HTML character decoding in Objective-C / Cocoa Touch

Solution 1:

Check out my NSString category for HTML. Here are the methods available:

- (NSString *)stringByConvertingHTMLToPlainText;
- (NSString *)stringByDecodingHTMLEntities;
- (NSString *)stringByEncodingHTMLEntities;
- (NSString *)stringWithNewLinesAsBRs;
- (NSString *)stringByRemovingNewLinesAndWhitespace;

Solution 2:

The one by Daniel is basically very nice, and I fixed a few issues there:

  1. removed the skipping character for NSSCanner (otherwise spaces between two continuous entities would be ignored

    [scanner setCharactersToBeSkipped:nil];

  2. fixed the parsing when there are isolated '&' symbols (I am not sure what is the 'correct' output for this, I just compared it against firefox):

e.g.

    &#ABC DF & B'  & C' Items (288)

here is the modified code:

- (NSString *)stringByDecodingXMLEntities {
    NSUInteger myLength = [self length];
    NSUInteger ampIndex = [self rangeOfString:@"&" options:NSLiteralSearch].location;

    // Short-circuit if there are no ampersands.
    if (ampIndex == NSNotFound) {
        return self;
    }
    // Make result string with some extra capacity.
    NSMutableString *result = [NSMutableString stringWithCapacity:(myLength * 1.25)];

    // First iteration doesn't need to scan to & since we did that already, but for code simplicity's sake we'll do it again with the scanner.
    NSScanner *scanner = [NSScanner scannerWithString:self];

    [scanner setCharactersToBeSkipped:nil];

    NSCharacterSet *boundaryCharacterSet = [NSCharacterSet characterSetWithCharactersInString:@" \t\n\r;"];

    do {
        // Scan up to the next entity or the end of the string.
        NSString *nonEntityString;
        if ([scanner scanUpToString:@"&" intoString:&nonEntityString]) {
            [result appendString:nonEntityString];
        }
        if ([scanner isAtEnd]) {
            goto finish;
        }
        // Scan either a HTML or numeric character entity reference.
        if ([scanner scanString:@"&" intoString:NULL])
            [result appendString:@"&"];
        else if ([scanner scanString:@"'" intoString:NULL])
            [result appendString:@"'"];
        else if ([scanner scanString:@""" intoString:NULL])
            [result appendString:@"\""];
        else if ([scanner scanString:@"<" intoString:NULL])
            [result appendString:@"<"];
        else if ([scanner scanString:@"&gt;" intoString:NULL])
            [result appendString:@">"];
        else if ([scanner scanString:@"&#" intoString:NULL]) {
            BOOL gotNumber;
            unsigned charCode;
            NSString *xForHex = @"";

            // Is it hex or decimal?
            if ([scanner scanString:@"x" intoString:&xForHex]) {
                gotNumber = [scanner scanHexInt:&charCode];
            }
            else {
                gotNumber = [scanner scanInt:(int*)&charCode];
            }

            if (gotNumber) {
                [result appendFormat:@"%C", (unichar)charCode];

                [scanner scanString:@";" intoString:NULL];
            }
            else {
                NSString *unknownEntity = @"";

                [scanner scanUpToCharactersFromSet:boundaryCharacterSet intoString:&unknownEntity];


                [result appendFormat:@"&#%@%@", xForHex, unknownEntity];

                //[scanner scanUpToString:@";" intoString:&unknownEntity];
                //[result appendFormat:@"&#%@%@;", xForHex, unknownEntity];
                NSLog(@"Expected numeric character entity but got &#%@%@;", xForHex, unknownEntity);

            }

        }
        else {
            NSString *amp;

            [scanner scanString:@"&" intoString:&amp];  //an isolated & symbol
            [result appendString:amp];

            /*
            NSString *unknownEntity = @"";
            [scanner scanUpToString:@";" intoString:&unknownEntity];
            NSString *semicolon = @"";
            [scanner scanString:@";" intoString:&semicolon];
            [result appendFormat:@"%@%@", unknownEntity, semicolon];
            NSLog(@"Unsupported XML character entity %@%@", unknownEntity, semicolon);
             */
        }

    }
    while (![scanner isAtEnd]);

finish:
    return result;
}