Automatic formatting.

[albu-diku]

Integrate black into the Makefile replacing previous autopep8 attempt.

We are currently trying to have consistent formatting in the codebase but
here is currently no standardised way to do this. Based on some growing
consensus bring in black and switch local formatting and linting to it.

A new `make fmt` target will format files with the appropriate style. The
`make lint` target (broken/unused) invokes the same but in checking mode
and thus will report violations.

Given we are likely to adopt this in stages we define a variable in the
Makefile listing the directories we wish to enforce cleanliness for.

[jonasbardino]

Looks good so far.
I would really like to also have isort included as black screws with all our imports and adding it later will give another round of significant diffs.

While it's nice to see the results being sane I'm slightly uncertain if we actually want to merge with the fmt target applied in the same PR, and more so if isort is not included.
What do you think?

Just to clarify I'll gladly help adding isort to the mix unless you prefer to do that yourself.
https://github.com/ucphhpc/migrid-sync/tree/addition/automatic-formatting-plus-isort
has my simple attempt to do so.

[jonasbardino]

The lint errors are not new as seen in
#482
so they are not show stoppers.

[albu-diku]

Looks good so far. I would really like to also have isort included as black screws with all our imports and adding it later will give another round of significant diffs.

While it's nice to see the results being sane I'm slightly uncertain if we actually want to merge with the fmt target applied in the same PR, and more so if isort is not included. What do you think?

Just to clarify I'll gladly help adding isort to the mix unless you prefer to do that yourself. https://github.com/ucphhpc/migrid-sync/tree/addition/automatic-formatting-plus-isort has my simple attempt to do so.

Happy to add isort too and make it part of this - kept it to this initially only as a caution to do only what we want (although only in one portion of the tree so perhaps less an issue).

Yah, my sense is it comes down to what level of assurance we as part of this really. On the one hand, we’re using extremely robust tools for all this so there’s no particular concern on my part about how much we include. So perhaps it’s more, do we see any value in going in stages or would it be preferable to change logic files to how we want in one go.

The advantage we have with this branch is it’ll let us see the results of tweaks. Let’s try i sort and see where we land - and perhaps the import thing mentioned earlier is another candidate modulo consensus.

[jonasbardino]

Could you split this one up into just the mechanics and a follow-up with the actual reformat please?