#import "HTMLParser.h"

enum {

HTMLParserSupportedTagBold = 1 << 0,

HTMLParserSupportedTagItalic = 1 << 1,

HTMLParserSupportedTagUnderline = 1 << 2,

HTMLParserSupportedTagLink = 1 << 3,

HTMLParserSupportedTagParagraph = 1 << 4

};

typedef NSInteger HTMLParserSupportedTags;

@interface HTMLParser ()

@property (assign, nonatomic) CGFloat fontSize;

@property (strong, nonatomic) NSString * fontName;

@property (strong, nonatomic) NSString * boldFontName;

@property (strong, nonatomic) NSString * italicFontName;

@property (strong, nonatomic) NSString * boldItalicFontName;

@end

@implementation HTMLParser

- (id)init

{

[self doesNotRecognizeSelector:_cmd];

return nil;

}

- (id)initWithFontSize:(CGFloat)fontSize fontName:(NSString *)fontName boldFontName:(NSString *)boldFontName italicFontName:(NSString *)italicFontName boldItalicFontName:(NSString *)boldItalicFontName

{

self = [super init];

if (self) {

self.fontSize = fontSize;

self.fontName = fontName;

self.boldFontName = boldFontName;

self.italicFontName = italicFontName;

self.boldItalicFontName = boldItalicFontName;

}

return self;

}

- (NSMutableAttributedString *)attributedStringFromHTMLString:(NSString *)htmlString

{

NSInteger const RegexCaptureRangeIndexStart = 1;

NSInteger const RegexCaptureRangeIndexOpenOrCloseTag = 2;

NSInteger const RegexCaptureRangeIndexTagName = 3;

NSInteger const RegexCaptureRangeIndexURL = 4;

// No need for capture 5

NSInteger const RegexCaptureRangeIndexClosingText = 6;

// Start with an empty attributed string.

NSMutableAttributedString *attributedString = [[NSMutableAttributedString alloc] initWithString:@""];

// Get the regex from a class method so it can be reused.

NSRegularExpression *splitRegex = [HTMLParser HTMLSplittingRegularExpression];

// Split the entire fragment and we'll deal with one match at a time.

NSArray *splitMatches = [splitRegex matchesInString:htmlString options:0 range:NSMakeRange(0, [htmlString length])];

// Keep track of which tags are open.

HTMLParserSupportedTags currentOpenTags = 0;

// Keep track of where in the string an open link/p tag started and what the URL was (if applicable).

NSInteger startLinkOffset = -1;

NSInteger startParagraphOffset = -1;

NSURL *lastCapturedLinkURL = nil;

for (NSTextCheckingResult *result in splitMatches) {

// Add an attributed string created with the start capture.

NSRange startCaptureRange = [result rangeAtIndex:RegexCaptureRangeIndexStart];

if (startCaptureRange.length != 0) {

NSString *startString = [[htmlString substringWithRange:startCaptureRange] gtm_stringByUnescapingFromHTML];

startString = [HTMLParser removeDuplicateWhitespaceFromString:startString];

[attributedString appendAttributedString:[self attributedStringFromString:startString withAttributes:currentOpenTags]];

}

// Get the URL out if it's there (only there if the match was a a tag)

NSURL *capturedURL = nil;

NSRange URLCaptureRange = [result rangeAtIndex:RegexCaptureRangeIndexURL];

if (URLCaptureRange.length != 0) {

// Get the URL string, trimming any whitespace around it

NSString *URLString = [[htmlString substringWithRange:URLCaptureRange] stringByTrimmingCharactersInSet:[NSCharacterSet whitespaceAndNewlineCharacterSet]];

// Escape the URL string so we can create an NSURL

URLString = [URLString stringByAddingPercentEscapesUsingEncoding:NSUTF8StringEncoding];

capturedURL = [NSURL URLWithString:URLString];

}

// Act appropriately based on tag type

HTMLParserSupportedTags tagType = 0;

NSRange tagCaptureRange = [result rangeAtIndex:RegexCaptureRangeIndexTagName];

// This should always be captured, but check length anyways so we avoid a crash if some bad data in was allowed in.

if (tagCaptureRange.length != 0) {

NSString *tagName = [htmlString substringWithRange:tagCaptureRange];

if ([tagName isEqualToString:@"strong"] || [tagName isEqualToString:@"b"]) {

tagType = HTMLParserSupportedTagBold;

} else if ([tagName isEqualToString:@"em"] || [tagName isEqualToString:@"i"]) {

tagType = HTMLParserSupportedTagItalic;

} else if ([tagName isEqualToString:@"u"]) {

tagType = HTMLParserSupportedTagUnderline;

} else if ([tagName isEqualToString:@"a"]) {

tagType = HTMLParserSupportedTagLink;

} else if ([tagName isEqualToString:@"p"]){

tagType = HTMLParserSupportedTagParagraph;

} else {

// Unsupported tag, just ignore it

}

}

// Determin if this is an opening or closing tag

NSRange openCloseCaptureRange = [result rangeAtIndex:RegexCaptureRangeIndexOpenOrCloseTag];

if (openCloseCaptureRange.length != 0) {

// This is a closing tag.

// Exclusive OR with the tagType so we can turn off the tag that just closed.

currentOpenTags ^= tagType;

// If this is a closing tag and 'a' we need to store the offset of the link and the URL it points to.

if (tagType == HTMLParserSupportedTagLink) {

// Avoid empty links

NSInteger endLinkOffset = [attributedString length];

if (startLinkOffset != endLinkOffset) {

// Set the attributes to create the link on the attributed string using the correct range and the last captured URL.

[attributedString addAttribute:NSLinkAttributeName value:lastCapturedLinkURL range:NSMakeRange(startLinkOffset, endLinkOffset - startLinkOffset)];

}

} else if (tagType == HTMLParserSupportedTagParagraph) {

// Avoid empty paragraphs

NSInteger endParagraphOffset = [attributedString length];

if (startParagraphOffset != endParagraphOffset) {

// Need to add a paragraph separator unicode character

[attributedString appendAttributedString:[[NSAttributedString alloc] initWithString:@"\u2029"]];

}

}

} else {

// This is a opening tag.

// Need to turn current tag to 'on' by OR'ing with the tag.

currentOpenTags |= tagType;

if (tagType == HTMLParserSupportedTagLink) {

// Need to keep track of where this link started

startLinkOffset = [attributedString length];

// Store the last link URL we have seen so we can use it when we find the closing 'a' tag.

lastCapturedLinkURL = capturedURL;

} else if (tagType == HTMLParserSupportedTagParagraph) {

// Need to keep track of where this paragraph started

startParagraphOffset = [attributedString length];

}

}

// Check if there's any trailing text without tags to add to the string.

// This might be the only capture in the fragment if the fragment did not have any tags.

NSRange closingTextCaptureRange = [result rangeAtIndex:RegexCaptureRangeIndexClosingText];

if (closingTextCaptureRange.length != 0) {

NSString *endString = [[htmlString substringWithRange:closingTextCaptureRange] gtm_stringByUnescapingFromHTML];

endString = [HTMLParser removeDuplicateWhitespaceFromString:endString];

// Whatever tags are still open are what style we will use. This means an unclosed tag applies to the end of the fragment.

[attributedString appendAttributedString:[self attributedStringFromString:endString withAttributes:currentOpenTags]];

}

}

return attributedString;

}

- (NSAttributedString *)attributedStringFromString:(NSString *)string withAttributes:(HTMLParserSupportedTags)openAttributes

{

NSMutableAttributedString *attributedString = [[NSMutableAttributedString alloc] initWithString:string];

UIFont *font = nil;

if ((openAttributes & HTMLParserSupportedTagBold) && (openAttributes & HTMLParserSupportedTagItalic)) {

font = [UIFont fontWithName:self.boldItalicFontName size:self.fontSize];

} else if (openAttributes & HTMLParserSupportedTagBold) {

font = [UIFont fontWithName:self.boldFontName size:self.fontSize];

} else if (openAttributes & HTMLParserSupportedTagItalic) {

font = [UIFont fontWithName:self.italicFontName size:self.fontSize];

} else {

// We don't have bold or italic on at the moment, use the default font

font = [UIFont fontWithName:self.fontName size:self.fontSize];

}

if (font) {

[attributedString addAttribute:NSFontAttributeName value:font range:NSMakeRange(0, [attributedString length])];

}

if (openAttributes & HTMLParserSupportedTagUnderline) {

[attributedString addAttribute:(NSString *)kCTUnderlineStyleAttributeName value:[NSNumber numberWithInt:1] range:NSMakeRange(0, [attributedString length])];

}

return attributedString;

}

#pragma mark Class methods

+ (NSString *)removeDuplicateWhitespaceFromString:(NSString *)string

{

// TODO: This should probably be in a category on NSString.

NSError *error = nil;

NSRegularExpression *regex = [NSRegularExpression regularExpressionWithPattern:@" +" options:NSRegularExpressionCaseInsensitive error:&error];

string = [regex stringByReplacingMatchesInString:string options:0 range:NSMakeRange(0, [string length]) withTemplate:@" "];

regex = [NSRegularExpression regularExpressionWithPattern:@"\n\n+" options:NSRegularExpressionCaseInsensitive error:&error];

string = [regex stringByReplacingMatchesInString:string options:0 range:NSMakeRange(0, [string length]) withTemplate:@"\n"];

return string;

}

+ (NSRegularExpression *)HTMLSplittingRegularExpression

{

static dispatch_once_t predicate = 0;

__strong static NSRegularExpression *sharedHTMLSplittingRegularExpression = nil;

dispatch_once(&predicate, ^{

// Split entire html fragment on strings/tags or end of fragment

NSError *error = nil;

// This looks complex but it's mostly just dealing with slight edge cases in the html like extra spaces. Everything we need is captured and then we checked what we got

// and process appropriately.

sharedHTMLSplittingRegularExpression = [NSRegularExpression

regularExpressionWithPattern:@"(?:(?:(.*?)<\\s*(/??)(\\w+)(?:(?:[^>]*?(?:(?:src)|(?:href))=(?:\'|\"){1}(.*?)(?:\'|\"){1}.*?/?)|(.*?))>)|(.*$))"

options:NSRegularExpressionCaseInsensitive | NSRegularExpressionDotMatchesLineSeparators // Needed to get to the start of the text which can be on a different line

error:&error];

// Find problems with this regex during dev, it should never error.

NSAssert1(sharedHTMLSplittingRegularExpression, @"Regex invalid: %@", error);

});

return sharedHTMLSplittingRegularExpression;

}

@end