Hi Suoqing, Okay, I think I have an idea. I believe other people have done this before. What you can do is compute the surface density, then supply this as a "field_parameter" to your data. For instance, you can compute a 2D grid of the surface density using the projection tool you already are, and convert this to a "fixed resolution buffer". Something like: proj = pf.h.proj(... frb = proj.to_frb(... Then the frb["Density"] value will be a 2D buffer, possibly transposed (i.e., you may need to do .transpose() on it), that corresponds to the surface density resulting from the projection. You can then set this as a field parameter: data_source = pf.h.all_data() data_source.set_field_parameter("surface_density", frb["Density"]) Now, in a derived field, you can access this: def my_field(field, data): surface_density = data.get_field_parameter("surface_density") ... And you should be able to project with this: pf.h.proj(..., source=data_source) Note that in 3.0, "source" becomes "data_source", but otherwise it should mostly be the same. I'd strongly encourage you to try this out, but also to test that your x,y values in the surface_density variable are correctly set up -- as in, that the origin is in the correct place for where you are, rather than with the x,y axes flipped or with the origin in upper left or something. Good luck, and let us know if it works, -Matt On Tue, Mar 18, 2014 at 5:27 PM, Suoqing JI wrote:

Hi Matt,

Thanks for your response!

I want to integrate the density from -zmin to +zmin for every cells located on (x, y), or in your words, on (r ,theta). So the cells with the same (x, y) location will have the same value of surface density, as the figures show. The surface density is a function of only (x, y).

Best wishes, -- Suoqing JI Ph.D Student Department of Physics University of California, Santa Barbara CA 93106, USA

On Mar 18, 2014, at 7:38 AM, Matthew Turk wrote:

Hi Suoqing,

If I understand correctly, you want to compute the *local* surface density (i.e., the integrated density up to that height) for each cell? As in, re you defining it such that it's a function of r, theta, z, or are you computing the surface density *once* for a disk object and using that (i.e., function of r, theta)?

-Matt

On Tue, Mar 18, 2014 at 2:24 AM, Suoqing JI wrote:

Hi,

I'm working on FLASH 3D Cartesian AMR data, and would like to define a derived field of surface density, so I can use surface density field to calculate other derived fields.

My current script does give desired results (slice_y: http://i.imgur.com/XigIYJc.png slice_z: http://i.imgur.com/kA2Fmlt.png, and the disk is cut from original data). However, it's extremely inefficient, because I cast a ray through disk height for every cells located on x-y plane surface of each AMR block.

So is there any clever way to define surface density as a derived field in YT, as what the figures attached show? Or some way without defining surface density as a derived field, but I can get access to surface density values when computing other derived fields?

Thanks a lot!

Best wishes, -- Suoqing JI Ph.D Student Department of Physics University of California, Santa Barbara CA 93106, USA

_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org

_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org

_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org