How to convert to "combining diacritical marks" on iOS

Viewed 678

In my app, I have characters that are followed by their "modifier diacritical marks" (e.g. "oˆ", where the "ˆ" is unicode 0x02c6) that I want to convert into fully precomposed characters (e.g. "ô" - unicode 0x00f4). I tried using the NSString method precomposedStringWithCanonicalMapping, but after several hours of beating my head against the wall trying to figure out why it wasn't working, I discovered that it only converts "combining diacritical marks" (http://www.unicode.org/charts/PDF/U0300.pdf) into precomposed characters. Ok, so all I need to do is convert all of my "modifier diacritical marks" into "combining diacritical marks", then perform a precomposedStringWithCanonicalMapping on the resulting string and I'm done. This does work, but I wonder if there's a less tedious/error prone way to do this? Here's my NSString category method that seems to fix most of the characters-

- (instancetype)combineDiacritics
{
    static NSDictionary<NSNumber *, NSNumber *> *sDiacriticalSubstDict; //unichar of diacritic -> unichar of combining diacritic
    static dispatch_once_t onceToken;
    dispatch_once(&onceToken, ^{
        //http://www.unicode.org/charts/PDF/U0300.pdf
        sDiacriticalSubstDict = @{ @(0x02cb) : @(0x0300), @(0x00b4) : @(0x0301), @(0x02c6) : @(0x0302), @(0x02dc) : @(0x0303), @(0x02c9) : @(0x0304),   //Grave, Acute, Circumflex, Tilde, Macron
                                   @(0x00af) : @(0x0305), @(0x02d8) : @(0x0306), @(0x02d9) : @(0x0307), @(0x00a8) : @(0x0308), @(0x02c0) : @(0x0309),   //Overline, Breve, Dot above, Diaeresis
                                   @(0x00b0) : @(0x030a), @(0x02da) : @(0x030b), @(0x02c7) : @(0x030c), @(0x02c8) : @(0x030d), @(0x02bb) : @(0x0312),   //Ring above, Double Acute, Caron, Vertical line above, Cedilla above
                                   @(0x02bc) : @(0x0313), @(0x02bd) : @(0x0314), @(0x02b2) : @(0x0321), @(0x02d4) : @(0x0323), @(0x02b1) : @(0x0324),   //Comma above, Reversed comma above, Palatalized hook below, Dot below, Diaeresis below
                                   @(0x00b8) : @(0x0327), @(0x02db) : @(0x0328), @(0x02cc) : @(0x0329), @(0x02b7) : @(0x032b), @(0x02cd) : @(0x0331),   //Cedilla, Ogonek, Vert line below, Inverted double arch below, Macron below
                                   };
    });
    NSMutableString* __block buffer = [NSMutableString stringWithCapacity:self.length];
    [self enumerateSubstringsInRange:NSMakeRange(0, self.length) options:NSStringEnumerationByComposedCharacterSequences usingBlock: ^(NSString* substring, NSRange substringRange, NSRange enclosingRange, BOOL* stop) {
                          NSString *newString = nil;
                          if (substring.length == 1)    //The diacriticals are all Unicode BMP.
                          {
                              unichar uniChar = [substring characterAtIndex:0];
                              unichar newUniChar = [sDiacriticalSubstDict[@(uniChar)] integerValue];
                              if (newUniChar != 0)
                              {
                                  NSLog(@"Unichar %04x => %04x", uniChar, newUniChar);
                                  newString = [NSString stringWithCharacters:&newUniChar length:1];
                              }
                          }
                          if (newString)
                              [buffer appendString:newString];
                          else
                              [buffer appendString:substring];
                      }];

    NSString *precomposedStr = [buffer precomposedStringWithCanonicalMapping];
    return precomposedStr;
}

Does anyone know of more built-in way to make this conversion?

1 Answers
Related