[yt-dev] KeyError accessing field_units from profile?

Rasmi Elasmar re2300 at columbia.edu
Wed May 25 11:40:46 PDT 2016


Hi Nathan,

Thanks for your help. Compiling from source and changing the sphere size
did the trick.

I still can't seem to get the plots I want, though. I want three lines --
star, gas, and dark matter density as a function of radius. I tried
building up a list of labels and profiles using yt.create_profile(sp,
'radius', field) and doing yt.ProfilePlot.from_profiles(profiles,
labels=labels), but at the end I only end up with one line (dark matter,
which is presumably the last field iterated through). I tried stepping
through the code in IPython and it seems it's writing three plots (one for
each field) when I only want one (with three fields).

I also tried  yt.create_profile(sp, 'radius', fields=fields.values()) and
passing that to  yt.ProfilePlot.from_profiles(), and I got the same issue.
Same with yt.ProfilePlot(sp, 'radius', fields.values()) -- three plots are
saved, each one overwriting the previous. Any ideas?

Thanks,

Rasmi

On Mon, May 23, 2016 at 6:01 PM, Nathan Goldbaum <nathan12343 at gmail.com>
wrote:

> Hi Rasmi,
>
> Once I adjust the size of the sphere object you're using to be a bit large
> (in my case I made it 10 Mpc instead of 30 kpc), I'm able to run your
> script without errors. If I use 30 kpc, the script errors out since the
> sphere isn't big enough to accumulate values from more than a single cell
> in your simulation.
>
> I'm running the latest development version of yt - there were some issues
> related to profiles that were fixed relative to yt 3.2.3 in the development
> version - so I'd suggest switching to the development branch until yt 3.3
> is released.
>
> There is documentation on how to switch to the development version of yt
> here:
>
>
> http://yt-project.org/docs/dev/installing.html#installing-yt-using-pip-or-from-source
>
> If you're using anaconda, you may find using the conda nightly builds to
> be easier:
>
> http://yt-project.org/docs/dev/installing.html#nightly-conda-builds
>
> Hope that's helpful,
>
> Nathan
>
>
> On Mon, May 23, 2016 at 3:50 PM, Rasmi Elasmar <re2300 at columbia.edu>
> wrote:
>
>> I tried fooling around with the test data (the Enzo cosmology datasets in
>> particular) and couldn't find one that gave me a reasonable number of
>> halos. This is obnoxious, but here
>> <https://drive.google.com/open?id=0BwK-7Z3S5X_yMDNyZ1RTcG56SFU>'s a
>> ~3.5gb archive of the data I've been using (with halo catalog and test
>> script). If that's not useful, let me know!
>>
>>
>> Rasmi
>>
>> On Mon, May 23, 2016 at 2:41 PM, Nathan Goldbaum <nathan12343 at gmail.com>
>> wrote:
>>
>>> Hi Rasmi,
>>>
>>> Is there any chance you can make a small test script that triggers the
>>> error you're seeing, using one of the public test datasets on
>>> yt-project.org/data? There are several Enzo datasets there that should
>>> be usable.
>>>
>>> That will make it much easier for one of us to reproduce the error
>>> you're seeing.
>>>
>>> -Nathan
>>>
>>> On Mon, May 23, 2016 at 1:36 PM, Rasmi Elasmar <re2300 at columbia.edu>
>>> wrote:
>>>
>>>> Hi all,
>>>>
>>>> I'm using yt 3.2.3. I've created some derived fields (specifically
>>>> 'dark_matter' and 'stars'), and I'm trying to plot radial profiles of
>>>> ('gas', 'density'), ('deposit', 'dark_matter_density'), and ('deposit',
>>>> 'stars_density'). I'm using yt.create_profile(sphere, 'radius', field)
>>>> (where field is one of the above) and appending each of those profiles to a
>>>> list called profiles. Those are created fine, and accessing
>>>> profile[1].field_units (for example) gives '{('deposit', 'stars_density'):
>>>> g/cm**3}', which is what I expect.
>>>>
>>>> The issue occurs when I call yt.ProfilePlot.from_profiles(profiles,
>>>> labels=labels) -- I get an error in profile_plotter.py when accessing
>>>> profile.field_units on a profile:
>>>>
>>>>> Traceback (most recent call last):
>>>>>   File "/work/03330/tg826294/applications/scripts/halo_profiles.py",
>>>>> line 50, in <module>
>>>>>     plot = yt.ProfilePlot.from_profiles(profiles, labels=labels)
>>>>>   File
>>>>> "/work/03330/tg826294/applications/pythonenv/lib/python2.7/site-packages/yt/visualization/profile_plotter.py",
>>>>> line 415, in from_profiles
>>>>>     return cls._initialize_instance(obj, profiles, labels, plot_specs,
>>>>> y_log)
>>>>>   File
>>>>> "/work/03330/tg826294/applications/pythonenv/lib/python2.7/site-packages/yt/visualization/profile_plotter.py",
>>>>> line 365, in _initialize_instance
>>>>>     obj._setup_plots()
>>>>>   File
>>>>> "/work/03330/tg826294/applications/pythonenv/lib/python2.7/site-packages/yt/visualization/profile_plotter.py",
>>>>> line 339, in _setup_plots
>>>>>     xtitle, ytitle = self._get_field_title(fname, profile)
>>>>>   File
>>>>> "/work/03330/tg826294/applications/pythonenv/lib/python2.7/site-packages/yt/visualization/profile_plotter.py",
>>>>> line 633, in _get_field_title
>>>>>     y_unit = profile.field_units[field_y]
>>>>> KeyError: ('gas', 'density')
>>>>
>>>>
>>>> I tried explicitly passing the units in yt.create_profile(sphere,
>>>> 'radius', field, units={field: units[field]}) (where units is a dictionary
>>>> and units[field] gives 'g/cm**3'), but I still get the error. I also tried
>>>> adding the units to ds.field_units, and that didn't help either.
>>>>
>>>> You can see my code here
>>>> <https://gist.github.com/rasmi/6d936b077527e245025b73059e59b9c0>.
>>>>
>>>> I also tried creating all the profiles at once with:
>>>>
>>>>> profile = yt.Profile1D(sp, 'radius', 32, minimum, maximum, True,
>>>>> weight_field=None)
>>>>> profile.add_values(fields.values())
>>>>
>>>>
>>>> But everything was 0, and I got errors with other approaches (I can
>>>> detail them if you're curious, but I think the
>>>> yt.ProfilePlot.from_profiles() is probably the best approach, right?).
>>>>
>>>> Any ideas? Thanks for your help!
>>>>
>>>> Rasmi
>>>>
>>>> _______________________________________________
>>>> yt-dev mailing list
>>>> yt-dev at lists.spacepope.org
>>>> http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
>>>>
>>>>
>>>
>>> _______________________________________________
>>> yt-dev mailing list
>>> yt-dev at lists.spacepope.org
>>> http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
>>>
>>>
>>
>> _______________________________________________
>> yt-dev mailing list
>> yt-dev at lists.spacepope.org
>> http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
>>
>>
>
> _______________________________________________
> yt-dev mailing list
> yt-dev at lists.spacepope.org
> http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.spacepope.org/pipermail/yt-dev-spacepope.org/attachments/20160525/64b064a6/attachment.htm>


More information about the yt-dev mailing list