[iphone] Convert HTML to NSAttributedString in iOS

I am using a instance of UIWebView to process some text and color it correctly, it gives the result as HTML but rather than displaying it in the UIWebView I want to display it using Core Text with a NSAttributedString.

I am able to create and draw the NSAttributedString but I am unsure how I can convert and map the HTML into the attributed string.

I understand that under Mac OS X NSAttributedString has a initWithHTML: method but this was a Mac only addition and is not available for iOS.

I also know that there is a similar question to this but it had no answers, I though I would try again and see whether anyone has created a way to do this and if so, if they could share it.

honoring font family, dynamic font I've concocted this abomination:

extension NSAttributedString
{
    convenience fileprivate init?(html: String, font: UIFont? = Font.dynamic(style: .subheadline))
    {
        guard let data = html.data(using: String.Encoding.utf8, allowLossyConversion: true) else {
        var totalString = html
        /*
         https://stackoverflow.com/questions/32660748/how-to-use-apples-new-san-francisco-font-on-a-webpage
            .AppleSystemUIFont I get in font.familyName does not work
         while -apple-system does:
         */
        var ffamily = "-apple-system"
        if let font = font {
            let lLDBsucks = font.familyName
            if !lLDBsucks.hasPrefix(".appleSystem") {
                ffamily = font.familyName
            }
            totalString = "<style>\nhtml * {font-family: \(ffamily) !important;}\n            </style>\n" + html
        }
        guard let data = totalString.data(using: String.Encoding.utf8, allowLossyConversion: true) else {
            return nil
        }
        assert(Thread.isMainThread)
        guard let attributedText = try?  NSAttributedString(data: data, options: [.documentType: NSAttributedString.DocumentType.html, .characterEncoding: String.Encoding.utf8.rawValue], documentAttributes: nil) else {
            return nil
        }
        let mutable = NSMutableAttributedString(attributedString: attributedText)
        if let font = font {
        do {
            var found = false
            mutable.beginEditing()
            mutable.enumerateAttribute(NSAttributedString.Key.font, in: NSMakeRange(0, attributedText.length), options: NSAttributedString.EnumerationOptions(rawValue: 0)) { (value, range, stop) in
                    if let oldFont = value as? UIFont {
                        let newsize = oldFont.pointSize * 15 * Font.scaleHeruistic / 12
                        let newFont = oldFont.withSize(newsize)
                        mutable.addAttribute(NSAttributedString.Key.font, value: newFont, range: range)
                        found = true
                    }
                }
                if !found {
                    // No font was found - do something else?
                }

            mutable.endEditing()
            
//            mutable.addAttribute(.font, value: font, range: NSRange(location: 0, length: mutable.length))
        }
        self.init(attributedString: mutable)
    }

}

alternatively you can use the versions this was derived from and set font on UILabel after setting attributedString

this will clobber the size and boldness encapsulated in the attributestring though

kudos for reading through all the answers up to here. You are a very patient man woman or child.

The above solution is correct.

[[NSAttributedString alloc] initWithData:[htmlString dataUsingEncoding:NSUTF8StringEncoding] 
                                 options:@{NSDocumentTypeDocumentAttribute: NSHTMLTextDocumentType,
                                           NSCharacterEncodingDocumentAttribute: @(NSUTF8StringEncoding)} 
                      documentAttributes:nil error:nil];

But the app wioll crash if you are running it on ios 8.1,2 or 3.

To avoid the crash what you can do is : run this in a queue. So that it always be on main thread.

Using of NSHTMLTextDocumentType is slow and it is hard to control styles. I suggest you to try my library which is called Atributika. It has its own very fast HTML parser. Also you can have any tag names and define any style for them.

Example:

let str = "<strong>Hello</strong> World!".style(tags:
    Style("strong").font(.boldSystemFont(ofSize: 15))).attributedString

label.attributedText = str

You can find it here https://github.com/psharanda/Atributika

There is a work-in-progress open source addition to NSAttributedString by Oliver Drobnik at Github. It uses NSScanner for HTML parsing.

The only solution you have right now is to parse the HTML, build up some nodes with given point/font/etc attributes, then combine them together into an NSAttributedString. It's a lot of work, but if done correctly, can be reusable in the future.

Helpful Extensions

Inspired by this thread, a pod, and Erica Sadun's ObjC example in iOS Gourmet Cookbook p.80, I wrote an extension on String and on NSAttributedString to go back and forth between HTML plain-strings and NSAttributedStrings and vice versa -- on GitHub here, which I have found helpful.

The signatures are (again, full code in a Gist, link above):

extension NSAttributedString {
    func encodedString(ext: DocEXT) -> String?
    static func fromEncodedString(_ eString: String, ext: DocEXT) -> NSAttributedString? 
    static func fromHTML(_ html: String) -> NSAttributedString? // same as above, where ext = .html
}

extension String {
    func attributedString(ext: DocEXT) -> NSAttributedString?
}

enum DocEXT: String { case rtfd, rtf, htm, html, txt }

Here is the Swift 5 version of Mobile Dan's answer:

public extension NSAttributedString {
    convenience init?(_ html: String) {
        guard let data = html.data(using: .unicode) else {
                return nil
        }

        try? self.init(data: data, options: [.documentType: NSAttributedString.DocumentType.html], documentAttributes: nil)
    }
}

Swift 3:
Try this:

extension String {
    func htmlAttributedString() -> NSAttributedString? {
        guard let data = self.data(using: String.Encoding.utf16, allowLossyConversion: false) else { return nil }
        guard let html = try? NSMutableAttributedString(
            data: data,
            options: [NSDocumentTypeDocumentAttribute: NSHTMLTextDocumentType],
            documentAttributes: nil) else { return nil }
        return html
    }
}  

And for using:

let str = "<h1>Hello bro</h1><h2>Come On</h2><h3>Go sis</h3><ul><li>ME 1</li><li>ME 2</li></ul> <p>It is me bro , remember please</p>"

self.contentLabel.attributedText = str.htmlAttributedString()

Creating an NSAttributedString from HTML must be done on the main thread!

Update: It turns out that NSAttributedString HTML rendering depends on WebKit under the hood, and must be run on the main thread or it will occasionally crash the app with a SIGTRAP.

New Relic crash log:

Below is an updated thread-safe Swift 2 String extension:

extension String {
    func attributedStringFromHTML(completionBlock:NSAttributedString? ->()) {
        guard let data = dataUsingEncoding(NSUTF8StringEncoding) else {
            print("Unable to decode data from html string: \(self)")
            return completionBlock(nil)
        }

        let options = [NSDocumentTypeDocumentAttribute : NSHTMLTextDocumentType,
                   NSCharacterEncodingDocumentAttribute: NSNumber(unsignedInteger:NSUTF8StringEncoding)]

        dispatch_async(dispatch_get_main_queue()) {
            if let attributedString = try? NSAttributedString(data: data, options: options, documentAttributes: nil) {
                completionBlock(attributedString)
            } else {
                print("Unable to create attributed string from html string: \(self)")
                completionBlock(nil)
            }
        }
    }
}

Usage:

let html = "<center>Here is some <b>HTML</b></center>"
html.attributedStringFromHTML { attString in
    self.bodyLabel.attributedText = attString
}

Output:

This is a String extension written in Swift to return a HTML string as NSAttributedString.

extension String {
    func htmlAttributedString() -> NSAttributedString? {
        guard let data = self.dataUsingEncoding(NSUTF16StringEncoding, allowLossyConversion: false) else { return nil }
        guard let html = try? NSMutableAttributedString(data: data, options: [NSDocumentTypeDocumentAttribute: NSHTMLTextDocumentType], documentAttributes: nil) else { return nil }
        return html
    }
}

To use,

label.attributedText = "<b>Hello</b> \u{2022} babe".htmlAttributedString()

In the above, I have purposely added a unicode \u2022 to show that it renders unicode correctly.

A trivial: The default encoding that NSAttributedString uses is NSUTF16StringEncoding (not UTF8!).

Swift initializer extension on NSAttributedString

My inclination was to add this as an extension to NSAttributedString rather than String. I tried it as a static extension and an initializer. I prefer the initializer which is what I've included below.

Swift 4

internal convenience init?(html: String) {
    guard let data = html.data(using: String.Encoding.utf16, allowLossyConversion: false) else {
        return nil
    }

    guard let attributedString = try?  NSAttributedString(data: data, options: [.documentType: NSAttributedString.DocumentType.html, .characterEncoding: String.Encoding.utf8.rawValue], documentAttributes: nil) else {
        return nil
    }

    self.init(attributedString: attributedString)
}

Swift 3

extension NSAttributedString {

internal convenience init?(html: String) {
    guard let data = html.data(using: String.Encoding.utf16, allowLossyConversion: false) else {
        return nil
    }

    guard let attributedString = try? NSMutableAttributedString(data: data, options: [NSAttributedString.DocumentReadingOptionKey.documentType: NSAttributedString.DocumentType.html], documentAttributes: nil) else {
        return nil
    }

    self.init(attributedString: attributedString)
}
}

Example

let html = "<b>Hello World!</b>"
let attributedString = NSAttributedString(html: html)

Made some modification on Andrew's solution and update the code to Swift 3:

This code now use UITextView as self and able to inherit its original font, font size and text color

Note: toHexString() is extension from here

extension UITextView {
    func setAttributedStringFromHTML(_ htmlCode: String, completionBlock: @escaping (NSAttributedString?) ->()) {
        let inputText = "\(htmlCode)<style>body { font-family: '\((self.font?.fontName)!)'; font-size:\((self.font?.pointSize)!)px; color: \((self.textColor)!.toHexString()); }</style>"

        guard let data = inputText.data(using: String.Encoding.utf16) else {
            print("Unable to decode data from html string: \(self)")
            return completionBlock(nil)
        }

        DispatchQueue.main.async {
            if let attributedString = try? NSAttributedString(data: data, options: [NSDocumentTypeDocumentAttribute: NSHTMLTextDocumentType], documentAttributes: nil) {
                self.attributedText = attributedString
                completionBlock(attributedString)
            } else {
                print("Unable to create attributed string from html string: \(self)")
                completionBlock(nil)
            }
        }
    }
}

Example usage:

mainTextView.setAttributedStringFromHTML("<i>Hello world!</i>") { _ in }

The built in conversion always sets the text color to UIColor.black, even if you pass an attributes dictionary with .forgroundColor set to something else. To support DARK mode on iOS 13, try this version of the extension on NSAttributedString.

extension NSAttributedString {
    internal convenience init?(html: String)                    {
        guard 
            let data = html.data(using: String.Encoding.utf16, allowLossyConversion: false) else { return nil }

        let options : [DocumentReadingOptionKey : Any] = [
            .documentType: NSAttributedString.DocumentType.html,
            .characterEncoding: String.Encoding.utf8.rawValue
        ]

        guard
            let string = try? NSMutableAttributedString(data: data, options: options,
                                                 documentAttributes: nil) else { return nil }

        if #available(iOS 13, *) {
            let colour = [NSAttributedString.Key.foregroundColor: UIColor.label]
            string.addAttributes(colour, range: NSRange(location: 0, length: string.length))
        }

        self.init(attributedString: string)
    }
}

Swift 4

• NSAttributedString convenience initializer
• Without extra guards
• throws error

extension NSAttributedString {

    convenience init(htmlString html: String) throws {
        try self.init(data: Data(html.utf8), options: [
            .documentType: NSAttributedString.DocumentType.html,
            .characterEncoding: String.Encoding.utf8.rawValue
        ], documentAttributes: nil)
    }

}

Usage

UILabel.attributedText = try? NSAttributedString(htmlString: "<strong>Hello</strong> World!")

Swift 3.0 Xcode 8 Version

func htmlAttributedString() -> NSAttributedString? {
    guard let data = self.data(using: String.Encoding.utf16, allowLossyConversion: false) else { return nil }
    guard let html = try? NSMutableAttributedString(data: data, options: [NSDocumentTypeDocumentAttribute: NSHTMLTextDocumentType], documentAttributes: nil) else { return nil }
    return html
}