Open cdeil opened 9 years ago
I think it might also make sense to include the erfa test from there. While in principal astropy is using erfa to do this, it does so via some subtly different paths, and at the very least we want to know how much floating point rounding kicks in there (I think that's the main effect right now, because it's at the micro-arcsec level)
As mentioned in https://github.com/astropy/astropy/pull/3296#issuecomment-71523968, @eteq has written coordinate accuracy tests here: https://github.com/eteq/astropy_astrometry_tests
We should integrate what's still missing in this
coordinates-benchmark
, e.g. the IDL (see #36) and TPM tests.