thanks. I was under the impression
that Length() would fail if the object is a single item rather than a
list, so in my own wrapper code (similar to yours) I first check
whether Length() has returned a list:
value <- Lookup( name, ParameterNames, ParameterValues )
if( IsList(value) ) {
SignalFatalError("Parameter "+name+" not found, or multiply defined")
}
return value
Anyway, there are always more ways to accomplish a result.
Thanks for the filename tip, it will definitely be useful!
Regarding including my tests with PEBL, I am not programming any
specific test, only a generic program that reads a couple of
configuration files to decide how to configure an experiment. I plan
to release it eventually but I wonder whether it would be appropriate
to distribute it with PEBL because it is not a finished product but
rather an infrastructure to develop experiments. Anyway, if you are
interested you can have a look at drghirlanda/alex on github. Mind
that the documentation is minimal right now...
Best,
Stefano
On Mon, Jan 26, 2015 at 9:10 AM, Shane Mueller <[email protected]> wrote:
> Stefano,
>
> Thanks for you feedback, (especially from a power user). If any of the
> tests you are developing would be appropriate for including in the PEBL
> test battery, please let me know and I'll see about trying to include
> them.
>
> For the file-name issue, you are in luck. I had the same concern, and for
> 0.14 I gave file streams a property called 'filename' that let's you
> access the actual file name. so, for example,
>
> Print("Opening tmp.dat for writing")
> xFile <- FileOpenWrite("tmp.dat")
> PrintProperties(xfile)
> xfile2 <- FileOpenWrite("tmp.dat")
> PrintProperties(xfile2)
> Print(xfile.filename)
> Print(xfile2.filename)
>
> The output should look something like this:
> Opening tmp.dat for writing
> Non-fatal Warning in line 7 of file ./files.pbl:
> File [tmp.dat] already exists. Using [tmp0.dat] instead
> ----------
> [DIRECTION]: WRITE
> [FILENAME]: tmp0.dat
> ....
>
> You should be able to access the filename with xfile2.filename. I'm not
> sure if this feature made it into the documentation though, so I'll put
> that on my list of things to do.
>
> In addition to this, you can always use FileOpenOverwrite(), which uses
> exactly the filename you specify (i.e., if it is for temporary files or
> something), and there is a new all-in-one function called GetNewDataFile()
> that I use in all the test battery tests. It handles headers, file types,
> re-use of subject codes and files for multiple sessions, separate
> subdirectories for each subject data file, and re-asks if you try to
> overwrite. This may handle some of your problems as well.
>
>
> Now, for the Lookup function, I have used this throughout the test battery
> for doing data recoding, and it is meant particularly for cases where you
> know the cue is one of the looked-up items. For example, if you want to
> recode '<lshift>' or '<rshift>' to "left" and "right". Even if it would
> make sense to do, I'm reluctant to change how it works as I think it might
> break things that would be hard to track down, especially if someone wants
> to upgrade to a new version of PEBL but use the same .pbl scripts they
> have before. But you can easily write a wrapper that does exactly what you
> want, and below I have an example that I use in the launcher:
>
> ## a robust version of lookup:
> define GetConfig(key, keys, database,empty)
> {
> line <- Lookup(key,keys,database)
> if(Length(line)==0)
> {
> ret <- empty
> }else{
> ret <- Second(line)
> }
> return ret
> }
>
>
> If it is empty, it will return whatever you pass into the 'empty'
> argument. If it is not empty, it returns the value of a paired list (you
> might need to adapt this). I'm not sure what happens if it matches
> multiple keys. I use this for setting default values in a parameter
> lookup. You may be able to use this directly or adapt it a bit for your
> needs.
>
> Shane
>
--
Stefano Ghirlanda - www.intercult.su.se/~stefano - drghirlanda.com
Full Professor of Psychology, Biology, and Anthropology, Brooklyn College
Founder and Fellow, Stockholm University Centre for the Study of
Cultural Evolution
------------------------------------------------------------------------------
Dive into the World of Parallel Programming. The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net/
_______________________________________________
Pebl-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/pebl-list
