Stable version of EDPS

[chyan26]

I found the EDPS 1.5.0 is different from 1.4.6. While I say -m science, 1.5.0 only process the meta-target science. However, 1.4.6 process all related tasked. I would suggest we use 1.4.6 for now. As on this page, https://ftp.eso.org/pub/dfs/pipelines/repositories/stable/src/edps/. The stable version should be 1.4.6.

[hugobuddel]

Can you elaborate? EDPS 1.5.0 seems to be working as expected. E.g. if I look at the recent github actions log, everything is processed to create the requested targets.

You replaced the -m science with -m all in bae8021, so even if there was a problem, it seems the problem is now gone, is that correct?

Or is there still a reason to use 1.4.6?

[chyan26]

Yes, correct. I replace -m science to -m all. However, the -m science will process all the tasks in 1.4.6 because there is no any pre-existing products. Therefore, -m science will run 12 tasks in 1.4.6 and run 9 tasks marked with science meta-target in 1.5.0.

[hugobuddel]

So shall we close this? I'm verifying we are indeed in agreement; I think so, but I'm not entirely sure.

Maybe the default changed in the edps from processing all to processing the science meta target.

1.5.0 should still process any intermediate data (e.g. tagged with calib) that would be required to create the products tagged with science right? For example, the dark_task is executed, which is not tagged science. So we can remove the science tag from background_std_task and standard_flux_task too.

If 1.4.6 also ran 3 other recipes that would not be required for the science targets, then that would have been a bug that is apparently fixed.