[yt-users] reading Gadget-HDF5 format in yt-3.0

[Matthew Turk]

Hi Brendan,

On Fri, Sep 27, 2013 at 4:22 PM, Matthew Turk <matthewturk at gmail.com> wrote:
> Hi Brendan,
>
> Thanks for writing.
>
> On Thu, Sep 26, 2013 at 9:41 PM, Brendan Griffen
> <brendan.f.griffen at gmail.com> wrote:
>> Hi,
>>
>> I noticed in the update log of yt-3.0 you can now read Gadget HDF5 format
>> (for OWLS project), but it is possible for Rockstar to read the exact same
>> formats? I have some HDF5 data and I'd like to run Rockstar on it through yt
>> if it has this capability in 3.0.
>
> Good question!  So there are a couple of things to address in this.
>
> The first is that of the HDF5 formats for reading data.  I have issued
> a pull request, with Desika Narayanan's help, to add support for
> Gadget HDF5 data files, which are subtly different than the OWLS data
> files.  I would *very* much appreciate it if you could help by testing
> this out.  You can do so by doing:
>
> cd $YT_DEST/src/yt-hg   # or wherever "yt instinfo" tells you, if you
> have run the install script
> hg pull -r d996eff https://bitbucket.org/MatthewTurk/yt-3.0
> hg up -r d996eff
> python2.7 setup.py build_ext -i -f
>
> This will update you to that version of the code, and you should now
> be able to execute this command:
>
> from yt.mods import *
> ds = load("gadget_filename.hdf5")
>
> Things like projecting and analyzing particles and so on should now
> work.  *Please* let us know if this works for your data, and if so,
> click on the "approve" button or leave a comment on the pull request:
>
> https://bitbucket.org/yt_analysis/yt-3.0/pull-request/99/implement-gadget-hdf5-and-some-helpful/diff
>
> I think we really only need one more external tester (Desika has
> tested it) and we can bring it in.
>
> The second thing is about Rockstar.  I've been working on a pull
> request against my fork of Rockstar (
> http://bitbucket.org/MatthewTurk/rockstar ) that enables deeper and
> better yt integration; these changes have been submitted upstream so
> hopefully I won't need a fork for much longer and we can all go back
> to tarballs of R*.
>
> I've been testing this in my bookmark "rockstar_work" (which I haven't
> merged with HDF5-gadget work, sorry!) and I am getting weird results.
> The process is working, but the halo masses are scaled incorrectly and
> I am going to be investigating units to figure out the problem.  This
> is actually a regression, as it had previously well, and I will be
> investigating it in more detail.  I hope to have some results shortly.
>
> Let me know if you get a chance to check the Gadget-HDF5, and
> hopefully by then I'll have the Rockstar stuff up and running!
>

Okay, it turns out I was wrong and I was looking at unsorted halo
masses and wondering why they were all so low-mass.  Sorting fixed it.
 :)

I've issued a pull request here, which hopefully we can get
incorporated soon.  There are still additional things on my Rockstar
checklist, for what it's worth.

https://bitbucket.org/yt_analysis/yt-3.0/pull-request/105/rockstar-updates/diff

-Matt

> -Matt
>
>>
>> Are there any examples of carrying out such a process? Thanks.
>>
>> Thanks.
>>
>> Brendan
>>
>> _______________________________________________
>> yt-users mailing list
>> yt-users at lists.spacepope.org
>> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>>