Repository navigation
Inconsistent angle string unit in coordinates/angles gives incorrect errors / values #1387
Description
In the 0.2.4 docs and code, this is seen:
>>> ICRSCoordinates('3d4m5s +6h7m8s', unit=(u.hour, u.degree))
ValueError: parse_hours: Invalid input string, can't parse to HMS. (3d4m5s)
In master one gets:
>>> ICRSCoordinates('3d4m5s +6h7m8s', unit=(u.hour, u.degree))
BoundsError: The angle(s) [ 5156.62015618 6016.05684887 6875.49354157] falls outside of the specified bounds (-5156.62015618, 5156.62015618) (in deg)
>>> ICRSCoordinates('0d0m0s +0h0m0s', unit=(u.hour, u.degree))
ValueError: Unknown format code 'f' for object of type 'str'
This may be related to this problem in Angle where giving an HMS string with unit=u.degree results in a length=3 array:
Angle('+6h7m8s', unit=u.degree)
<Angle [u'90d00m00.00000s' u'105d00m00.00000s' u'120d00m00.00000s']>
Activity
Remove non-crucial example that fails now in master (see astropy#1387)
mdboom commentedon Aug 30, 2013
That last example is definitely wrong, and I thought it was covered by the tests. Let me look into that.
mdboom commentedon Sep 3, 2013
With #1388:
I think that's a legit error.
And
taldcroft commentedon Sep 3, 2013
Agreed that 6h7m is outside of 90 degrees, but what the heck is with the angle of 5258.79 and bounds of +/-5156.62? I think that #1388 is good for merge, but that this issue #1387 should stay open until we resolve the above.
mdboom commentedon Sep 3, 2013
Oh, gee. Of course. Something is wrong with the unit of the error message, I suspect.
Partial fix for astropy#1387: Angles are not properly parsed when the…
mdboom commentedon Sep 3, 2013
The bug in the error message is actually gone in master, since #1310 was merged (and the error message happens in a completely different way). Rebasing #1387 fixed it.
taldcroft commentedon Sep 3, 2013
@mdboom - Excellent. So it looks like you are good to merge #1388 and close #1387, at least from my perspective.