Simple DOM based XML parser #cocoa #xml #parser #utils #dom

TJXMLParser.h

#import <Foundation/Foundation.h>

@interface TJXMLElement : NSObject

@property (nonatomic, readonly) NSString*       name;
@property (nonatomic, readonly) NSDictionary*   attributes;
@property (nonatomic, retain) NSMutableArray*   children;
@property (nonatomic, retain) NSMutableString*  text;

- (id) initWithName:(NSString*)name withAttributes:(NSDictionary*)attributes;
- (TJXMLElement*) childWithName:(NSString*)name;
- (NSArray*) childrenWithName:(NSString*)name;

@end

@interface TJXMLParser : NSObject <NSXMLParserDelegate> {
@private
    NSXMLParser*        _parser;
    NSMutableArray*     _parentList;
}

- (id) initWithContentsOfURL:(NSURL *)url;
- (id) initWithData:(NSData *)data;

- (TJXMLElement*) parse;

@property (nonatomic, retain) TJXMLElement* currentElement;
@property (nonatomic, retain) TJXMLElement* rootElement;
@property (nonatomic, retain) NSError* parseError;

@end

TJXMLParser.m

#import "TJXMLParser.h"

@implementation TJXMLElement
@synthesize name        = _name;
@synthesize attributes  = _attributes;
@synthesize children    = _children;
@synthesize text        = _text;

- (id) initWithName:(NSString*)name withAttributes:(NSDictionary*)attributes
{
    self = [super init];
    if (self != nil) {
        _name = [name retain];
        _attributes = [attributes retain];
        _children = [[NSMutableArray alloc] init];
        _text = [[NSMutableString alloc] init];
    }
    return self;
}

- (void) dealloc 
{
    [_name release]; _name = nil;
    [_attributes release]; _attributes = nil;
    self.children = nil;
    self.text = nil;
    [super dealloc];
}

- (NSString *)description
{
    NSMutableString* desc = [[[NSMutableString alloc] initWithString:[super description]] autorelease];
    [desc appendFormat:@"\nName: %@\nAttributes: %@\nText: %@\nChildren: %d item(s)\n", 
     self.name, 
     [self.attributes count] > 0 ? [self.attributes description] : @"nil", 
     [self.text length] > 0 ? self.text : @"nil", 
     [self.children count]];
    return desc;
}

- (TJXMLElement*) childWithName:(NSString*)name
{
    for (TJXMLElement* child in self.children) {
        if ([name isEqualToString:child.name])
            return child;
    }
    return nil;
}

- (NSArray*) childrenWithName:(NSString*)name
{
    NSMutableArray* results = [NSMutableArray array];
    for (TJXMLElement* child in self.children) {
        if ([name isEqualToString:child.name])
            [results addObject:child];
    }
    return results;
}

@end

@interface TJXMLParser ()

- (void) construct;

@end

@implementation TJXMLParser

@synthesize currentElement = _currentElement;
@synthesize rootElement = _rootElement;
@synthesize parseError = _parseError;

#pragma mark Init & dealloc

- (id) initWithContentsOfURL:(NSURL *)url
{
    self = [super init];
    if (self != nil) {
        _parser = [[NSXMLParser alloc] initWithContentsOfURL:url];
        [self construct];
    }
    return self;
}

- (id) initWithData:(NSData *)data
{
    self = [super init];
    if (self != nil) {
        _parser = [[NSXMLParser alloc] initWithData:data];
        [self construct];
    }
    return self;
}

- (void) construct
{
    _parser.delegate = self;
    _parentList = [[NSMutableArray alloc] init];
}

- (void) dealloc
{
    [_parser release]; _parser = nil;
    [_parentList release]; _parentList = nil;
    self.currentElement = nil;
    self.rootElement = nil;
    self.parseError = nil;
    [super dealloc];
}

#pragma mark Parsing

- (TJXMLElement*) parse
{
    BOOL result = [_parser parse];
    if (result) {
        return [[self.rootElement retain] autorelease];
    } else {
        return nil;
    }
}

#pragma mark NSXMLParserDelegate implementation

- (void)parserDidStartDocument:(NSXMLParser *)parser
{
    self.currentElement = nil;
    self.rootElement = nil;
    self.parseError = nil;
}

- (void)parser:(NSXMLParser *)parser 
didStartElement:(NSString *)elementName 
  namespaceURI:(NSString *)namespaceURI 
 qualifiedName:(NSString *)qualifiedName 
    attributes:(NSDictionary *)attributeDict 
{
    TJXMLElement* newElement = [[[TJXMLElement alloc] initWithName:elementName withAttributes:attributeDict] autorelease];
    if (self.currentElement) {
        TJXMLElement* parentElement = self.currentElement;
        [_parentList addObject:parentElement];
        [parentElement.children addObject:newElement];
    } else {
        self.rootElement = newElement;
    }
    self.currentElement = newElement;
}

- (void)parser:(NSXMLParser *)parser foundCharacters:(NSString *)string
{
    [self.currentElement.text appendString:string];
}

- (void)parser:(NSXMLParser *)parser 
 didEndElement:(NSString *)elementName 
  namespaceURI:(NSString *)namespaceURI 
 qualifiedName:(NSString *)qName
{
    if (![elementName isEqualToString:self.currentElement.name]) {
        // this is just to be sure, the parser would report an error before
        [parser abortParsing];
    }
    
    self.currentElement = [_parentList lastObject];
    [_parentList removeLastObject];
}

- (void)parserDidEndDocument:(NSXMLParser *)parser
{
    if (self.currentElement != nil || [_parentList count] > 0) {
        // this is just to be sure, the parser would report an error before
        [parser abortParsing];
    }
}

- (void)parser:(NSXMLParser *)parser parseErrorOccurred:(NSError *)parseError
{
    self.parseError = parseError;
    self.rootElement = nil;
}

@end