Handle nested dictionaries in config parameter storage

[Ian0sborne]

This PR fixes #941 whereby some arguments in write_city_configuration were over the 300 character threshold. This issue occurred when passing dictionary arguments without unpacking them. This PR makes the write_city_configuration function sensitive to dictionary arguments and unpacks them, assigning sensible names. I have also added a case for this in the test function.

[gonzaponte]

This looks pretty good.

I would it to be a bit more general, allowing an arbitrary level of nesting. I would go for a recursive approach.

[Ian0sborne]

I have tried to come up with an iterative approach by adding a new function. Not sure if this is ugly though?

def flatten_dict(arg : dict,
                parent_key : str= "",
                dev : str= "."):
    flat_dict = {}
    for k, v in arg.items():
        new_key = f"{parent_key}{dev}{k}" if parent_key else k
        if isinstance(v, dict):
            flat_dict.update(flatten_dict(v, new_key))
        else:
            flat_dict[new_key] = str(v)
    return flat_dict

This would output something like:

Let me know what you think and I will implement it

[gonzaponte]

Apologies for the late reply. I've had a hectic couple of weeks.

The proposed solution is excellent, I think. Just a minor comment: you either remove the dev argument (what does it stand for?) or pass it as an argument when you call the function from within the loop. I don't think we want to make it this generic, but both approaches are fine by me.

[Ian0sborne]

No need to apologise!

You're absolutely right Gonzalo: the "dev" was meant to stand for "divider" (but I made a typo). There's no need for it to be an argument so I'll implement that and the rest of the changes now.

[gonzaponte]

Sorry, I thought I have already commented here. This is almost ready, let's rewrite the commit history following these recommendations. After that, I'll approve it.

[Ian0sborne]

Hi Gonzalo! Is the commit history okay now?

[gonzaponte]

The second commit says "Alter function & test...", but the changes are more like "Introduce function and alter test ...". I suggest you put the second version of the test (the one prior to "simplification") already in the first commit and then introduce the final version of the function in the second commit, leaving the third one as it is now. Let me know if you need help with this.

[Ian0sborne]

I'll try this now, will let you know if I need help. Thanks!

[Ian0sborne]

I think this should be better now.

I was not able to put the second version of the testing function in the first commit because I committed it at the same time as the "final version" of the unpacking function. For this reason, I thought they would be better in the second commit together.

Let me know what you think of these changes

[gonzaponte]

The history has been rewritten with peras to learn some magit.

[gonzaponte]

Handles nested dictionaries elegantly in the config output. And it's tested. Nice job!

[gonzaponte]

@Luismiguelvillar please merge!