parsing HTML on the iPhone [closed]

Can anyone recommend a C or Objective-C library for HTML parsing? It needs to handle messy HTML code that won't quite validate.

Does such a library exist, or am I better off just trying to use regular expressions?

I found using hpple quite useful to parse messy HTML. Hpple project is a Objective-C wrapper on the XPathQuery library for parsing HTML. Using it you can send an XPath query and receive the result .

Requirements:

-Add libxml2 includes to your project

1. Menu Project->Edit Project Settings
2. Search for setting "Header Search Paths"
3. Add a new search path "${SDKROOT}/usr/include/libxml2"
4. Enable recursive option

-Add libxml2 library to to your project

1. Menu Project->Edit Project Settings
2. Search for setting "Other Linker Flags"
3. Add a new search flag "-lxml2"

-From hpple get the following source code files an add them to your project:

1. TFpple.h
2. TFpple.m
3. TFppleElement.h
4. TFppleElement.m
5. XPathQuery.h
6. XPathQuery.m

-Take a walk on w3school XPath Tutorial to feel comfortable with the XPath language.

Code Example

#import "TFHpple.h"

NSData *data = [[NSData alloc] initWithContentsOfFile:@"example.html"];

// Create parser
xpathParser = [[TFHpple alloc] initWithHTMLData:data];

//Get all the cells of the 2nd row of the 3rd table 
NSArray *elements  = [xpathParser searchWithXPathQuery:@"//table[3]/tr[2]/td"];

// Access the first cell
TFHppleElement *element = [elements objectAtIndex:0];

// Get the text within the cell tag
NSString *content = [element content];  

[xpathParser release];
[data release];

Known issues

As hpple is a wrapper over XPathQuery which is another wrapper, this option probably is not the most efficient. If performance is an issue in your project, I recommend to code your own lightweight solution based on hpple and xpathquery library code.

Looks like libxml2.2 comes in the SDK, and libxml/HTMLparser.h claims the following:

This module implements an HTML 4.0 non-verifying parser with API compatible with the XML parser ones. It should be able to parse "real world" HTML, even if severely broken from a specification point of view.

That sounds like what I need, so I'm probably going to use that.

Just in case anyone has got here by googling for a nice XPath parser and gone off and used TFHpple, Note that TFHpple uses XPathQuery. This is pretty good, but has a memory leak.

In the function *PerformXPathQuery, if the nodes are found to be nil, it jumps out before cleaning up.

So where you see this bit of code: Add in the two cleanup lines.

  xmlNodeSetPtr nodes = xpathObj->nodesetval;
  if (!nodes)
    {
      NSLog(@"Nodes was nil.");
        /* Cleanup */
        xmlXPathFreeObject(xpathObj);
        xmlXPathFreeContext(xpathCtx);
      return nil;
    }

If you are doing a LOT of parsing, it's a vicious leak. Now.... how do I get my night back :-)

I wrote a lightweight wrapper around libxml which maybe useful:

Objective-C-HMTL-Parser