Question

Possible Duplicate:
How to parse and process HTML with PHP?

I am looking for a php function that decreases html hx (h1, h2, h3, …, h6) tags by one.

• h1 becomes h2
• h2 becomes h3 and so on
• …
• h6 gets replaced by ‘ ‘

do you know such a function?

This is how I started stripping the h6 tags:

$string = preg_replace('#<(?:/)?\s*h6\s*>#', ' ', $string);

Answer 1

Here’s one for DOM, which iterates trough all mappings and then replace tags or copies children.

<?php

// New tag mappings:
//     null => extract childs and push them into parent contrainer
// Make sure that they are in this order, otherwise they would match wrongly
// between each another
$mapping = array(
    'h6' => null,
    'h5' => 'h6',
    'h4' => 'h5',
    'h3' => 'h4',
    'h2' => 'h3',
    'h1' => 'h2'
);

// Load document
$xml = new DOMDocument();
$xml->loadHTMLFile('http://stackoverflow.com/questions/12883009/php-code-to-decrease-html-hx-tags') or die('Failed to load');

$xPath = new DOMXPath( $xml);

foreach( $mapping as $original => $new){
    // Load nodes
    $nodes = $xPath->query( '//' . $original);

    // This is a critical error and should NEVER happen
    if( $nodes === false){
        die( 'Malformed expression: //' . $original);
    }

    echo $original . ' has nodes: ' . $nodes->length . "\n";

    // Process each node
    foreach( $nodes as $node){
        if( $new == null){
            // Append all the childs before self and remove self afterwards
            foreach( $node->childNodes as $child){
                $node->parentNode->insertBefore( $child->cloneNode( true), $node);
            }
            $node->parentNode->removeChild( $node);

        } else {
            // Create new empty node and push all childrens to it
            $newNode = $xml->createElement( $new);
            foreach( $node->childNodes as $child){
                $newNode->appendChild( $child);
            }
            $node->parentNode->replaceChild( $newNode, $node);
        }
    }
}

echo $xml->saveHTML();

You may also do some xPath optimalizations like using //* or //h3|//h2 and checking DOMElement::tagName, but I wanted this to be straight forward.

---

Edit: Solution that goes trough nodes only once and doesn’t care about order:

<?php
// The beginning (everything to the first foreach loop) remains the same
// Load nodes
$nodes = $xPath->query( '//*');

// This is a critical error and should NEVER happen
if( $nodes === false){
    die( 'Malformed expression: //' . $original);
}

// Process each node
foreach( $nodes as $node){
    // Check correct $node class
    if( !($node instanceof DOMElement)){
        continue;
    }
    $tagName = $node->tagName;  

    // Do we have a mapping?
    if( !array_key_exists( $tagName, $mapping)){
        continue;
    }
    $new = $mapping[$tagName];
    echo 'Has element: ' . $tagName . ' => ' . $new . "\n";

    if( $new == null){
        // Append all the childs before self and remove self afterwards
        foreach( $node->childNodes as $child){
            $node->parentNode->insertBefore( $child->cloneNode( true), $node);
        }
        $node->parentNode->removeChild( $node);

    } else {
        // Create new empty node and push all childrens to it
        $newNode = $xml->createElement( $new);
        foreach( $node->childNodes as $child){
            $newNode->appendChild( $child);
        }
        $node->parentNode->replaceChild( $newNode, $node);
    }
}

echo $xml->saveHTML();

And the last optimalization I can think about is using:

$xPathQuery = '//' . implode( array_keys($mapping), '|//');
$nodes = $xPath->query( $xPathQuery);