Visualization API slices disagree with plotted coordinates
Annelia Anderson
14 Nov '22
I am using the vis_query API to visualize gas density in 10 kpc slices. Upon comparing them with identical slices which plot gas coordinates from downloaded halo files, I find some discrepancies. The overall structure in halos is consistent, but there are some regions where the API visualization appears more dense, and some where the plotted coordinates appear more dense.
Could you provide any insight regarding this difference?
Additionally, would it be possible to have access to the code that vis_query uses to create visualizations to better understand its functionality?
Thank you!
Dylan Nelson
15 Nov '22
Are you using the "projection depth fraction" option to achieve a 10 kpc slice?
Note that this is a fraction, relative to the image size, e.g. its height or width. Would this explain the issue?
Unfortunately this visualization code is wrapped up in many other things at the moment - I hope to release it in the future, but can't at the moment.
You can of course switch the "render method" to "histo" in which case all that is happening is a np.histogram2d(), so you should be able to re-create the exact image in that case without difficulty. Note you can also download the raw data of the image, as an HDF5 file, to compare.
I am using the vis_query API to visualize gas density in 10 kpc slices. Upon comparing them with identical slices which plot gas coordinates from downloaded halo files, I find some discrepancies. The overall structure in halos is consistent, but there are some regions where the API visualization appears more dense, and some where the plotted coordinates appear more dense.
Could you provide any insight regarding this difference?
Additionally, would it be possible to have access to the code that vis_query uses to create visualizations to better understand its functionality?
Thank you!
Are you using the "projection depth fraction" option to achieve a 10 kpc slice?
Note that this is a fraction, relative to the image size, e.g. its height or width. Would this explain the issue?
Unfortunately this visualization code is wrapped up in many other things at the moment - I hope to release it in the future, but can't at the moment.
You can of course switch the "render method" to "histo" in which case all that is happening is a np.histogram2d(), so you should be able to re-create the exact image in that case without difficulty. Note you can also download the raw data of the image, as an HDF5 file, to compare.