Re: task pages: Moved many NGS packages from bio to bio-ngs
Hi Andreas,
On 22.09.17 11:47, Andreas Tille wrote:
> Hi Steffen,
>
> On Thu, Sep 21, 2017 at 07:29:47PM +0200, Steffen Möller wrote:
>> This effort is a bit moot since we should likely follow one of the
>> classifications that are already out there (like EDAM ontology topics or
>> OMICtools categories),
> I do not think that the effort is really moot since there is no code yet
> (and nobody volunteered to code this - I for myself feel not really able
> to since I'm lacking the biological background). So for the moment we
> need to rely on manual editing.
Agreed.
Should we put the auto-transfer of other-than-bio task pages from the
external catalogs on our agenda for the next Sprint? There should be
APIs out there to use for us.
>> but this will take some time. So, I finally got
>> around to manually moving a series of NGS-centric packages over to the
>> bio-ngs task. I have no exact clue about how quickly the task pages will
>> be updated, just possibly give it a check tomorrow about what I have
>> missed or about where I was too ambitious while doing the transfer.
> Tasks pages are rebuild twice per day and after a change of the tasks
> (+30min or so). So this should not last that long and the latest version
> is from
>
> Last update: Fri, 22 Sep 2017 05:44:29 -0000
>
> (see bottom line of the page).
>
> However, I'm not really sure whether your change is that lucky.
> *Currently* the idea behind the tasks pages is:
>
> bio: Dump *everything* related to bioinformatics in here
> Metapackage is created
> bio-ngs: List programs related to NGS ask task in the web
> **NO** Metapackage is created
> bio-phylogeny: List phylogeny related programs as task in the web
> **NO** Metapackage is created
> cloud: Everything from bio that has no GUI
> Metapackage is created
>
> Your change changes the philosophy behind bio and might make people
> wondering why all those NGS related packages might have vanished.
Right. And since about all sequencing is now performed with some NGS
method, we basically sort out everything that is about genomics, which
is not what we want.
>
> Moreover there is no install method any more to install NGS packages
> (at least if we would render metapackages from the current state).
>
> That's why I'm tempted to restore the old bio task with all packages
> included - at least as long as we have a better solution.
>
> What do you think?
This makes sense. And you should then also add all the packages that are
in bio-ngs but not in bio, I tend to think. Also we may be happier by
tagging packages within bio rather than maintaining multiple files.
Anyway ... let us get complete the references to catalog(ue)s in and
then see how this goes.
Best,
Steffen
Reply to: