Added the ability to read xdmf dumps from Idefix

[dutta-alankar]

@neutrinoceros As discussed a few days ago, here is the first implementation of a working reader to parse Idefix xdmf dumps.

While developing the HDF5 IO module on Idefix, I have taken a special attention to make the files contain additional attributes that make them self-contained and hence don't need definitions.hpp or idefix.ini to process the data. I am aware that my implementation is far from being the cleanest, and simply does the job for the time being. But I believe that it is a good head start and can be trimmed and matured to a polished version. I'm open to take in comments to improve this.

To test, I'm using the following code with a KHI data dump from the following PR. I'm also sharing the data used in this test here in this link.

import yt
import yt_idefix

num = 1
ds = yt.load(f"data.{num:04d}.flt.h5")

# Create gas density slices in all three axes.
p = yt.SlicePlot(ds, "z", ("gas", "density"))
p.set_cmap(field=("gas", "density"), cmap="viridis")
p.save()

which gives the following output:

I compared this with the following ParaView output:

which seems to agree well with each other.

The yt code output is the following

yt : [INFO     ] 2024-10-04 11:42:12,973 Parameters: current_time              = 0.01003431438520486
yt : [INFO     ] 2024-10-04 11:42:12,974 Parameters: domain_dimensions         = [1024  256    1]
yt : [INFO     ] 2024-10-04 11:42:12,974 Parameters: domain_left_edge          = [0. 0. 0.]
yt : [INFO     ] 2024-10-04 11:42:12,975 Parameters: domain_right_edge         = [4. 1. 1.]
yt : [INFO     ] 2024-10-04 11:42:12,975 Parameters: cosmological_simulation   = 0
yt : [INFO     ] 2024-10-04 11:42:13,137 xlim = 0.000000 4.000000
yt : [INFO     ] 2024-10-04 11:42:13,138 ylim = 0.000000 1.000000
yt : [INFO     ] 2024-10-04 11:42:13,141 xlim = 0.000000 4.000000
yt : [INFO     ] 2024-10-04 11:42:13,141 ylim = 0.000000 1.000000
yt : [INFO     ] 2024-10-04 11:42:13,143 Making a fixed resolution buffer of (('gas', 'density')) 800 by 800
yt : [INFO     ] 2024-10-04 11:42:13,586 Saving plot data.0001.flt.h5_Slice_z_density.png

[neutrinoceros]

This is awesome, thanks a lot for doing this !
I will make sure to review in detail soon, but just a couple very quick comments right off the bat:

In you test script, note that

import yt_idefix

can be omitted if you have yt>=4.2.0 (which I think yt_idefix requires anyway).
In the upcoming yt 4.4.0, you'll also be able to install as pip install 'yt[idefix]' so users don't even need to know that yt_idefix is a separate package.

Otherwise, regarding the PR itself, keep in mind we'll need to add at least one formal test for it, but as it'll require using git-lfs (which I'm allergic to), we can do that part later / separately.

[neutrinoceros]

quick update: I have a lot going on at the moment and I don't think I'll be able to prioritize this review before next week, but I'm not forgetting about it. Thanks for your patience !

[dutta-alankar]

No hurries at all.