This code used in test_ephem.py and test_utils.py is broken under astropy 4.2 because the last column is no longer parsed as name "col10":
# Configure a CSV reader for the Horizons output format.
csv_reader = astropy.io.ascii.Csv()
csv_reader.header.comment = r'[^ ]'
csv_reader.data.start_line = 35
csv_reader.data.end_line = 203
# Read moon ephemerides for the first week of 2020.
path = astropy.utils.data._find_pkg_data_path(
os.path.join('data', 'horizons_2020_week1_moon.csv'),
package='desisurvey')
cls.table = csv_reader.read(path)
# Horizons CSV file has a trailing comma on each line.
cls.table.remove_column('col10')
I suggest moving this horizons_2020_week1_moon.csv reader into a single place that both test_ephem.py and test_utils.py can use, ensuring that it works with both astropy 4.0.x and 4.2.x, and switch from using astropy.utils.data._find_pkg_data_path to pkg_resources.resource_filename to avoid using a _underscore_hidden function from astropy that may break again in a future release.
This code used in test_ephem.py and test_utils.py is broken under astropy 4.2 because the last column is no longer parsed as name "col10":
I suggest moving this
horizons_2020_week1_moon.csv
reader into a single place that both test_ephem.py and test_utils.py can use, ensuring that it works with both astropy 4.0.x and 4.2.x, and switch from usingastropy.utils.data._find_pkg_data_path
topkg_resources.resource_filename
to avoid using a_underscore_hidden
function from astropy that may break again in a future release.