Write proposal and run numbers to virtual CXI files.

[turkot]

Although proposal and run numbers are not part of the CXI format, I find them useful additional information.

[daniloefl]

Looks great! Feel free ignore all my comments. They are mostly only ideas for alternative implementations, in case you think it is worth going in those directions.

[daniloefl]

I can see you are just following a previous standard, but is there a reason for entry_1?

[turkot]

According to CXI format convention, entry_1 should correspond to the experiment.

[turkot]

Thank you for your comments, Danilo! The are all very good. We should try to follow CXI format guide:
https://raw.githubusercontent.com/cxidb/CXI/master/cxi_file_format.pdf
But since these kind of data is outside of their specification, it is up to us where (and how) to store it.
I like your idea to put proposal and run numbers as attributes, with some default values (e.g. -1?) in case it is unknown. But I'm not yet sure where it belongs - root, /entry_1 or /entry_1/instrument_1...

[daniloefl]

I also do not have an opinion here. Your usage of entry_1 is fine. I was just asking in case there was a specific reason for that. If the CXI data format allows, I am biased towards using attrs instead, but I am not sure this is a viable option.

[philsmt]

You may want to consider whether attributes are already a concept for metadata in CXI or not, and whether you want this EuXFEL-specific information to be visible or rather hidden (note I do not have answers or opinions on this out of ignorance).

[turkot]

I'm adding name of the detector as 'detector_1' attribute.

[takluyver]

This broadly LGTM.

One more naming bikeshed question, though: do we want to make it explicit that the proposal number is an EuXFEL proposal number? EuXFEL proposal number 12345 is unique, and you have an immediate hint where to find more context.