HTML 解析

xml,json都有大量的库来解析，我们如何解析html呢？

TFHpple是一个小型的封装，可以用来解析html，它是对libxml的封装，语法是xpath。
今天我看到一个直接用libxml来解析html，参看：http://www.cocoanetics.com/2011/09/taming-html-parsing-with-libxml-1/#comment-3090 那张图画得一目了然，很值得收藏。这个文章中的源码不能遍历所有的html，我做了一点修改可以将html遍历打印出来

// NSData data contains the document data 
// encoding is the NSStringEncoding of the data 
// baseURL the documents base URL, i.e. location

CFStringEncoding cfenc = CFStringConvertNSStringEncodingToEncoding(encoding); 
CFStringRef cfencstr = CFStringConvertEncodingToIANACharSetName(cfenc); 
const char *enc = CFStringGetCStringPtr(cfencstr, 0);

htmlDocPtr _htmlDocument = htmlReadDoc([data bytes], 
[[baseURL absoluteString] UTF8String], 
enc, 
XML_PARSE_NOERROR | XML_PARSE_NOWARNING); 
if (_htmlDocument) 
{ 
xmlFreeDoc(_htmlDocument); 
}

xmlNodePtr currentNode = (xmlNodePtr)_htmlDocument;

while (currentNode) 
{ 
// output node if it is an element

if (currentNode->type == XML_ELEMENT_NODE) 
{ 
NSMutableArray *attrArray = [NSMutableArray array];

for (xmlAttrPtr attrNode = currentNode->properties; attrNode; attrNode = attrNode->next) 
{ 
xmlNodePtr contents = attrNode->children;

[attrArray addObject:[NSString stringWithFormat:@"%s='%s'", attrNode->name, contents->content]]; 
}

NSString *attrString = [attrArray componentsJoinedByString:@" "];

if ([attrString length]) 
{ 
attrString = [@" " stringByAppendingString:attrString]; 
}

NSLog(@"<%s%@>", currentNode->name, attrString); 
} 
else if (currentNode->type == XML_TEXT_NODE) 
{ 
//NSLog(@"%s", currentNode->content); 
NSLog(@"%@", [NSString stringWithCString:(const char *)currentNode->content encoding:NSUTF8StringEncoding]); 
} 
else if (currentNode->type == XML_COMMENT_NODE) 
{ 
NSLog(@"/* %s */", currentNode->name); 
}

if (currentNode && currentNode->children) 
{ 
currentNode = currentNode->children; 
} 
else if (currentNode && currentNode->next) 
{ 
currentNode = currentNode->next; 
} 
else 
{ 
currentNode = currentNode->parent;

// close node 
if (currentNode && currentNode->type == XML_ELEMENT_NODE) 
{ 
NSLog(@"</%s>", currentNode->name); 
}

if (currentNode->next) 
{ 
currentNode = currentNode->next; 
} 
else 
{ 
while(currentNode) 
{ 
currentNode = currentNode->parent; 
if (currentNode && currentNode->type == XML_ELEMENT_NODE) 
{ 
NSLog(@"</%s>", currentNode->name); 
if (strcmp((const char *)currentNode->name, "table") == 0) 
{ 
NSLog(@"over"); 
} 
}

if (currentNode == nodes->nodeTab[0]) 
{ 
break; 
}

if (currentNode && currentNode->next) 
{ 
currentNode = currentNode->next; 
break; 
} 
} 
} 
}

if (currentNode == nodes->nodeTab[0]) 
{ 
break; 
} 
}

不过我还是喜欢用TFHpple，因为它很简单，也好用，但是它的功能不是很完完善。比如，不能获取children node,我就写了两个方法，一个是获取children node,一个是获取所有的contents. 还有node的属性content的key与node's content的key一样，都是@"nodeContent", 正确情况下属性的应是@"attributeContent"，
所以我写了这个方法，同时修改node属性的content key.
NSDictionary *DictionaryForNode2(xmlNodePtr currentNode, NSMutableDictionary *parentResult) 
{ 
NSMutableDictionary *resultForNode = [NSMutableDictionary dictionary];

if (currentNode->name) 
{ 
NSString *currentNodeContent = 
[NSString stringWithCString:(const char *)currentNode->name encoding:NSUTF8StringEncoding]; 
[resultForNode setObject:currentNodeContent forKey:@"nodeName"]; 
}

if (currentNode->content) 
{ 
NSString *currentNodeContent = [NSString stringWithCString:(const char *)currentNode->content encoding:NSUTF8StringEncoding];

if (currentNode->type == XML_TEXT_NODE) 
{ 
if (currentNode->parent->type == XML_ELEMENT_NODE) 
{ 
[parentResult setObject:currentNodeContent forKey:@"nodeContent"]; 
return nil; 
}

if (currentNode->parent->type == XML_ATTRIBUTE_NODE) 
{ 
[parentResult 
setObject: 
[currentNodeContent 
stringByTrimmingCharactersInSet:[NSCharacterSet whitespaceAndNewlineCharacterSet]] 
forKey:@"attributeContent"]; 
return nil;

} 
} 
}

xmlAttr *attribute = currentNode->properties; 
if (attribute) 
{ 
NSMutableArray *attributeArray = [NSMutableArray array]; 
while (attribute) 
{ 
NSMutableDictionary *attributeDictionary = [NSMutableDictionary dictionary]; 
NSString *attributeName = 
[NSString stringWithCString:(const char *)attribute->name encoding:NSUTF8StringEncoding]; 
if (attributeName) 
{ 
[attributeDictionary setObject:attributeName forKey:@"attributeName"]; 
}

if (attribute->children) 
{ 
NSDictionary *childDictionary = DictionaryForNode2(attribute->children, attributeDictionary); 
if (childDictionary) 
{ 
[attributeDictionary setObject:childDictionary forKey:@"attributeContent"]; 
} 
}

if ([attributeDictionary count] > 0) 
{ 
[attributeArray addObject:attributeDictionary]; 
} 
attribute = attribute->next; 
}

if ([attributeArray count] > 0) 
{ 
[resultForNode setObject:attributeArray forKey:@"nodeAttributeArray"]; 
} 
}

xmlNodePtr childNode = currentNode->children; 
if (childNode) 
{ 
NSMutableArray *childContentArray = [NSMutableArray array]; 
while (childNode) 
{ 
NSDictionary *childDictionary = DictionaryForNode2(childNode, resultForNode); 
if (childDictionary) 
{ 
[childContentArray addObject:childDictionary]; 
} 
childNode = childNode->next; 
} 
if ([childContentArray count] > 0) 
{ 
[resultForNode setObject:childContentArray forKey:@"nodeChildArray"]; 
} 
}

return resultForNode; 
}

TFHppleElement.m里加了两个key 常量
NSString * const TFHppleNodeAttributeContentKey = @"attributeContent"; 
NSString * const TFHppleNodeChildArrayKey = @"nodeChildArray";

并修改获取属性方法为：
- (NSDictionary *) attributes 
{ 
NSMutableDictionary * translatedAttributes = [NSMutableDictionary dictionary]; 
for (NSDictionary * attributeDict in [node objectForKey:TFHppleNodeAttributeArrayKey]) { 
[translatedAttributes setObject:[attributeDict objectForKey:TFHppleNodeAttributeContentKey] 
forKey:[attributeDict objectForKey:TFHppleNodeAttributeNameKey]]; 
} 
return translatedAttributes; 
}

并添加获取children node 方法：
- (BOOL) hasChildren 
{ 
NSArray *childs = [node objectForKey: TFHppleNodeChildArrayKey];

if (childs) 
{ 
return YES; 
}

return NO; 
}

- (NSArray *) children 
{ 
if ([self hasChildren]) 
return [node objectForKey: TFHppleNodeChildArrayKey]; 
return nil; 
}

参看：http://giles-wang.blogspot.com/2011/08/iphoneansi.html

原文：http://blog.csdn.net/favormm/article/details/6794487