Question

In Hebrew, there are certain vowels that NSPredicate fails to ignore even when using the ‘d’ (diacritic insensitive) modifier in the predicate. I was told that the solution is to use regular expressions to do the search.

How do I take a search string and “use regex” to search hebrew text that contains vowels, ignoring those vowels?

Edit:

In other words, If I wanted to search the following text, disregarding dashes and asterisks, how would I do so using regex?

Example Text:

I w-en*t t-o the st*o*r*-e yes-ster*day.

Edit 2:

Essentially, I want to:

1. Take an input string from a user
2. Take a string to search
3. Use a regex based on the user’s search string to search for “contains” matches in the larger block of text. The regex should ignore vowels as shown above.

Edit 3:

Here’s how I’m implementing my search:

//
//  The user updated the search text
//

- (BOOL)searchDisplayController:(UISearchDisplayController *)controller 
shouldReloadTableForSearchString:(NSString *)searchString{

    NSMutableArray *unfilteredResults = [[[[self.fetchedResultsController sections] objectAtIndex:0] objects] mutableCopy];

    if (self.filteredArray == nil) {
        self.filteredArray = [[[NSMutableArray alloc ] init] autorelease];
    }

    [filteredArray removeAllObjects];

    NSPredicate *predicate;

    if (controller.searchBar.selectedScopeButtonIndex == 0) {
        predicate = [NSPredicate predicateWithFormat:@"articleTitle CONTAINS[cd] %@", searchString];
    }else if (controller.searchBar.selectedScopeButtonIndex == 1) {
        predicate = [NSPredicate predicateWithFormat:@"articleContent CONTAINS[cd] %@", searchString];            
    }else if (controller.searchBar.selectedScopeButtonIndex == 2){
        predicate = [NSPredicate predicateWithFormat:@"ANY tags.tagText CONTAINS[cd] %@", searchString];
    }else{
        predicate = [NSPredicate predicateWithFormat:@"(ANY tags.tagText CONTAINS[cd] %@) OR (dvarTorahTitle CONTAINS[cd] %@) OR (dvarTorahContent CONTAINS[cd] %@)", searchString,searchString,searchString];
    }

    for (Article *article in unfilteredResults) {

        if ([predicate evaluateWithObject:article]) {
            [self.filteredArray addObject:article];
        }

    }

    [unfilteredResults release];


    return YES;
}

Edit 4:

I am not required to use regex for this, was just advised to do so. If you have another way that works, go for it!

Edit 5:

I’ve modified my search to look like this:

NSInteger length = [searchString length];

NSString *vowelsAsRegex = @"[\\u5B0-\\u55C4]*";

NSMutableString *modifiedSearchString = [searchString mutableCopy];

for (int i = length; i > 0; i--) {
    [modifiedSearchString insertString:vowelsAsRegex atIndex:i];
}

if (controller.searchBar.selectedScopeButtonIndex == 0) {
            predicate = [NSPredicate predicateWithFormat:@"articleTitle CONTAINS[cd] %@", modifiedSearchString];
        }else if (controller.searchBar.selectedScopeButtonIndex == 1) {
            predicate = [NSPredicate predicateWithFormat:@"articleContent CONTAINS[cd] %@", modifiedSearchString];            
        }else if (controller.searchBar.selectedScopeButtonIndex == 2){
            predicate = [NSPredicate predicateWithFormat:@"ANY tags.tagText CONTAINS[cd] %@", modifiedSearchString];
        }else{
            predicate = [NSPredicate predicateWithFormat:@"(ANY tags.tagText CONTAINS[cd] %@) OR (dvarTorahTitle CONTAINS[cd] %@) OR (dvarTorahContent CONTAINS[cd] %@)", modifiedSearchString,modifiedSearchString,modifiedSearchString];
        }

for (Article *article in unfilteredResults) {
  if ([predicate evaluateWithObject:article]) {
    [self.filteredArray addObject:article];
  }          
 }

I’m still missing something here, what do I need to do to make this work?

Edit 6:

Okay, almost there. I need to make two more changes to be finished with this.

I need to be able to add other ranges of characters to the regex, which might appear instead of, or in addition to the character in the other set. I’ve trie changing the first range to this:

[\u05b0-\u05c, \u0591-\u05AF]?

Something tells me that this is incorrect.

Also, I need the rest of the regex to be case insensitive. What modifier do I need to use with the .* regex to make it case insensitive?

Answer 1

This answer picks up where the question left off. Please read that for context.

As it turns out, iOS can make regular expressions case insensitive using an Objective-C modifier to NSPredicate. All that’s left is to combine the two ranges. I realized that they are actually two consecutive ranges. My final code looks like this:

NSInteger length = [searchString length];

NSString *vowelsAsRegex = @"[\u0591-\u05c4]?[\u0591-\u05c4]?"; //Cantillation: \u0591-\u05AF Vowels: \u05b0-\u05c

NSMutableString *modifiedSearchString = [searchString mutableCopy];

for (int i = length; i > 0; i--) {
    [modifiedSearchString insertString:vowelsAsRegex atIndex:i];
}

if (controller.searchBar.selectedScopeButtonIndex == 0) {
  predicate = [NSPredicate predicateWithFormat:@"articleTitle CONTAINS[cd] %@", modifiedSearchString];
}else if (controller.searchBar.selectedScopeButtonIndex == 1) {
    predicate = [NSPredicate predicateWithFormat:@"articleContent CONTAINS[c] %@", modifiedSearchString];            
}else if (controller.searchBar.selectedScopeButtonIndex == 2){
    predicate = [NSPredicate predicateWithFormat:@"ANY tags.tagText CONTAINS[c] %@", modifiedSearchString];
}else{
    predicate = [NSPredicate predicateWithFormat:@"(ANY tags.tagText CONTAINS[c] %@) OR (dvarTorahTitle CONTAINS[c] %@) OR (dvarTorahContent CONTAINS[c] %@)", modifiedSearchString,modifiedSearchString,modifiedSearchString];
}

[modifiedSearchString release];

for (Article *article in unfilteredResults) {
  if ([predicate evaluateWithObject:article]) {
    [self.filteredArray addObject:article];
  }          
}

Note that the range portion of the regular expression repeats itself. This is because there can be both a cantillation mark and a vowel on a single letter. Now, I can search uppercase and lowercase English, and Hebrew with or without vowels and cantillation marks.

Awesome!