Function: sort_text

--Usage--
> sort_text(order: order specification, some_text)
> sort_rule(order: order specification)(some_text)

Sort a string or filter it by keeping specific characters.

When the @order@ parameter is not given, the input is sorted in alphabetical order.

Otherwise the @order@ parameter determines what characters to show in the output.
The @order@ is a string consisting of multiple ''parts'', each part selects a specific thing from the input.
Parts
! Part			<tt>sort("zzyyxxy")</tt>	Description
| @"x"@			@"xx"@			A normal character, selects all matching characters from the input
| @" "@			@""@			Whitespace is ignored.
| @"\\<"@		@""@			Characters can be escaped with a backslash, normally the character @"<"@ has special behaviour.
			 			Note that you also need to escape the backslash for the [[type:string]].
| @"<wxy>"@		@"xy"@			Selects the characters from the input, but only once.
			 			@"<xy>"@ is the same as @"<x><y>"@.
| @"[wxy]"@		@"yyyxx"@		Selects the given characters, but keep them in the same order as in the input.
| @"ordered(wxy)"@	@"xxyyy"@		The same as just @"wxy"@.
| @"once(wxy)"@		@"xy"@			The same as @"<wxy>"@.
| @"mixed(wxy)"@		@"yyyxx"@		The same as @"[wxy]"@.
| @"compound(yx)"@	@"yx"@			Selects the literal substring @"xy"@ only where it matches exactly. Selects it as many times as it appears
| @"any()"@		@"zzyyxxy"@		Selects the remaining part of the input.
| @"cycle(xwz)"@	@"zzxx"@		Selects the given characters, and put them in the order that minimizes the distance.
			 			around a 'circle'. Imagine w x and y in a circle, then "zx" is shorter then "x.z" because the latter passes the "w" space.
			 			This sorting option is used to sort colored mana symbols in Magic.
| @"reverse_order(x y)"@	@"yyyxx"@	Inside the parentheses can be multiple parts (that can again include part-functions).
			 			@"reverse_order(x y)"@ is ''not'' the same as just @"y x"@, because in the first case x is evaluated first.
			 			For example @"x compound(xy)"@ doesn't work, because the xs are all matched by the time the compound part is evaluated.
			 			So instead you can use @"reverse_order(compound(xy) x)"@.
| @"pattern(.z. xyz)"@	@"yzz"@			Selects all things that match the pattern, where @"."@ is a wildcard.
			 			The things matching the wildcards are then sorted using the given pattern (separated by a space), and subsituted back in.
			 			So in <tt>"zzyyxxy"</tt> the pattern matches <tt>"<b>zzy</b>xxy"</tt> with wildcards <tt>"zy"</tt> these sort as <tt>"yz"</tt> and in the pattern this becomes <tt>"yzz"</tt>.

The parts are read from left to right, each part that matches something removes it from the input, so for example
> sort_text(order: "xx")
behaves the same as
> sort_text(order: "x")
because the first @"x"@ already selects all xs from the input.

This function is available in [[script:rule form]].

--Parameters--
! Parameter	Type				Description
| @input@	[[type:string]]			Text to sort
| @order@	[[type:string]] (optional)	Order specification

--Examples--
> sort_text("banana")                                      ==  "aaabnn"
> sort_text(order: "na",                        "banana")  ==  "nnaaa"
> sort_text(order: "n a",                       "banana")  ==  "nnaaa"
> sort_text(order: "<na>",                      "banana")  ==  "na"
> sort_text(order: "once(na)",                  "banana")  ==  "na"
> sort_text(order: "[na]",                      "banana")  ==  "anana"
> sort_text(order: "mixed(na)",                 "banana")  ==  "anana"
> sort_text(order: "compound(na)",              "banana")  ==  "nana"
> sort_text(order: "reverse_order(na)",         "banana")  ==  "aaann"
> sort_text(order: "a n <a>",                   "banana")  ==  "aaann"
> sort_text(order: "reverse_order(<a> n a)",    "banana")  ==  "aanna"
> sort_text(order: "pattern(./. cycle(wubrg))", "wgw/g")   ==  "g/w"
> sort_text(order: "[1234567890]cycle(wubrg)",  "21wg")    ==  "21gw"
>
> f := sort_rule(order: "[1234567890]cycle(wubrg)")
> f("21wg")  ==  "21gw"