iphone - iOS NSXMLParsing unformed HTML

Question

This is my method body for parsing "img src" image links from poorly formed html generated by an RSS feed... I am aware that NSXML only parses XML, but I have this hope that it can stumble through the mess to find these miniscule image links from messy html.

I'm trying to retrieve ONLY the FIRST image link found in the src attribute I find in each element name called IMG in nsData that has a src attribute and then save it to a NSString *img in another class. The img tags are not all the same, for instance an instance of nsData will contain only one image instance like any one of these:

< img class="ms-rteStyle-photoCredit" src="www.imagelinkthatineed.com" stuff I don't need

< img alt="" src="www.imagelinkineedfortableimagecellpreview" stuff I don't need

< img class="ms-rteStyle-photoCredit" src="www.IneedThisLink.com" more stuff I don't need

The only class that seems to generate NSLog output is the first one.

How can I get the parser methods to actually run ?

Given that there's a way, is there a different, simpler way you recommend?

#import "HtmlParser.h"
#import "ArticleItem.h"

@implementation HtmlParser
@synthesize elementArray;

- (HtmlParser *) InitHtmlByString:(NSString *)string {
//    NSString *description = [NSString string];
NSData *nsData = [[NSData alloc] initWithContentsOfFile:(NSString *)string];
elementArray = [[NSMutableArray alloc] init];
parser = [[NSXMLParser alloc] initWithData:nsData];
parser.delegate = self;
[parser parse];

If I NSLog(@"%@", nsData); in this method body, the output spits out the raw HTML.

currentHTMLElement = [ArticleItem alloc];
return self;
}
- (void)parser:(NSXMLParser *)parser didStartElement:(NSString *)elementName namespaceURI:(NSString *)namespaceURI qualifiedName:(NSString *)qualifiedName attributes:(NSDictionary *)attributeDict
{
if ([elementName isEqualToString:@"img src"]) {
    currentHTMLElement = [[ArticleItem alloc] init];
}
NSLog(@"%@ found a %@ element", self, elementName);
}
- (void) parser:(NSXMLParser *)parser foundCharacters:(NSString *)string
{
if (!currentHTMLElement)
    currentHTMLElement = [[NSMutableString alloc] initWithString:string];   
NSLog(@"Processing Value: %@", currentHTMLElement);
}
- (void) parser:(NSXMLParser *)parser didEndElement:(NSString *)elementName namespaceURI:(NSString *)namespaceURI qualifiedName:(NSString *)qName   
{
if ([elementName isEqualToString:@"img src"])
    {
        currentHTMLElement.img = elementName;
        [elementArray addObject:currentHTMLElement];
        currentHTMLElement = nil;
        currentNodeContent = nil;
    }
else
{
    if (currentHTMLElement !=nil && elementName != nil && ([elementName isEqualToString:@"img src"]))
    {
        [currentHTMLElement setValue:currentHTMLElement forKey:elementName];
    }
}
    currentHTMLElement = nil;
}                
@end

Thank you for your thoughts.

Answer 1

Given that HTML is generally not well-formed XML, NSXMLParser might not work. If you want to parse HTML, you might refer to this Ray Wenderlich article, How to Parse HTML on iOS. If you've followed those instructions and have added Hpple to your project, you can then retrieve the image src attributes like so:

#import "TFHpple.h"

- (void)retrieveImageSourceTagsViaHpple:(NSURL *)url
{
    NSData *data = [NSData dataWithContentsOfURL:url];

    TFHpple *parser = [TFHpple hppleWithHTMLData:data];

    NSString *xpathQueryString = @"//img";
    NSArray *nodes = [parser searchWithXPathQuery:xpathQueryString];

    for (TFHppleElement *element in nodes)
    {
        NSString *src = [element objectForKey:@"src"];
        NSLog(@"img src: %@", src);
    }
}

---

Alternatively, and I say this bracing myself for the onslaught of anti-NSRegularExpression responses (in the vein of my all-time favorite Stack Overflow answer), if you want a list of img tags in an html file, you can use the following somewhat complicated regular expression:

- (void)retrieveImageSourceTagsViaRegex:(NSURL *)url
{
    NSString *string = [NSString stringWithContentsOfURL:url
                                                encoding:NSUTF8StringEncoding
                                                   error:nil];

    NSError *error = NULL;
    NSRegularExpression *regex = [NSRegularExpression regularExpressionWithPattern:@"(<img\s[\s\S]*?src\s*?=\s*?['"](.*?)['"][\s\S]*?>)+?"
                                                                           options:NSRegularExpressionCaseInsensitive
                                                                             error:&error];

    [regex enumerateMatchesInString:string
                            options:0
                              range:NSMakeRange(0, [string length])
                         usingBlock:^(NSTextCheckingResult *result, NSMatchingFlags flags, BOOL *stop) {

                             NSString *src = [string substringWithRange:[result rangeAtIndex:2]];
                             NSLog(@"img src: %@", src);
                         }];
}

---

If you wanted to use NSXMLParser, it would look like so:

- (void)retrieveImageSourceTagsViaNSXMLParser:(NSURL *)url
{
    NSXMLParser *parser = [[NSXMLParser alloc] initWithContentsOfURL:url];
    parser.delegate = self;
    [parser parse];
}

#pragma mark - NSXMLParserDelegate methods

- (void)parser:(NSXMLParser *)parser didStartElement:(NSString *)elementName namespaceURI:(NSString *)namespaceURI qualifiedName:(NSString *)qName attributes:(NSDictionary *)attributeDict
{
    if ([elementName isEqualToString:@"img"])
    {
        NSString *src = attributeDict[@"src"];

        NSLog(@"img src: %@", src);
    }
}

The problem is, in my experience, NSXMLParser is less successful in parsing HTML than LibXML2/Hpple is. I find that on some simple pages, the above works great. But in other situations, it doesn't. Bottom line, While NSXMLParser is great at parsing well-formed XML, I'd be wary of using it for the parsing of HTML.