Named for the artificial minds which control stompy awesome steam powered robots in Privateer Press's game Warmachine, Cortex is a tool for rapidly computing attack values and statistics.

I built Cortex as a learning aid, so that as I play Warmachine I can rapidly evaluate the probable outcomes of an absurd charge compared to a more conservative sniper stance in terms of damage dealt, and this purpuse Cortex serves rather well.

Cortex does not at present provide any assistance in movement planning or board position analysis. These tasks are both complex and require that Cortex somehow achieve a total knowlege stance over the game board which nobody can or should consider fair given the "fog of war" intent of Warmachine's distance measuring rules.

Simply download or clone this repository and execute the file "src/cortex.py", and you drop into the Cortex read/eval/print loop.

Implemented Commands:
	 "help"
		 prints a list of all installe commands

     "help <command>"
		 prints usage information for the mentioned command
		 
     "load models <filename>"
		 loads a model descriptor file, being a list of Model compatible
		 JSON objects and enters them into the interpreter's model table.
    
    "alias model <alias> <model name>"
		 creates an alias for a model, hopefully something easier to type
    
    "ls models"
        enumerates the loaded models by name
    
    "ls aliases"
        enumerates the loaded aliases by abbrev -> real name

	"stats <model>"
		prints model statistics as previously loaded from a datafile

	"attack <model> [with ([def|arm|pow|mat|rat|str] [-|+|=]<number>)+ end]
            <model> [with ([def|arm|pow|mat|rat|str] [-|+|=]<number>)+ end]"
        enumerates and evaluates all attack options for one model
        against another, evaluating every weapon the attacking model
        has against the target at range and in charge. Attacking
        warcasters and warjacks will compute the probable outcomes of
        different boosting strategies as well. This command has a
        boatload of options because it provides support for inline
        stat modifiers.
		
    "reset"
        clears out all the state of the repl, essentially restarting
        it without quitting and reloading the program.
    
    "quit" | "exit" | ":q"
        exits the Cortex program.

Cortex is designed to store models in simple JSON, with a data file being a list containing well formed objects representing models. This is an example of a well formed model object for a fictional model.

{"name":"Walking Dishwasher",
 "type":"warjack",
 "size":"large",
 "attrs":{"modules":["l","r","m","c"],
	      "pc":9,
		  "armor":20,
		  "defence":10,
		  "rat":6,
		  "mat":7,
		  "strength":12,
		  "faction":"arrdem",
		  "arcs":["front"],
		  "c":{"front":[[3,0],[3,1],[4,0]]},
		  "m":{"front":[[1,0],[2,0],[2,1]]},
		  "r":{"front":[[5,0],[4,1],[4,2]]},
		  "l":{"front":[[0,0],[1,1],[1,2]]},
		  "boxes":{"front":[4,5,6,6,5,4]},
		  "speed":5,
		  "reach":true},
 "weapons":[{"name":"Flung Brick",
	         "type":"ranged","attrs":{},
			 "rng":16,"rof":1,"pow":15},
			{"name":"Heavy Noodle",
			 "type":"melee","attrs":{"reach":true},
			 "rng":2.0,"pow":5}]}

Many of these datapoints, such as the arcs and boxes details, are currently ignored. However should cortex ever grow beyond its current scope and attempt to plan actions they would become relivant and are hence specified.

I cannot and do not and will not provide the datafiles describing the models, as they are the sole IP of Privateer Press. These legalities I am very much aware of, and I have taken pains to purge the data files which I develop and test against from this project's distributable repository.

However, I do provide a well documented format which such data could take, and may extend cortex to include a tool to aid in the compilation of such data.

• Model position/movement AI
• Target selection AI
• Vision based board analysis
• Trivial list management/point counting

This code is made available under the terms of the EPL 1.0, and is held in sole copyright by Reid "arrdem" McKenzie