Slicing or indexing Geocentric position+velocity array corrupts the velocity

[dmopalmer]

If you have an array of Geocentric position and velocity (e.g. from a satellite.at(timearray)), taking a slice or element out of it loses some of the type context and also converts the velocities to nan.

import skyfield.api
import os
import numpy as np

loader = skyfield.api.Loader(os.path.expanduser("~/.config/skyfield"))
ts = loader.timescale()

stations_url = 'http://celestrak.com/NORAD/elements/stations.txt'
satellites = loader.tle(stations_url)
satellite = satellites['ISS (ZARYA)']
# print(satellite)

now = ts.now()
nowandthen = ts.tai_jd(now.tai + np.arange(2)/86400)

atbyseconds = satellite.at(nowandthen)


print("Unsliced, sliced, and indexed:\n")
for posandvel in (atbyseconds, atbyseconds[:], atbyseconds[0]):
    print(type(posandvel), posandvel)
    print("Position")
    print(posandvel.position.km)
    print("Velocity")
    print(posandvel.velocity.km_per_s)
    print("-------------------------------------------")

Result:

Unsliced, sliced, and indexed:
<class 'skyfield.positionlib.Geocentric'> <Geocentric position and velocity at date t center=399 target=<object object at 0x10fe16ef0>>
Position
[[6161.12839736 6159.77530714]
 [2749.56861045 2754.20941124]
 [ 735.51733732  729.5672061 ]]
Velocity
[[-1.34916969 -1.3570354 ]
 [ 4.64268485  4.63917042]
 [-5.95007741 -5.95101607]]
-------------------------------------------
<class 'skyfield.positionlib.Geocentric'> <Geocentric position and velocity at date t>
Position
[[6161.12839736 6159.77530714]
 [2749.56861045 2754.20941124]
 [ 735.51733732  729.5672061 ]]
Velocity
[[nan nan]
 [nan nan]
 [nan nan]]
-------------------------------------------
<class 'skyfield.positionlib.Geocentric'> <Geocentric position and velocity at date t>
Position
[6161.12839736 2749.56861045  735.51733732]
Velocity
[nan nan nan]
-------------------------------------------