iPhone Objective C - How to remove URLs from an NSString

Question

I am looking for an efficient way to replace URLs in an NSString (or NSMutableString) with the replacement word 'link', for example ...

@"This is a sample with **http://bitbanter.com** within the string and heres another **http://spikyorange.co.uk** for luck"

I would like this to become...

@"This is a sample with **'link'** within the string and heres another **'link'** for luck"

Ideally, I would like this to be some sort of method that accepts regular expressions, but, this needs to work on the iPhone, preferably without any libraries, or, I could be persuaded if the library was tiny.

Other features that would be handy, replace @"OMG" with @"Oh my God" , but not when it's part of a word, ie @"DOOMGAME" shouldn't be touched.

Any suggestions appreciated.

Regards, Rob.

Answer 1

This was actually quite a bit of fun to play with and hopefully the solution is somehow what you were looking for. This is flexible enough to cater not only for links, but also other patterns where you may want to replace a word for another using certain conditions:

I have commented most of the code so it should be pretty self explanatory. If not, feel free to leave a comment and I will try my best to help:

- (NSString*)replacePattern:(NSString*)pattern withReplacement:(NSString*)replacement forString:(NSString*)string usingCharacterSet:(NSCharacterSet*)characterSetOrNil
{
    // Check if a NSCharacterSet has been provided, otherwise use our "default" one
    if (!characterSetOrNil)
    characterSetOrNil = [NSCharacterSet characterSetWithCharactersInString:@" !?,()]"];

    // Create a mutable copy of the string supplied, setup all the default variables we'll need to use
    NSMutableString *mutableString = [[[NSMutableString alloc] initWithString:string] autorelease];
    NSString *beforePatternString = nil;
    NSRange outputrange = NSMakeRange(0, 0);

    // Check if the string contains the "pattern" you're looking for, otherwise simply return it.
    NSRange containsPattern = [mutableString rangeOfString:pattern];
    while (containsPattern.location != NSNotFound)
    // Found the pattern, let's run with the changes
    {
        // Firstly, we grab the full string range
        NSRange stringrange = NSMakeRange(0, [mutableString length]);
        NSScanner *scanner = [[NSScanner alloc] initWithString:mutableString];

        // Now we use NSScanner to scan UP TO the pattern provided
        [scanner scanUpToString:pattern intoString:&beforePatternString];

        // Check for nil here otherwise you will crash - you will get nil if the pattern is at the very beginning of the string
        // outputrange represents the range of the string right BEFORE your pattern
        // We need this to know where to start searching for our characterset (i.e. end of output range = beginning of our pattern)
        if (beforePatternString != nil)
            outputrange = [mutableString rangeOfString:beforePatternString];

        // Search for any of the character sets supplied to know where to stop.
        // i.e. for a URL you'd be looking at non-URL friendly characters, including spaces (this may need a bit more research for an exhaustive list)
        NSRange characterAfterPatternRange = [mutableString rangeOfCharacterFromSet:characterSetOrNil options:NSLiteralSearch range:NSMakeRange(outputrange.length, stringrange.length-outputrange.length)];

        // Check if the link is not at the very end of the string, in which case there will be no characters AFTER it so set the NSRage location to the end of the string (== it's length)
        if (characterAfterPatternRange.location == NSNotFound)
            characterAfterPatternRange.location = [mutableString length];

        // Assign the pattern's start position and length, and then replace it with the pattern
        NSInteger patternStartPosition = outputrange.length;
        NSInteger patternLength = characterAfterPatternRange.location - outputrange.length;
        [mutableString replaceCharactersInRange:NSMakeRange(patternStartPosition, patternLength) withString:replacement];
        [scanner release];

        // Reset containsPattern for new mutablestring and let the loop continue
        containsPattern = [mutableString rangeOfString:pattern];
    }
    return [[mutableString copy] autorelease];
}

And to use your question as an example, here's how you could call it:

NSString *firstString = @"OMG!!!! this is the best convenience method ever, seriously! It even works with URLs like http://www.stackoverflow.com";
NSCharacterSet *characterSet = [NSCharacterSet characterSetWithCharactersInString:@" !?,()]"];
NSString *returnedFirstString = [self replacePattern:@"OMG" withReplacement:@"Oh my God" forString:firstString usingCharacterSet:characterSet];
NSString *returnedSecondString = [self replacePattern:@"http://" withReplacement:@"LINK" forString:returnedFirstString usingCharacterSet:characterSet];
NSLog (@"Original string = %@\nFirst returned string = %@\nSecond returned string = %@", firstString, returnedFirstString, returnedSecondString);

I hope it helps! Cheers, Rog

Answer 2

As of iOS 4, NSRegularExpression is available. Amongst other things, you can enumerate all matches within a string via a block, allowing you to do whatever you want to each, or have the regular expression perform some kinds of substitution directly for you.

Direct string substitutions (like 'OMG' -> 'Oh my God') can be performed directly by an NSString, using -stringByReplacingOccurencesOfString:withString: , or replaceOccurrencesOfString:withString:options:range: if your string is mutable.