[yt-users] Area of Surface Cell

[Nathan Goldbaum]

Many apologies Matt, you're completely right.  I spoke too soon and missed
the obvious answer - shame on me for not trying it out first! The key is to
use either the triangles or vertices attributes of the surface object -
whichever makes it easier for you to think about it since they both contain
the same information about the surface.

I've pasted a fully worked example of what Melinda is trying to do here:

http://paste.yt-project.org/show/4904/

An equivalent script for yt-3.0 is pasted here:

http://paste.yt-project.org/show/4905/

I think the cython routine does the same thing as this - just in a manner
that is much faster and scales better in parallel.

On my machine this prints 0.99589582454 - I'm not quite sure why its not 1
- maybe there is some accumulation of floating point round-off error
somewhere?  FWIW, I get the same answer in 2.x and 3.0.

Interestingly, due to the issue from this trello card:

https://trello.com/c/SFuSSoIo

and the fact that surfaces need vertex centered data, 3.0 is substantially
slower at this than 2.x.

Anyway, hope that helps Melinda.

Nathan


On Sat, Jul 12, 2014 at 2:23 PM, Msoares.physics at gmail.com <
msoares.physics at gmail.com> wrote:

> Thanks, Michael. I did that last night, but it raised some concerns when
> the total flux I computed was off by a factor of 1.6 from the calculate
> flux command. I assumed the triangles were equal. Thanks again.
>
> On Jul 12, 2014, at 2:19 PM, Michael Zingale <
> michael.zingale at stonybrook.edu> wrote:
>
> This sound like what you want, in which case, I would also suggest taking
> the cross product of two edges to evaluate the normal direction of that
> local triangle.  I'll likely want to play with this later in the summer
> (after I'm done traveling), since this is good functionality for stellar
> simulations.
>
>
> On Sat, Jul 12, 2014 at 4:14 PM, Matthew Turk <matthewturk at gmail.com>
> wrote:
>
>> Hi Melinda and Nathan,
>>
>> I'm currently boarding my flight, but I just saw this. You can get the
>> area by examining the triangle vertices which hang off the surface object.
>> You can compute the area this way by using them as corners of the triangles
>> and inputting that into an area calculation.
>>
>> Matt
>> On Jul 12, 2014 2:40 PM, "Nathan Goldbaum" <nathan12343 at gmail.com> wrote:
>>
>>> Hi Melinda,
>>>
>>> Looking at the code, it looks like yt doesn't expose the individual flux
>>> or area values, only returning the total flux from the
>>> `calculate_isocontour_flux` function.
>>>
>>> That said, these values are calculated inline in the
>>> march_cubes_grid_flux function but are not cached or returned to the user.
>>>  In principle it should be possible to cache and return them to do the
>>> analysis you're trying to do, but it will require modifying yt.
>>>  march_cubes_grid_flux is defined in the cython file
>>> yt/utilities/lib/marching_cubes.pyx.
>>>
>>> I think this would be a generally useful modification to yt, so if you
>>> do go down the path of doing this we would certainly welcome your patch as
>>> a pull request.
>>>
>>> That said, since you only care about fluxes across a spherical surface,
>>> it should be possible to take advantage of the symmetry of the sphere to
>>> calculate the fluxes without doing the marching cubes isosurface
>>> calculation.  In the past, I've had a lot of success calculating fluxes
>>> across cylindrical surfaces by first exporting AMR data from yt to a
>>> uniform fixed resolution using a covering_grid data object and then
>>> postprocessing using custom code.
>>>
>>> Take a look at my radial_flux_analyzer, which does this calculation:
>>> https://bitbucket.org/ngoldbaum/galaxy_analysis_fork/src.
>>>
>>> I think it would be generally useful to extend what it does to spheres
>>> and other simple geometric shapes.  In principle it should also be possible
>>> to do the same fast calculation inside of yt its self using the
>>> multiresolution AMR data, but that has so far proven unnecessary for my
>>> purposes.
>>>
>>> Sorry to not have a more concrete solution for you.
>>>
>>> -Nathan
>>>
>>>
>>>
>>>
>>> On Sat, Jul 12, 2014 at 12:18 AM, Melinda Soares-Furtado <
>>> msoares.physics at gmail.com> wrote:
>>>
>>>> I created a surface using *surf = p1.h.surface(sp,"Radius",cluster_radius)*
>>>>
>>>> and then gathered the density values for the cells in this surface with
>>>>
>>>> *dens = surf['Density'], *which I can export as a textfile.
>>>>
>>>> I now need to get the area of each cell, but I'm having trouble doing this.
>>>>
>>>> Any advice?
>>>>
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> yt-users mailing list
>>>> yt-users at lists.spacepope.org
>>>> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>>>>
>>>>
>>>
>>> _______________________________________________
>>> yt-users mailing list
>>> yt-users at lists.spacepope.org
>>> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>>>
>>>
>> _______________________________________________
>> yt-users mailing list
>> yt-users at lists.spacepope.org
>> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>>
>>
>
>
> --
> Michael Zingale
> Associate Professor
>
> Dept. of Physics & Astronomy • Stony Brook University • Stony Brook, NY
> 11794-3800
> *phone*:  631-632-8225
> *e-mail*: Michael.Zingale at stonybrook.edu
> *web*: http://www.astro.sunysb.edu/mzingale
>
> _______________________________________________
> yt-users mailing list
> yt-users at lists.spacepope.org
> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>
>
> _______________________________________________
> yt-users mailing list
> yt-users at lists.spacepope.org
> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.spacepope.org/pipermail/yt-users-spacepope.org/attachments/20140712/a7626fd7/attachment.html>