How to read names in a string using NSLinguisticTagger

Swift version: 5.1

Paul Hudson @twostraws May 28th 2019

Foundation has a built-in class to parse strings of text, and it includes some useful options to extra names of people, places, organizations, and more.

To try it out, consider this string:

let text = "Apple Computer was established in Cupertino by Steve Jobs, Steve Wozniak, and Ronald Wayne."

That contains a company name, a place name, and three names of people all in one, and we can use NSLinguisticTagger to pull them all out.

First you create a linguistic tagger and tell it to look for the names of things inside that text string:

let tagger = NSLinguisticTagger(tagSchemes: [.nameType], options: 0)
tagger.string = text

Next you create the range to scan. This is done using the older NSRange type, like this:

let range = NSRange(location: 0, length: text.utf16.count)

Third, you tell NSLinguisticTagger what it should look for and how it should scan. One useful option here is .joinNames, which means it will return “Steve Jobs” as a single name rather than as two individual names:

let options: NSLinguisticTagger.Options = [.omitPunctuation, .omitWhitespace, .joinNames]
let tags: [NSLinguisticTag] = [.personalName, .placeName, .organizationName]

Finally, you tell NSLinguisticTagger to enumerate the tags in the input string, filter out any that aren’t in the tags array we’re looking for, convert the NSRange back to a Swift range, then print out each match:

tagger.enumerateTags(in: range, unit: .word, scheme: .nameType, options: options) { tag, tokenRange, stop in
    if let tag = tag, tags.contains(tag) {
        if let range = Range(tokenRange, in: text) {
            let name = text[range]
            print("\(name): \(tag)")
        }
    }
}

That will find the company, organization, and three people names in our string – nice!