[yt-users] volume rendering annotation error for a derived field

[Nathan Goldbaum]

On Sat, Aug 27, 2016 at 8:43 PM, tazkera haque <h.tazkera at gmail.com> wrote:

> Hi yt people,
>
> I just updated to yt 3.3.1 to use all the new features of volume
> rendering. I am trying to render ejected density which is a derived field
> and draw the coordinate triads, domain boundary along with labels showing
> the log range of ejected density. my code is attached. While my code worked
> fine with default field 'density', the following figure was produced for
> 'density', but I am getting error when the source field is set to
> 'ejected_density'.
>

Hi Tazkera,

First, very nice rendering!

As for your error, this is due to a bug in yt --- thank you for the report!
I have issued a pull request so this bug can be fixed for yt 3.3.2:

https://bitbucket.org/yt_analysis/yt/pull-requests/2351

For your specific use case, the easiest way to work around this without
applying my bugfix is to change how you're creating the ejected_density
derived field. Instead of using the @derived_field decorator, it will work
correctly if you use the ds.add_field function. To make that explicit,
where you currently have:

    @derived_field(name = "ejected_density", units = "g/cm**3")
    def _get_ejected_density(field, data):
        E = 0.5* data["cell_mass"]*
(data["velx"]**2+data["vely"]**2+data["velz"]**2)+
data["gpot"]*data["cell_mass"]
        return ((np.array(E)>=0)*1)*data["density"]

    ds = yt.load(...)

You will instead have:

    def _get_ejected_density(field, data):
        E = 0.5* data["cell_mass"]*
(data["velx"]**2+data["vely"]**2+data["velz"]**2)+
data["gpot"]*data["cell_mass"]
        return ((np.array(E)>=0)*1)*data["density"]

    ds = yt.load(...)

    ds.add_field(function=_get_ejected_density, name='ejected_density',
units='g/cm**3')

What's happening is that if you use the @derived_field decorator, yt's
field machinery doesn't associate the derived field you add to a specific
dataset. This is normally fine, but when the field tries to generate a nice
LaTeX label for the field, it assumes that a dataset is available (e.g. for
fields with units like 'code_mass/code_length**3' where we need to have a
dataset available to know what 'code_mass' or 'code_length' mean). This
isn't true for locally defined fields, causing the crash you're seeing.

Sorry about the need for the workaround, but thank you again for the
detailed report and script, which were key for me to be able to see what
was going wrong and reproduced the bug locally.

Also the 'density' figure still does not show either the coordinate triads
> or the domain boundary. I would appreciate your help to fix the error and
> thanks in advance.
>

This is probably because the alpha values you're setting in annotate_axes
and annotate_domain are too low for this rendering. Try increasing them.


>
> *Traceback (most recent call last):*
> *  File "volume_render_test.py", line 117, in <module>*
> *    text_annotate=[[(.1, 1.05), text_string]])*
> *  File
> "/work/03858/thaque56/sw/yt-new-3.3/yt-conda/lib/python2.7/site-packages/yt/visualization/volume_rendering/scene.py",
> line 395, in save_annotated*
> *    label = rs.data_source.ds._get_field_info(rs.field).get_label()*
> *  File
> "/work/03858/thaque56/sw/yt-new-3.3/yt-conda/lib/python2.7/site-packages/yt/fields/derived_field.py",
> line 225, in get_label*
> *    units = Unit(self.units, registry=self.ds.unit_registry)*
> *AttributeError: 'NoneType' object has no attribute 'unit_registry'*
>
>
> *Thanks*
> *Tazkera*
>
> _______________________________________________
> 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/20160828/c7d3ff9b/attachment.htm>