[yt-users] Cosmological parameters necessary to run synthetic spectrum?

Nicholas Earl nchlsearl at gmail.com
Fri Dec 12 10:58:01 PST 2014 

Well done! This is excellent, it works great. I'm glad you were able to
work around some of my naive assumptions -- sorry about those.

Beautiful absorption lines, indeed! Thanks for taking the time to help out
like this -- you rock!

On Fri Dec 12 2014 at 8:10:48 AM Britton Smith <brittonsmith at gmail.com>
wrote:

> Hi Nick,
>
> Ok, I think we're now sorted.  There was a combination of issues at play
> here.
>
> In your script, some of the fields were not in the correct units, so the
> column densities were not coming out right.  Since yt is now unit aware,
> one has to make sure that field arrays have the correct units.  For
> example, the density and dl arrays here were being written in code units
> while the spectrum generator was expecting CGS.  Additionally, I told you
> to write out the density field, when I really mean number density.  Sorry
> about that!
>
> On top of that, it appears that the absorption spectrum generator itself
> was not truly yt-3.0 ready.  I apologize for that as I was the one who
> signed off on it.  I have issued a pull request that makes things work for
> real.  You can check it out here:
>
>
> https://bitbucket.org/yt_analysis/yt/pull-request/1356/bugfix-updating-absorption-spectrum
>
> Finally, after fixing all of this, I was still getting a spectrum free of
> absorption.  I determined that at the spectral resolution given, all of the
> line components were unresolved, meaning that each individual line was less
> than the size of one spectral element.  Previously, the behavior was to
> quietly ignore these components, which I can see now is the wrong thing to
> do.  I have added a check for this, and when encountered, the code will
> issue a warning and tell you to increase the resolution.
>
> After all this, you now have a beautiful absorption line:
> [image: Inline image 1]
>
> Here is a copy of your script with the few changes I made.
>
> http://paste.yt-project.org/show/5258/
>
> That and the changes from the pull request should take care of it.  Sorry
> for inconvenience!
>
> Britton
>
> On Thu, Dec 11, 2014 at 1:51 AM, Nicholas Earl <nchlsearl at gmail.com>
> wrote:
>
>> Hey all,
>>
>> Britton: thanks a lot for the run down -- I really appreciate it! I seem
>> to be running into an issue, however: the spectrum builder doesn't want to
>> give me anything other than a flat line for the flux.
>>
>> Testing the min/max values the ray returns seems to indicate everything
>> is pretty much in working order; the temperature, densities, etc seem
>> reasonable. I'm hoping perhaps you guys may be able to spot something
>> I've missed.
>>
>> For reference, here's is the current script I'm using to try and generate
>> the spectrum: https://gist.github.com/nmearl/85c70d6f8dc8e9e092a6. And
>> here is a single output of the test data I'm playing with:
>> https://www.dropbox.com/s/ldq4xyfnbzbbcj0/test_run.tar.gz?dl=0.
>>
>> Thanks,
>> Nick
>>
>> On Tue Dec 09 2014 at 9:51:00 AM Britton Smith <brittonsmith at gmail.com>
>> wrote:
>>
>>> Hi Nick,
>>>
>>> There are a number of places in the LightRay tool that assume that the
>>> simulation data being used is cosmological.  This is certainly something
>>> that could and should be improved, but it will be a non-trivial upgrade.
>>> Your best bet is to bypass the LightRay tool and manually create files that
>>> can be used by the absorption spectrum generator.  This should be quite
>>> straightforward.
>>>
>>> The LightRay tool works by creating ray data containers, gathering field
>>> data for all the cells intersected by the ray, and saving them to an hdf5
>>> file.  If you have some dataset loaded, you can do:
>>>
>>> import yt
>>> ds = yt.load(...)
>>> ray = ds.ray(start_position, end_position)
>>>
>>> Then, you can get fields you want to use in your spectrum.  You will
>>> want the density fields for whatever species you want to make an absorption
>>> line for.  You will also want to get the temperature field, since that
>>> controls the width of the line.  You will also need to know the physical
>>> path length of the ray through each cell.  This is multiplied by the
>>> density fields to get column densities that are then used to generate the
>>> Voigt profiles that make the absorption lines.  The ray object has an extra
>>> field called "dts" which is that same length normalized such that the total
>>> sum over the ray is 1.0.  Just multiply that by the physical length of the
>>> ray and you'll have what you need.
>>>
>>> If you want the lines to be shifted via the Doppler effect, you will
>>> have to calculate the line of sight velocity, which is just the dot product
>>> of the velocity field with your line of sight vector.
>>>
>>> Once you have all of these arrays created for your ray, just write them
>>> out to an hdf5 file.  Name the line of sight velocity field "los_velocity"
>>> and the path length field "dl" (lowercase L).  The absorption spectrum
>>> generator will also look for a redshift field, so just write out an array
>>> of zeros to the hdf5 file with the name "redshift".  If you have made the
>>> hdf5 file in this way, you can feed it to the absorption spectrum generator
>>> and it will work.
>>>
>>> To summarize, you need to write out the following fields:
>>> - densities (for each species you want to make lines for)
>>> - temperature
>>> - line of sight velocity "los_velocity" (optional)
>>> - path length "dl"
>>> - array of zeros called "redshift"
>>>
>>> Finally, I highly suggest you do this in yt-3.0.  While there may be
>>> bugs (although I believe all are fixed as of now), it will be much easier
>>> for us to fix them in 3.0.  The code has changed significantly since
>>> version 2.x, so porting fixes back is not easy.
>>>
>>> Britton
>>>
>>>
>>> On Tue, Dec 9, 2014 at 2:06 AM, Nicholas Earl <nchlsearl at gmail.com>
>>> wrote:
>>>
>>>> Thanks, everyone, for the responses.
>>>>
>>>> It is as Brian has said: I am dealing specifically with the case of an
>>>> absorption spectrum generator. The simulation is indeed idealized and
>>>> non-cosmological.
>>>>
>>>> Attempting to load the data and pass it to SpectrumBuilder yields the
>>>> 'CosmologyInitialRedshift' KeyError seen here
>>>> http://paste.yt-project.org/show/5248/. Editing the sfr_spectrum.py
>>>> file to accept hard-coded values (for the omegas and initial redshift)
>>>> yields a second KeyError seen here
>>>> http://paste.yt-project.org/show/5249/.
>>>>
>>>> As per Nathan's suggestion, I will give the yt-3.0 version of the
>>>> generator a go in the mean time.
>>>>
>>>> Thanks again for your guys' attention,
>>>> Nick
>>>>
>>>> On Mon Dec 08 2014 at 5:19:21 PM Cameron Hummels <chummels at gmail.com>
>>>> wrote:
>>>>
>>>>> I've used the code on individual blobs of gas in an ambient medium
>>>>> with short sightlines, but I used it on sim outputs that were
>>>>> cosmological.  I think it would be relatively straightforward to generalize
>>>>> the code to allow one to *set* the cosmological parameters, instead of
>>>>> reading those values from the dataset.  I think Britton would be the main
>>>>> person to answer these questions, but I'll take a look at it this evening
>>>>> and see how hard this might be.
>>>>>
>>>>> Cameron
>>>>>
>>>>> On Mon, Dec 8, 2014 at 2:59 PM, Nathan Goldbaum <nathan12343 at gmail.com
>>>>> > wrote:
>>>>>
>>>>>> Hi Brian and Nick,
>>>>>>
>>>>>> Can you share with us the errors you're seeing?
>>>>>>
>>>>>> It's likely that no one has tried to do what you're doing before.  To
>>>>>> get it to work, it might just be necessary to make some adjustments to the
>>>>>> absorption spectrum machinery in the case when we detect a non-cosmological
>>>>>> dataset. That will require making some modifications to yt, but that's ok!
>>>>>>
>>>>>> It looks like the absorption spectrum analysis module *has* seen some
>>>>>> updates for yt-3.0, so you may also want to try again in the latest version
>>>>>> of the codebase.  I haven't personally used it - it looks like Hillary Egan
>>>>>> was the last person to touch the code.
>>>>>>
>>>>>> -Nathan
>>>>>>
>>>>>> On Mon Dec 08 2014 at 1:52:15 PM Brian O'Shea <bwoshea at gmail.com>
>>>>>> wrote:
>>>>>>
>>>>>>> Hi folks,
>>>>>>>
>>>>>>> Just to send a quick followup on this: Nick is referring
>>>>>>> specifically to the *absorption spectrum* generator, and he is trying to
>>>>>>> make spectra of idealized, non-cosmological simulations (think a blob in an
>>>>>>> ambient medium) that has no redshift as a simulation parameter.  The
>>>>>>> question is, how does one do this?
>>>>>>>
>>>>>>> Cheers,
>>>>>>> Brian
>>>>>>>
>>>>>>>
>>>>>>> On Mon, Dec 8, 2014 at 11:59 AM, Nicholas Earl <nchlsearl at gmail.com>
>>>>>>> wrote:
>>>>>>>
>>>>>>>> Hey everyone,
>>>>>>>>
>>>>>>>> Is there a way to fake the cosmological parameters that the
>>>>>>>> synthetic spectrum generator looks for in a simulation? Or rather, a
>>>>>>>> "sanctioned" way? Having the package reference hardcoded values when the
>>>>>>>> parameters themselves are not found yields other missing fields like
>>>>>>>> 'creation_time'. I am wondering if this is a consequence of the simulation
>>>>>>>> not being inherently cosmological?
>>>>>>>>
>>>>>>>> I am using yt 2.x currently, as I'm aware that there seems to be an
>>>>>>>> issue with the generator in the 3.0 release.
>>>>>>>>
>>>>>>>> Thanks for your time,
>>>>>>>> Nick
>>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>>> 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
>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Cameron Hummels
>>>>> Postdoctoral Researcher
>>>>> Steward Observatory
>>>>> University of Arizona
>>>>> http://chummels.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/20141212/1754fcd2/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: now_it_works.png
Type: image/png
Size: 15155 bytes
Desc: not available
URL: <http://lists.spacepope.org/pipermail/yt-users-spacepope.org/attachments/20141212/1754fcd2/attachment-0001.png>
-------------- next part --------------
_______________________________________________
yt-users mailing list
yt-users at lists.spacepope.org
http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org

More information about the yt-users mailing list