[Yt-dev] Ellipsoid update

Geoffrey So gsiisg at gmail.com
Wed Aug 10 18:05:30 PDT 2011


Hi Matt and Stephen,

I was able to fork YT (hg clone https://gsiisg@bitbucket.org/gsiisg/el-yt)
and implemented the

class AMREllipsoidBase(AMR3DData):

in data_containers.py, and right now I'm trying to put in the different
methods of finding the ellipsoid parameters from the halo parameters.  I'm
going to put in the geometry method first, and that requires the halo's

-center of mass x,y,z
-particle positions x,y,z
-grid edges left and right (for checking periodicity)

>From Matt's previous email:
"I think how this would work the best would be to make two separate
methods on the Halo base class."

I was wondering where to put the addition to Halo base class, should it be
in the halo_objects.py, or make a new file?

I'm still need to figure out (and this is where help would save me a ton of
time), how to access the two files inside a class in halo_objects.py.  Both
of which are created by

haloes.dump()

One of the file (.out) has the center of mass, and the other (.h5) has the
particle positions.  In my script they were either in memory already when I
ran the halo finder parallelHF, or loaded into memory by the single function
LoadHaloes().  I did a h5ls on the .h5 files and don't see the center of
mass information, so I assume it opened both the .out and .h5 and loaded
them into memory.

And last but not least, I need to make another "dump" mechanism, or add to
the existing one so it output the ellipsoid parameters.  Yes I did take
Britton's advice and cut out the e2xyz, e3xyz and use a single tilt angle to
replace them, which should cut out 5 columns of extra information.  The
ellipsoid object will take *center_of_mass, A,B,C,*e1vector,tilt_angle as
input, the ones with * means the input is a numpy array.

From
G.S.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.spacepope.org/pipermail/yt-dev-spacepope.org/attachments/20110810/4f023373/attachment.htm>


More information about the yt-dev mailing list