[yt-users] OffAxisProjection axes problems

Nathan Goldbaum nathan12343 at gmail.com
Mon Jun 6 14:47:34 PDT 2016 

Hi Stephanie,

So I think I know what's happening here. Here's a reduced, simplified
version of your script that illustrates the issue:

http://paste.yt-project.org/show/6589/

In this example I've tried to construct the SlicePlot and
OffAxisProjectionPlot to be as similar as possible to ease comparison.

In your slices, when we you do:

ds.coordinates.x_axis[1] = 0
ds.coordinates.x_axis['y'] = 0
ds.coordinates.y_axis[1] = 2
ds.coordinates.y_axis['y'] = 2

You're explicitly asking for a non-right handed representation of the
simulation's coordinate system. Note that the SlicePlot and
OffAxisProjecitonPlot coordinate system *would* line up, if you didn't
adjust ds.coordinates.x_axis and ds.coordinates.y_axis, so long as you set
north_vector = [1, 0, 0].

The trouble for your use comes comes when you try to generate a plot with
an inverted coordinate system using OffAxisProjectionPlot. That doesn't
work because OffAxisProjectionPlot doesn't have a way to declare the
orientation of the plot coordinate system. As you said, if it were possible
to specify an east_vector (or a coordinate system orientation) you would be
able to work around this. Unfortunately, that's not supported right now.

Assuming there isn't anything special about the 'y' axis as opposed to the
'x' axis, I think a simpler way to handle this would be to make plots along
the 'x' axis instead of the 'y' axis. Here's an example script that
illustrates how the issues you're running into go away when you do that:

http://paste.yt-project.org/show/6590/

If you *need* to slice along 'y', *and* for the coordinate systems in the
two plots to have identical orientations, then we would need to modify
OffAxisProjectionPlot to add an API hook that causes the plot orientation
to be flipped. I think this is a totally reasonable thing to request as a
feature - it's pretty common that people want to do this for SlicePlot and
ProjectionPlot along the 'y' axis. Having a general interface for this for
all plots would be very useful.

-Nathan


On Mon, Jun 6, 2016 at 4:12 PM, Stephanie Tonnesen <stonnes at gmail.com>
wrote:

> Hi Nathan (and anyone reading this string!),
>
> Sorry it took a little while, but I wrote a short version of the code that
> shows the same problem/difference between the y slice and
> OffAxisProjection.  The difference is most easily seen comparing the
> velocity_x panels it writes out.  You will probably have to change the
> exact file path to wherever you keep your IsolatedGalaxy data, but
> otherwise it should happily run!  Also, just a note that the zoom is a bit
> extreme in the OffAxisProjection, but you still see the problem.
>
> Thanks,
> Stephanie
>
>
> --
> Dr. Stephanie Tonnesen
> Alvin E. Nashman Postdoctoral Fellow
> Carnegie Observatories, Pasadena, CA
> stonnes at gmail.com
>
> On Fri, Jun 3, 2016 at 3:44 PM, Nathan Goldbaum <nathan12343 at gmail.com>
> wrote:
>
>> Is there any chance you can make a self-contained example using one of
>> the test datasets on yt-project.org/data? That will make it easier for
>> one of us to take a shot at figuring out what's going wrong. I usually use
>> "IsolatedGalaxy" - it has a disk in the center of the simulation box with
>> an angular momentum vector aligned with the z axis.
>>
>> On Fri, Jun 3, 2016 at 5:10 PM, Stephanie Tonnesen <stonnes at gmail.com>
>> wrote:
>>
>>> Right, I have my north vector, which I *think* is why my z-axis (Image
>>> y) looks correct.  Unfortunately I can't specify an east_vector.  I will
>>> keep playing and if I come up with something clever I will let ppl know.
>>>
>>> Thanks for thinking about this with me!
>>>
>>> Stephanie
>>>
>>>
>>>
>>>
>>>
>>> --
>>> Dr. Stephanie Tonnesen
>>> Alvin E. Nashman Postdoctoral Fellow
>>> Carnegie Observatories, Pasadena, CA
>>> stonnes at gmail.com
>>>
>>> On Fri, Jun 3, 2016 at 2:10 PM, Nathan Goldbaum <nathan12343 at gmail.com>
>>> wrote:
>>>
>>>>
>>>>
>>>> On Fri, Jun 3, 2016 at 4:08 PM, Stephanie Tonnesen <stonnes at gmail.com>
>>>> wrote:
>>>>
>>>>> Hi Nathan,
>>>>>
>>>>>
>>>>> So I tried two of your ideas, and don't get the last one.
>>>>>
>>>>> ds.coordinates.x_axis[1] = 0
>>>>>>     ds.coordinates.x_axis['y'] = 0
>>>>>>     ds.coordinates.y_axis[1] = 2
>>>>>>     ds.coordinates.y_axis['y'] = 2
>>>>>>
>>>>>
>>>>> Is the orientation correct (albeit with z along the horizontal axis
>>>>> and x along the vertical axis) if you don't do this ^ ?
>>>>>
>>>>> If I don't do that but still have the same L and north_vector, I get
>>>>> the same results.
>>>>>
>>>>> If I change L = [0,1,0] and north_vector=[0,0,1], I get the same
>>>>> results as well.
>>>>>
>>>>> Ugh!
>>>>>
>>>>> I am not sure what "x" and "z" fields you think I should be making
>>>>> OffAxisProjection s for?  Do you mean have L = [1,0,0] and [0,0,1], or have
>>>>> it project an "x" and "z" field onto my off-axis projection?
>>>>>
>>>>
>>>> 'x' and 'z' don't really have a meaning for OffAxisProjectionPlot,
>>>> that's why they're dispalyed as "image x" and "image y". However, if you
>>>> specify a north_vector, that should disambiguate things.
>>>>
>>>> It's entirely possible there's a bug here.
>>>>
>>>>
>>>>>
>>>>> Thanks,
>>>>>
>>>>> Stephanie
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Dr. Stephanie Tonnesen
>>>>> Alvin E. Nashman Postdoctoral Fellow
>>>>> Carnegie Observatories, Pasadena, CA
>>>>> stonnes at gmail.com
>>>>>
>>>>> On Thu, Jun 2, 2016 at 5:14 PM, Nathan Goldbaum <nathan12343 at gmail.com
>>>>> > wrote:
>>>>>
>>>>>>
>>>>>>
>>>>>> On Thu, Jun 2, 2016 at 6:45 PM, Stephanie Tonnesen <stonnes at gmail.com
>>>>>> > wrote:
>>>>>>
>>>>>>> Hi all,
>>>>>>>
>>>>>>> The quick summary of what I want to do is switch the coordinates of
>>>>>>> the x-axis on a OffAxisProjection plot.  Basically, what I am trying to do
>>>>>>> is get a LOS look at a galaxy , so I need a 20 degree tilt in the y-z
>>>>>>> plane.  For some reason this is swapping the x-axis direction from my plain
>>>>>>> old y slice.  The relevant parts of my code look like this:
>>>>>>>
>>>>>>> ds.coordinates.x_axis[1] = 0
>>>>>>>     ds.coordinates.x_axis['y'] = 0
>>>>>>>     ds.coordinates.y_axis[1] = 2
>>>>>>>     ds.coordinates.y_axis['y'] = 2
>>>>>>>
>>>>>>
>>>>>> Is the orientation correct (albeit with z along the horizontal axis
>>>>>> and x along the vertical axis) if you don't do this ^ ?
>>>>>>
>>>>>> What happens if you make an off-axis projection aligned with the y
>>>>>> axis, i.e. L = [0, 1, 0], north_vector = [0, 0, 1]?
>>>>>>
>>>>>> Can you try making an off-axis projection of the 'x' and 'z' fields?
>>>>>>
>>>>>>
>>>>>>>
>>>>>>>     plot6=yt.SlicePlot(ds,"y","density")
>>>>>>>     plot6.save()
>>>>>>>
>>>>>>>     plot8=yt.SlicePlot(ds,"y","velocity_z")
>>>>>>>     plot8.save()
>>>>>>>
>>>>>>>     plot9 = yt.SlicePlot(ds,"y","velocity_y")
>>>>>>>     plot9.save()
>>>>>>>
>>>>>>>     plot10 = yt.SlicePlot(ds,"y","velocity_x")
>>>>>>>     plot10.save()
>>>>>>>
>>>>>>>     ds.periodicity=(True,True,True)
>>>>>>>
>>>>>>>     L = [0.0,0.93969,-0.342]
>>>>>>>     #W = [0.5,0.5,0.5]
>>>>>>>     #c = [0.5,0.5,0.5]
>>>>>>>     #N = 1028
>>>>>>>     north_vector = [0,0.342,0.93969]
>>>>>>>     plot12 =
>>>>>>> yt.OffAxisProjectionPlot(ds,L,'density',width=(100,'kpc'),north_vector=north_vector)
>>>>>>>     plot12.save()
>>>>>>>
>>>>>>>     plot13 =
>>>>>>> yt.OffAxisProjectionPlot(ds,L,'velocity_x',width=(100,'kpc'),north_vector=north_vector)
>>>>>>>     plot13.save()
>>>>>>>     plot14 =
>>>>>>> yt.OffAxisProjectionPlot(ds,L,'velocity_y',width=(100,'kpc'),north_vector=north_vector)
>>>>>>>     plot14.save()
>>>>>>>     plot15 =
>>>>>>> yt.OffAxisProjectionPlot(ds,L,'velocity_z',width=(100,'kpc'),north_vector=north_vector)
>>>>>>>     plot15.save()
>>>>>>>
>>>>>>> I have attached four files:  the slices of velocity_y and velocity_x
>>>>>>> and the OffAxisProjections of the same.  You can see that the x-axis seems
>>>>>>> to be flipped.  You will also notice that the OffAxisProjection is more
>>>>>>> zoomed than the Slice, but don't let this distract you :)
>>>>>>>
>>>>>>> I have no idea what has gone wrong here.  Any help, as always, is
>>>>>>> really appreciated.
>>>>>>>
>>>>>>> Thanks,
>>>>>>> Stephanie
>>>>>>> --
>>>>>>> Dr. Stephanie Tonnesen
>>>>>>> Alvin E. Nashman Postdoctoral Fellow
>>>>>>> Carnegie Observatories, Pasadena, CA
>>>>>>> stonnes at gmail.com
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> 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
>>>>>
>>>>>
>>>>
>>>> _______________________________________________
>>>> 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
>>
>>
>
> _______________________________________________
> 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/20160606/22cd2b09/attachment.html>

More information about the yt-users mailing list