Bug: apparent disagreement on number of HDUs
weaverba137 opened this issue · 3 comments
When verifying DESI_SPECTRO_REDUX, some files on disk appear to disagree with the data model on the number of HDUs. However when manually inspecting both the data file and the model file, the number of HDUs does agree. Track this down so that DESI_SPECTRO_REDUX can be fully verified.
Problem identified: some text in some data model files contains a certain phrase, e.g.:
See :doc:`coadd files <coadd-SURVEY-PROGRAM-PIXNUM>` for a coadded
However that is a signal to the data model software that the referenced document is the actual data model for the file, i.e. a cross reference. We use this to cross reference the data model for fiberassign and fibermap files, for example, but in the example above, when spectra-SURVEY-PROGRAM-PIXNUM.rst is read, this is interpreted as a command to use coadd-SURVEY-PROGRAM-PIXNUM.rst as the data model instead.
@sbailey, please take note.
This can be solved with a simple rephrasing of the text, which will appear in #169.
I've also found cases where a cross-reference is intended, but it did not follow the phrasing above.
Closed by #169.