I'm pretty uncomfortable with the changes to pipe_tasks. Passing around mappers and accessing the internal standardization methods seems dangerous. I understand the desire not to duplicate code, so maybe that's why you didn't use a translate function. That can be avoided by putting the code in a free function and calling it in both translate_taiObs and std_raw_md. That would mean you do not need the changes in pipe_tasks.

Another general comment is that the coding standard asks for docstrings for every method and class. There are many methods in this changes et that have no docstring at all and many others where the docstring does not adhere to our standard (i.e. naming all input and output variables).

I'm not quite done with the review, but I'll post this now so you can look at them.

I've completed my review. I'd like to come up with a solution that does not include the pipe_tasks changes you have on this ticket. I would also like to make sure docstrings are up to standard. The rest of my suggestions are relatively minor.

Simon Krughoff I believe I have dealt with all your comments now. Could you take a look and confirm please?

Looks great. The only minor thing is that I don't see how you've dealt with the fact that the same code exists in the notebook as exists in the pre-burner script. It's not a big deal to me, but thought I'd mention it. Feel free to merge.

Edit: I was confused. The duplication was resolved. I was looking at another script that is not a duplication of the notebook.

Merged.