After a private discussion, we decided to get rid of the custom Orekit-based API and make use of its pure Python wrapper. This new implementations includes the following features:
The orekit/src/ has been completely removed as it is no longer required.
Test files and functions are renamed validate_*.py. This outputs a cool namespace in the CI actions, from my point of view.
Regarding the orekit-data.zip, it is ignore from the VCS no matter its location and will be downloaded if not found. This behaviour is coded under the tests/conftest.py.
The Orekit's virtual machine needs to be executed in each one of the scripts. Otherwise, I have experienced segmentation-fault errors, strange non-traceable issues and non-sense numerical values.
After a private discussion, we decided to get rid of the custom Orekit-based API and make use of its pure Python wrapper. This new implementations includes the following features:
orekit/src/
has been completely removed as it is no longer required.validate_*.py
. This outputs a cool namespace in the CI actions, from my point of view.orekit-data.zip
, it is ignore from the VCS no matter its location and will be downloaded if not found. This behaviour is coded under thetests/conftest.py
.(Edit: Supersedes #15)