Feature/smhi class to match 0.2.0

[docNord]

Naive update of parent class to new structure. To be updated later. Ambition here is simply to support 0.2.0 release.

• Restructure base code to new format
• Fix superstructure (find station from gps and place name, interpolation etc)
• Test functionality
• Re-write tests
• Update documentation

[mgcth]

A simplification.

[docNord]

I am not in a position to fix the unit tests without support.

[mgcth]

I hate mutating an object's state. I'd rather a method called get_ returns something and not also mutated an object's internal state.

I'd also write the unit tests in a way that will be easier for us when updating the code in the future. That is, as little patching as possible and assert calls and stuff. The important underlying stuff has been tested and we can assume it'll run fine. I know this isn't strictly a unit test, but they are too brittle to upkeep for our small organisation. Define some input data at some defined interface and some output data your module should produce, mock requests.get, the Metobs modules or so, and let the code run. Call that the unit test of the module. In the integration test, we can actually call the API and run through.

I've simplified it a bit in 57c3e05. I'd probably add a separate method called get_data_by_city (and also allow for interpolation) or so and refactor some of the code in get_data so it can be reused.

[mgcth]

I still think you should refactor the _interpolate method in c4dfcca.

[docNord]

Sorry @mgcth, didn't realize until now that you had also commited here. I will look into your work and update. Agree we shouldn't mutate the object.

[mgcth]

Great job, looks better!

At this point, I'd focus on docs and one or two tests of the whole SMHI object (call them integration, but without calling the API).

[docNord]

Thanks for the input. I will try to get an integration test going and sort out the remaining stuff to the extent that I can.

[docNord]

@mgcth it seems your fix to use IntEnum changed the order of the list so that default is now "corrected-archive" again, was that deliberate?

[mgcth]

@mgcth it seems your fix to use IntEnum changed the order of the list so that default is now "corrected-archive" again, was that deliberate?

Yes. Do you prefer it the other way around?

[docNord]

@mgcth it seems your fix to use IntEnum changed the order of the list so that default is now "corrected-archive" again, was that deliberate?

@mgcth it seems your fix to use IntEnum changed the order of the list so that default is now "corrected-archive" again, was that deliberate?

Yes. Do you prefer it the other way around?

Not at all - I thought you wanted it the other way around after that discussion we had earlier, but I have absolutely no opinion on the matter. I will continue with this way of working, seems good to me. 👍

[docNord]

Ping @mgcth

[mgcth]

Looks good. Just minor comments from the phone. Fix them and I'll approve tomorrow when I have a computer at hand.

[mgcth]

Great job! 🚀

[docNord]

I don't know if I managed to get the figure in the build properly @mgcth, could you have a look please? Other than that, I am quite happy.