plotting ra,dec from eFEDS FITS file (without units?)

[dnelson86]

I copied the example code for the SDSS FITS file, for a different file:

import matplotlib.pyplot as plt
import numpy as np
import scida

path = '/virgotng/mpia/obs/eROSITA/eFEDS/'
filename = 'fm00_merged_020_EventList_c001.fits'

ds = scida.load(path + filename)

fig = plt.figure(figsize=(10,5))
ax = fig.add_subplot(111, projection="aitoff")
sc = ax.scatter(ds['RA'], ds['DEC'], s=0.05, c=ds['PI'], rasterized=True)
fig.colorbar(sc, label="Photon Energy [eV]")
ax.set_xticklabels(['14h','16h','18h','20h','22h','0h','2h','4h','6h','8h','10h'])
ax.set_xlabel("RA")
ax.set_ylabel("DEC")
ax.grid(True)
plt.savefig("eFEDS.png", dpi=150)

which produces the following image:

but the distribution of RA,DEC in this file (in theory) are localized to a small sky region like this:

Is this possibly because the fields in this FITS file have no units?

WARNING:scida.io._base:FITS file support is work-in-progress.
WARNING:scida.interfaces.mixins.units:Cannot determine units from neither unit file nor metadata for '/DEC'.
WARNING:scida.interfaces.mixins.units:Cannot determine units from neither unit file nor metadata for '/PI'.
WARNING:scida.interfaces.mixins.units:Cannot determine units from neither unit file nor metadata for '/RA'.

or possibly something else?

[dnelson86]

Documentation for the structure of such files:

https://erosita.mpe.mpg.de/dr1/eSASS4DR1/eSASS4DR1_ProductsDescription/extensions_description_dr1.html

[dnelson86]

Same result with:

sc = ax.scatter(ds['RA']*ds.ureg['deg'], ds['DEC']*ds.ureg['deg'], s=0.05, c=ds['PI'], rasterized=True)

WARNING:pint.util:Calling the getitem method from a UnitRegistry is deprecated. use `parse_expression` method or use the registry as a callable.
WARNING:pint.util:Calling the getitem method from a UnitRegistry is deprecated. use `parse_expression` method or use the registry as a callable.
/u/dnelson/.local/envs/py3/lib/python3.9/site-packages/numpy/ma/core.py:2820: UnitStrippedWarning: The unit of the quantity is stripped when downcasting to ndarray.
  _data = np.array(data, dtype=dtype, copy=copy,
/u/dnelson/.local/envs/py3/lib/python3.9/site-packages/matplotlib/axes/_axes.py:4353: UnitStrippedWarning: The unit of the quantity is stripped when downcasting to ndarray.
  c = np.asanyarray(c, dtype=float)

[cbyrohl]

One can test this without units:

ds_nounits = scida.load(path + filename, units=False)
ds = scida.load(path + filename)

which gives the same result except for carrying "unknown units" as a pint.Quantity for

assert np.all(ds.data["RA"].magnitude.compute() == ds.data["RA"].compute())

I would expect this problem to be unrelated to the unit attachment?

[dnelson86]

In [5]: RA.min()
Out[5]: 126.20363814596136

In [6]: RA.max()
Out[6]: 145.82010938548257

In [7]: DEC.min()
Out[7]: -3.223283315519501

In [8]: DEC.max()
Out[8]: 6.099576083087285

[dnelson86]

The result is the same with ds = scida.load(path + filename, units=False).

[cbyrohl]

Isn't the computed data range correct?

[dnelson86]

Yes, just the visual plot is wrong. Turning off the aitoff projection with ax = fig.add_subplot(111) fixes everything, and produces this image (7.6M photons in eFEDS):

Is this just some matplotlib bug, or some strange interaction between scida and matplotlib?

[cbyrohl]

It is the same when you pass these arrays cast to numpy, so I do not think this has anything to do with scida.

Regarding units for eFEDS: If there is no metadata, we do not make an assumption for the units here. We can provide a minimal unit file.

[dnelson86]

With correct x-axis, and sorting by photon energy:

[dnelson86]

All 172M photons of the full eRASS1, unfortunately I think (vs Merloni+24 Fig 6) it suffers from a similar bug, i.e. the data should not fill the sky:

[cbyrohl]

Did you check the matplotlib docs? I could imagine it uses radians rather than degrees.