[CK3] Title parsing

[robmart]

CK3 titles are structured like this

e_<empire_name> = {
	k_<kingdom_name> = {
		d_<duchy_name> = {
			c_<county_name> = {
				b_<barony_name> = {
				}
			}
		}
	}
}

However... Any title tier can be the root, and none of the of the tiers are required. The only requirement is that a title can only be the root or inside a higher tier title, which means that

k_<kingdom_name> = {
	d_<duchy_name> = {
		b_<barony_name> = {
		}
	}
}

e_<empire_name> = {
	d_<duchy_name> = {
		b_<barony_name> = {
		}
	}
}

are a valid definitions, though

e_<empire_name> = {
	e_<empire_name> = {
		d_<duchy_name> = {
			c_<county_name> = {
				b_<barony_name> = {
				}
			}
		}
	}
}

is invalid.

As far as I know there is no way to accurately represent this in configs.

[DragonKnightOfBreeze]

For config matching e_<empire_name> - you can just use e_<empire_name>, if empire_nameis a valid definition type (like event), this should works
For the nested structure - you can just combine with alias / single alias, this should be not hard to support.

[robmart]

Ah sorry the <empire_name> isn't it's own type - I should have used a different symbol. That's just the stand-in I used for having any string there.

The main problem is that each nested title has to be its own type. So,

e_{empire_name} = { <-- This needs to be its own type
	k_{kingdom_name} = { <-- As does this
		d_{duchy_name} = { <-- And these
			c_{county_name} = { <-- And these
				b_{barony_name} = { <-- And these
				}
			}
		}
	}
}

I could get just the validation done by doing it with subtypes and single alias, this is true, however that wouldn't make any of the "sub-titles" types of their own, which they need to be so they can be references elsewhere.

[DragonKnightOfBreeze]

So the empire_name and others must be definitions?
Or be dynamic values (like event targets) is ok?
Or be complex enum values is ok?

Are there any actual examples? Plz paste them.

[robmart]

Here's the definition for britannia from vanilla https://gist.github.com/robmart/542e3aa40af4fbf860c30a63de5b9e04

[DragonKnightOfBreeze]

Seems like there is already some special options in vanilla cwt type configs (## start_with or ## starts_with?), and I misunderstood its purpose.
Nested declarations should be already supported.
For this situation, seems like while the root key is e_xxx, the definition name should be xxx.

[DragonKnightOfBreeze]

Will check them in dev version. Since I do not own ck3, more info may be required ...

[robmart]

It seems like it was hardcoded for CK2 in CWT https://github.com/cwtools/cwtools/blob/60f085aea0ed64f3304b98f1b03cf03543196073/CWTools/Game/CK2/CK2Game.fs#L84

[DragonKnightOfBreeze]

.....

[DragonKnightOfBreeze]

I dislike the idea that use hardcorded-code in cwt config repos...

[kaiser-chris]

I dislike the idea that use hardcorded-code in cwt config repos...

Yeah please do not hardcode this part. shudder

[robmart]

I agree that it's a bad solution

[robmart]

An alternate solution could maybe be instead of a nested type, a recursive type? Not sure if it's a good idea but something like:

## recursive = yes
type[landed_title] = {
    path = "common/landed_titles"
    ## starts_with = e_
    ## recursive_starts_with = { k_ d_ c_ b_ }
    subtype[empire] = {
    }
    ## starts_with = k_
    ## recursive_starts_with = { d_ c_ b_ }
    subtype[kingdom] = {
    }
    ## starts_with = d_
    ## recursive_starts_with = { c_ b_ }
    subtype[duchy] = {
    }
    ## starts_with = c_
    ## recursive_starts_with = { b_ }
    subtype[county] = {
    }
    ## starts_with = b_
    subtype[barony] = {
    }
}

So recursive would indicate that the type could be inside itself, and recursive_starts_with (which could be on either the type or the subtypes) would indicate that only the elements that start with that prefix should be treated as the type.