RAAN from LTAN calculation off-by 180 degrees

[noc0lour]

The pull request which added RAAN from ltan function to twobody.angles module
shifts the angle by 180 degrees and it ends up having LTAN and LTDN switched around.

🐞 Problem

Verification with actual orbits obtained from TLEs shows this discrepancy

Example Sentinel-1:

In [1]: import astropy.time

In [2]: import astropy.units as u

In [3]: from poliastro import twobody

In [4]: s1_epoch = astropy.time.Time(astropy.time.Time('2020-01-01 00:00').to_value('mjd') + 45.05318108 - 1, format='mjd')

In [5]: twobody.angles.raan_from_ltan(s1_epoch, 18.00 * u.hourangle)
Out[5]: <Angle 593.23886941 deg>

In [6]: twobody.angles.raan_from_ltan(s1_epoch, 18.00 * u.hourangle).wrap_at(360 * u.deg)
Out[6]: <Angle 233.23886941 deg>

In [7]: twobody.angles.raan_from_ltan(s1_epoch, 18.00 * u.hourangle).wrap_at(360 * u.deg) - 180 * u.deg
Out[7]: <Angle 53.23886941 deg>

Value from Celestrak TLE for this: 53.8031

💡 Possible solutions

Removing -12 * u.hourangle from alpha

    alpha = (coordinates.Angle(ltan).wrap_at(24 * u.hourangle) - 12 * u.hourangle).to(
        u.rad
    )

I crosschecked with Sentinel-2, METOP-B and if I change this it checks out (+- 0.5 degree).
Calculation of the epoch coming from the TLE is a bit special since 2020-01-01 00:00 is expressed as 20001.0000 and thus I need to subtract 1.0 before calculating the appropriate MJD.

I've created this issue to discuss if this is actually wrong or if I'm just confusing something...

[astrojuanlu]

Ooops, good catch @noc0lour! Labeling this as a bug