This was triggered by https://github.com/open-gamma-ray-astro/gamma-astro-data-formats/issues/118 - I've decided to keep the HDU index table as-is. It's not possible to make it compatible with the HDUCLAS header keys, without braking backward-compatibility completely or making the index table very complex. These index tables have been used and are in use since 3 or 4 years. They will likely be completly replaced by something better within the next year or two. I've added a section at the end of the page explaining the differences to HDUCLAS.
One open TODO in the spec was the use of BASE_DIR. I have clarified this in the spec here, and improved the implementation in Gammapy to make this work properly in https://github.com/gammapy/gammapy/pull/1585 .
This PR improves the HDU index table spec.
This was triggered by https://github.com/open-gamma-ray-astro/gamma-astro-data-formats/issues/118 - I've decided to keep the HDU index table as-is. It's not possible to make it compatible with the HDUCLAS header keys, without braking backward-compatibility completely or making the index table very complex. These index tables have been used and are in use since 3 or 4 years. They will likely be completly replaced by something better within the next year or two. I've added a section at the end of the page explaining the differences to HDUCLAS.
One open TODO in the spec was the use of BASE_DIR. I have clarified this in the spec here, and improved the implementation in Gammapy to make this work properly in https://github.com/gammapy/gammapy/pull/1585 .