[Neurodebian-devel] How can Blends techniques be adapted to NeuroDebian

[Andreas Tille]

Hi Michael,

On Tue, Aug 21, 2012 at 09:59:53PM +0200, Michael Hanke wrote:
> > To make my point clear:  I would rather prefer to create NeuroDebian-ish
> > web pages not only for NeuroDebian but for any Blend - I can really not
> > assume that the NeuroDebian user audience is the only one who deserves
> > good user oriented pages.
> 
> Fair enough. Let see how we can proceed. I guess we would all be happy
> if NeuroDebian starts using as much blends framework (code) as possible.
> I was thinking at which level of information exchange/aggregation is
> most useful, while avoid duplication with services like UDD/DDE.
> 
> I believe that we would benefit most at the level where all the
> package-information that is present on the blends pages is processed to
> a degree where it is ready for insertion into a (language-agnostic) page
> template. In the blends pages case the is HTML, in our case this is RST.

The tasks (and bugs) pages of the Blends web sentinel are actually using
Genshi templates (to finally produce HTML).  As you know I considered
some way to export RST via Genshi but this does not seem to be possible
directly.

> If we can establish this layer as an interface, we could write/change
> our code to create a portal like neuro.debian.net from this information.

This is what we discussed in Grenoble.  Due to my experiments Genshy
just outpust some xml and you either need to postprocess the data to
create rst or you just need to choose an alternative templating system.
What I do have is some Python code that holds the Blends data.  If there
is no reasonable templating system I could write an rst_output method
which spits out the data you need to a file.

> NeuroDebian could then switch to blend data as its primary source of
> information and discontinue any non-neuroscience specific data
> aggregation. If we find additional information is needed/useful we could
> directly integrate additional data grabber/aggregators into the blends
> framework.

Yes.

> And of course, if the blends code breaks, more people should
> be motivated to fix it ;-)

Damn, NeuroDebian people do actually have mind reading abilities! ;-)
 
> While this doesn't immediately give us the all-in-wonder web portal for
> any blends, it would help focus efforts on the important bits. If we can
> generate neuro.debian.net by simply pointing a script to a specific
> blend (which would be the upcoming neuro-* blend), we should be able to
> do it for any other one with relatively low effort.

Exactly.

> In terms of how to make this data accessible: Right now we are using
> JSON, but pretty much anything like that should do it.
> 
> What do you think?

I think if you don't know a fitting templating system I could write a
rst_output method and we might proceed from here.

Kind regards

      Andreas.

-- 
http://fam-tille.de