GFZ / EL_PASO / 37337005847

73%

push

github

Random bug fixes (#185)

* fix: save both GFZ and netcdf output when save_strategy is 'both'

The hope electrons, ect combined, goes r mps high and goes realtime recipes
only assigned the strategy in each branch and called ep.save() once after,
so only the netcdf output was written. Call ep.save() inside each branch,
as the other recipes already do.

* fix: use dict.values() in Hashabledict.__hash__

itervalues() only exists in Python 2, so hashing a Hashabledict raised
AttributeError. This broke the cached construct_maginput() call when
indices_solar_wind was passed to compute_magnetic_field_variables().

* fix: correct posixtime to j2k conversion in j2k_posixtime_equiv

The backward leg called .total_seconds() on an ndarray of datetimes, which
raised AttributeError. Add posix_to_j2k and use it for the backward
conversion.

* fix: raise when pitch angles change in time in fold_pitch_angles_and_flux

* mean timebin should do nanmean instead of just mean

* min/max -> nan-min/-max)

* fix: avoid infinite loop in download with single_file cadence

* fix: guard empty InvK/InvMu in _load_variable for InvV

* fix: use UTC date in release mode timestamp

* fix: return Earth-fixed coordinates from calculate_geo_coords_from_tle

* fix: NaN-fill empty time bins in get_real_time_tipsod

* fix: keep sub-second precision in datenum conversions

* fix: treat naive datetimes as UTC in datetime_to_datenum

* fix: apply mag field name remap to all save strategies in arase mepe

* fix: correct alpha in density processing note and drop unused constants

* fix: remove duplicate B_Eq in poes meped variables

* fix: add float tolerance to model grid sanity checks

* mag_field conditional on gfz save_strategy

36 of 42 new or added lines in 16 files covered. (85.71%)

5270 of 7398 relevant lines covered (71.24%)

0.71 hits per line