Locale-sensitive string sort in Dart - sorting

Question

I need to sort a List<String>. Its elements might contain non-ASCII characters (think é, ž, Æ, ...). Is there a way to do this in Dart while being locale-aware?
Something like localeCompare in Javascript.
Some examples:
Czech (cs_CZ)
List<String> czechWords = ['čeština', 'cvrček', 'chleba',];
// expected after calling a localizedSort() or similar method
List<String> czechWordsSorted = ['cvrček', 'čeština', 'chleba',];

Answer 1

The Dart platform libraries do not currently have any locale-specific string sorting or comparison functionality.

Answer 2

If anyone still has the same problem, here is how I solved this (for Czech in the example).
Rationale
This works by substituting the diacritics chars away & then comparing on normal strings. The trick is to expand both the normalized & unnormalized strings, but so that the order is as You want it.
For example, I want to have u before ú which is before ů, so I can ensure this if I substitute all u for ua, ú for ub, and ů for uc.
Another thing is solving ch; this is done by substituting h for ha and ch for hb (so that each ch char is just after h).
Both cases need to be done in one pass, as they interact.
Solution
const Map<String, String> collTable = {
"a": "aa", "á": "ab", "A": "Aa", "Á": "Ab", "c": "ca", "č": "cb", "C": "Ca",
"Č": "Cb", "d": "da", "ď": "db", "D": "Da", "Ď": "Db", "e": "ea", "é": "eb",
"ě": "ec", "E": "Ea", "É": "Eb", "Ě": "Ec", "i": "ia", "í": "ib", "I": "Ia",
"Í": "Ib", "n": "na", "ň": "nb", "N": "Na", "Ň": "Nb", "o": "oa", "ó": "ob",
"O": "Oa", "Ó": "Ob", "r": "ra", "ř": "rb", "R": "Ra", "Ř": "Rb", "s": "sa",
"š": "sb", "S": "Sa", "Š": "Sb", "t": "ta", "ť": "tb", "T": "Ta", "Ť": "Tb",
"u": "ua", "ú": "ub", "ů": "uc", "U": "Ua", "Ú": "Ub", "Ů": "Uc", "y": "ya",
"ý": "yb", "Y": "Ya", "Ý": "Yb", "z": "za", "ž": "zb", "Z": "Za", "Ž": "Zb",
};
String collKeys = collTable.keys.join("");
final RegExp keysPat = RegExp(r"(cH|ch|CH|Ch|h|H)|([" + collKeys + r"])");
bool isUpperCase(String str){
return str.toUpperCase() == str;
}
String collStr(String str){
return str.replaceAllMapped(keysPat, (match){
assert((match[1] == null) != (match[2] == null));
if(match[1] != null) {
// ch group = either h or ch
String g = match[1]!;
if (g.length == 1) {
// h => ha, H => Ha
return "${g}a";
} else {
assert(g.length == 2);
// need to keep the case of the letters
String subc = isUpperCase(g[0]) ? 'H' : 'h';
// ch => hb, Ch => Hb, CH => Hb, cH => hb
return "${subc}b";
}
}
// diacritics
return collTable[match[2]]!;
});
}
int cmpColl(String a, String b){
return collStr(a).compareTo(collStr(b));
}
Usage
You can then use it like this:
List<String> czechWords = ['čeština', 'cvrček', 'děda', 'chleba', 'czechia', 'houska', 'imbecilní'];
for(var w in czechWords){
debugPrint("$w -> ${collStr(w)}");
}
czechWords.sort(cmpColl);
debugPrint(czechWords.toString());
Which prints:
čeština -> cbeasbtaianaaa
cvrček -> cavracbeak
děda -> daecdaaa
chleba -> hbleabaa
czechia -> cazaeahbiaaa
houska -> haoauasakaa
imbecilní -> iambeacaialnaib
[cvrček, czechia, čeština, děda, houska, chleba, imbecilní]
Disclaimer
This definitely needs more testing :-)
Edit:
The code for comparing ch is still not perfect. It is not clear, how the order should really be with capitalization, e.g. Ch vs. CH, cH vs ch, etc. Some of these might not be a valid czech letter... Right now, the Ch and CH are both treated as capital ch, so order is the same (== resulting order depends on whether dart sort is stable).