What is a meta_playground?
Can we play in just a playground ?
yeah but the word "playground" is already used by Ketrew (for the Host.t's temp-files and so on).
I started with meta-playground a bit as a joke to delay thinking about naming things...
I don't know what to use instead. Any suggestions?
I think "work" directory is reasonable? We're somehow using that for a subdirectory in the meta_playground, but I don't think we need a separate name for that subdirectory.
On Wed, Mar 16, 2016 at 6:06 PM Sebastien Mondet [email protected] wrote:
yeah but the word "playground" is already used by Ketrew (for the Host.t's temp-files and so on).
I started with meta-playground a bit as a joke to delay thinking about naming things...
I don't know what to use instead. Any suggestions?
— You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub https://github.com/hammerlab/biokepi/issues/174#issuecomment-197573286
work_dir ?
The "work" directory you (@ihodes) mention is the work_dir argument of the Pipeline compiler:
https://github.com/hammerlab/biokepi/blob/master/src/lib/pipeline.ml#L497
we usually put it inside the the meta_playground (to be able to rm -fr <meta-playground> :) )
@rleonid so yes work_dir is also already taken :)
Can't the Pipeline compiler use the meta playground? Why does it need its own directory?
On Wed, Mar 16, 2016 at 6:14 PM Sebastien Mondet [email protected] wrote:
The "work" directory you (@ihodes https://github.com/ihodes) mention is the work_dir argument of the Pipeline compiler: https://github.com/hammerlab/biokepi/blob/master/src/lib/pipeline.ml#L497
we usually put it inside the the meta_playground (to be able to rm -fr
:) ) @rleonid https://github.com/rleonid so yes work_dir is also already taken :)
— You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub https://github.com/hammerlab/biokepi/issues/174#issuecomment-197579621
@ihodes we could force the compiler's work_dir to be a subdir of meta_playground.
But since the meta_playground otherwise contains more "sample-independent" sutff (software and reference-data); I though it would be handy to separate the 2 (more hackable ways of saving disk-space and/or avoid recomputations for example by sharing among users).
I agree.
@iskander came over and we all chatted; each directory is a work_dir in their specific contexts; so they can all be called work_dir, and disambiguated when necessary with prefixes.
On Wed, Mar 16, 2016 at 6:20 PM Sebastien Mondet [email protected] wrote:
@ihodes https://github.com/ihodes we could force the compiler's work_dir to be a subdir of meta_playground. But since the meta_playground otherwise contains more "sample-independent" sutff (software and reference-data); I though it would be handy to separate the 2 (more hackable ways of saving disk-space and/or avoid recomputations for example by sharing among users).
— You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub https://github.com/hammerlab/biokepi/issues/174#issuecomment-197582215
@smondet the work_dir in the Pipeline is used as a prefix of where results are written. AFAIK, we don't actually "work" in that directory. So I think it is misnamed. How about calling that results_path?
The meta_playgrounds also don't necessarily have anything executed in them. I think we use them as a destination of where we install tools. How about calling those install_path ?
We can go back and change the original ketrew playgrounds to actual work_dir's if we really want.
@rleonid it's more than the results; the intermediary files go there; and some tools have to cd there (like the indel-realigner OTOH).
the meta_playground so far has the tools and the reference genomes' data (FASTAs + databases like COSMIC).
some tools have to run in there to pre-process the reference genome (like samtools faidx, bwa index, picard-create-dict) → that preprocessing is reused across samples/runs/experiments.
the meta_playground so far has the tools and the reference genomes' data (FASTAs + databases like COSMIC).
isn't this more reason to use a more specific name? I find meta_playground very uninformative.
isn't this more reason to use a more specific name? I find meta_playground very uninformative.
Yes, me too :)