Next try with WMO sensor names

[sfinkens]

Here's a new try, where the sensor attribute holds a set of WMO names. The
attribute itself is not modified. I added normalization and serialization methods
to make use of this attribute in filenames.

We can't use it like this, because changing the type of the sensor attribute
would breaks users' workflows. The plan is to add a new instruments
attribute that can be used instead. But for now I just used sensor
to keep the number of changes small and see if that's the direction we
want to go.

As I was going through the code base I found it easier working with the set
most of the time. For YAML filenames (which only contain a single sensor)
I added a function to normalize the sensor name. Only in the Writer
class I needed a serialization function for multiple sensors.

• Closes #xxxx
• Tests added
• Fully documented
• Add your name to AUTHORS.md if not there already

[sfinkens]

Grrr, I amended to an older commit, will fix this tomorrow, gotta go now

[djhoese]

You might need to remake this PR. It has the NDVI Hybrid green change that I made and my commit message, but includes your other changes. Very confusing. Not sure how this happened.

Otherwise, I think the biggest/easiest change I'd like on this is to rename all the helper functions and variables to be "instrument" or "inst" (for the shorter variable names). That's what I took from the slack conversations is that we should do instruments everywhere and get away from sensor.

[sfinkens]

Fixed the amended commit.

rename all the helper functions and variables to be "instrument"

Agreed! I didn't do that yet to make it easier to review. That's the next step once we're sure that this is the direction we want to go

[sfinkens]

Superseded by #3390