Python: bidimensional dictionary?

[Oberlus]

I don't know python, just the obvious stuff for a C programmer.
I'd like to access a nested dictionary, both levels indexed by fo.galaxySetupOptionGeneric (low, medium, high). Like this:

Code: Select all

    age_to_density_factor = {
        #                                            Low density                            Medium density                            High density
        #                Young galaxy:
        fo.galaxySetupOptionGeneric.low:    {fo.galaxySetupOptionGeneric.low: 0.54, fo.galaxySetupOptionGeneric.medium: 0.79, fo.galaxySetupOptionGeneric.high: 1.32},
        #                Mature galaxy:
        fo.galaxySetupOptionGeneric.medium: {fo.galaxySetupOptionGeneric.low: 0.78, fo.galaxySetupOptionGeneric.medium: 1.17, fo.galaxySetupOptionGeneric.high: 1.95},
        #                Ancient galaxy:
        fo.galaxySetupOptionGeneric.high:   {fo.galaxySetupOptionGeneric.low: 0.92, fo.galaxySetupOptionGeneric.medium: 1.46, fo.galaxySetupOptionGeneric.high: 2.42},
    }
    num_planets_in_galaxy = galaxy_size * age_to_density_factor[galaxy_age][planet_density]

The syntax is wrong, and gives varied KeyError, depending on what I try. I also tried using parenthesis in the inner level, which allowed me to hang the server in some kind of infinite loop flooding the command line with what seemed like freeoriond.log contents.

Maybe I should be doing something totally different, like using tuples of galaxySetupOptionGeneric to index a simple dictionary?

Code: Select all

    age_to_density_factor = {
        # Young Low:
        fo.galaxySetupOptionGeneric.low, fo.galaxySetupOptionGeneric.low: 0.54,
        fo.galaxySetupOptionGeneric.low, fo.galaxySetupOptionGeneric.medium: 0.79, 
        fo.galaxySetupOptionGeneric.low, fo.galaxySetupOptionGeneric.high: 1.32,
        # Mature:
        fo.galaxySetupOptionGeneric.medium, fo.galaxySetupOptionGeneric.low: 0.78,
        fo.galaxySetupOptionGeneric.medium, fo.galaxySetupOptionGeneric.medium: 1.17,
        fo.galaxySetupOptionGeneric.medium, fo.galaxySetupOptionGeneric.high: 1.95,
        # Ancient:
        fo.galaxySetupOptionGeneric.high, fo.galaxySetupOptionGeneric.low: 0.92,
        fo.galaxySetupOptionGeneric.high, fo.galaxySetupOptionGeneric.medium: 1.46,
        fo.galaxySetupOptionGeneric.high, fo.galaxySetupOptionGeneric.high: 2.42,
    }
    num_planets_in_galaxy = galaxy_size * age_to_density_factor[galaxy_age, planet_density]

This again allowed me to screw the server and forced me to manually kill zombie processes.

It seems I'm not making the right questions to Google when searching for an answer.

Can anyone help me? Vezzra or o01eg maybe?


I'm tempted to use an ugly if-then block of code...

[Grummel7]

You could use using a single dictionary with tuple-keys:

Code: Select all

    age_density_factor = {
        (fo.galaxySetupOptionGeneric.low, fo.galaxySetupOptionGeneric.low): 0.54,
        (fo.galaxySetupOptionGeneric.low, fo.galaxySetupOptionGeneric.medium): 0.79, 
        ...
    }
    num_planets_in_galaxy = galaxy_size * age_density_factor[(galaxy_age, planet_density)]

Your first implementation looks correct, though.
If you are getting key errors, I guess it means galaxy_age or planet_density have other values then fo.galaxySetupOptionGeneric.low/medium/high

[Oberlus]

Thank you, Grummel7, this worked perfectly.
Now I have to do something (else) with the FOCS part...

Edit: o01eg, once the species focs files are translated to python, would that make it possible to declare data like that? Or will I be forced anyways to use if-else and the such?

[o01eg]

Thank you, Grummel7, this worked perfectly.
Now I have to do something (else) with the FOCS part...

Edit: o01eg, once the species focs files are translated to python, would that make it possible to declare data like that? Or will I be forced anyways to use if-else and the such?

With Python FOCS it will be available not only if-else control flow but all other.

[Oberlus]

Awesome!

Then I'll wait for python species files (or I could try to help you with that, if you have any menial tasks to share with me) before continuing with the galaxy-size-bound upkeep formula.
Also, more time to explore alternatives that Geoff could approve.

[o01eg]

Awesome!

Then I'll wait for python species files (or I could try to help you with that, if you have any menial tasks to share with me) before continuing with the galaxy-size-bound upkeep formula.
Also, more time to explore alternatives that Geoff could approve.

I need re-create species with 98 effect groups for equality test and it's a quite big task before changing parser.

[Ophiuchus]

I'll wait for python species files ... before continuing with the galaxy-size-bound upkeep formula.

out of curiosity - how is that connected?

[Oberlus]

The upkeep is calculated in species/common/influence.macros.

[Ophiuchus]

With Python FOCS it will be available not only if-else control flow but all other.

maybe I am confused...

what i understand the python code will be evaluated at e.g. game start so we will be able to use python control flow to do focs metaprogramming.

But the influence upkeep oberlus talks about needs to be calculated at runtime, so the formula can not actually use python control flow, is that correct?

[Oberlus]

I think it's about the parser.

FOCS parser, being retired, can't do fancy stuff with tables, python FOCS parser will be able.

[Ophiuchus]

I think it's about the parser.

FOCS parser, being retired, can't do fancy stuff with tables, python FOCS parser will be able.

yes, but not while the game is running (because at runtime only the FOCS c++ data structures exist and if there is no support for fancy tables there.. you wont be able to use it). And invoking a python interpreter per planet just to calculate influence cost would sloooooooow everything, so that is probably not gonna happen.

wait for o01eg to clarify

[o01eg]

I think it's about the parser.

FOCS parser, being retired, can't do fancy stuff with tables, python FOCS parser will be able.

yes, but not while the game is running (because at runtime only the FOCS c++ data structures exist and if there is no support for fancy tables there.. you wont be able to use it). And invoking a python interpreter per planet just to calculate influence cost would sloooooooow everything, so that is probably not gonna happen.

wait for o01eg to clarify

Yes, both FOCS parsers run once at application starts, not each time calculations happened.