[gt-users] some questions

[Gordon Gremme]

Hi Giorgio,

On Sun, Jul 26, 2009 at 10:52 PM, Giorgio Gonnella<ggonnell at yahoo.it> wrote:
> (1) About 2-3 months ago I printed a list of tools using 'gt -help'; one of them was 'gt exercise', which conversely had a long list of exercise tool. I was wondering why it is not present anymore. I am not missing anything specific, just wondering.

with the publication of the AnnotationSketch paper we properly
introduced libgenometools which allows to program against the
GenomeTools library instead of including every ``exercise'' into the
GenomeTools binary. Furthermore, these exercise tools didn't really
fit in with the overall system and we wanted to remove some old
baggage.


> (2) Is there a 'tool' you may use to extract subsequences of a given lenght?

Yes, the `gt extractseq` tool with the parameters -frompos and -topos.


> If the answer to the second question is 'no' and if it doesn't turn out to be too difficult for a complete beginner in genometools coding and almost C newbie like me, I was thinking to try taking some confidence to the project and implement such a functionality in a new tool as a first step to a basic sequencing reads simulator - I don't really need it, but I think is a simple enough thing to start with. Do you think it makes sense to do it?
>
> To do it I think I should: (is that right or am I missing some steps?)
> - clone the repo from git://genometools.org/genometools.git
> - open a new branch in my local repository
> - I was thinking of taking another simple tool, for example gt_seq_filter as an example
> - write my tool in a similar way / debug it
> - probably write some unit tests? are there tests for each tool?
> - check that everything still compiles correctly and that tests run correctly
> - when everything looks fine, ask you if would like to have a look at it for an eventual merge in the project

The overall approach would be correct.
Concerning the tests: We have unit tests for some classes and modules
(not for all of them, as it would be desirable).
They are implemented as single methods in C and are registered in the
gtt_unit_tests() methods in src/gtt.c
Most tools are tested by the testsuite in the testsuite/ directory
where we usually call the tool with different inputs and compare the
output to known results.

Gordon