Need ways to speed up PHP methods

Question

I ended up creating my own HTML class, mainly to keep all our output consistent. The code is below.

<?php

/**
 * A class to generate html tags
 * @author Glen Solsberry
 */
class HTML {

    private $isOpen;
    /**
     * holds all information about tags
     * @var $tags array
     */
    private $tags;
    private $current_tag = 0;

    private $depth = 0;
    private $output = "";

    private $separator = " ";
    private $pretty_print = true;

    /**
     * Set the pretty_print status
     * @author Glen Solsberry
     */
    public function setPretty($new_value) {
        $this->pretty_print = (bool)$new_value;
    }

    /**
     * Set the "separator" (the string that will be printed before tags on a new line)
     * @author Glen Solsberry
     */
    public function setSeparator($new_value) {
        $this->separator = $new_value;
    }

    /**
     * add a tag to the "DOM"
     * @author Glen Solsberry
     */
    public function tag($tag) {
        $this->updateChildrenAndParent();

        $this->tags[$this->current_tag]['tag'] = $tag;
        $this->tags[$this->current_tag]['open'] = true;

        return $this;
    }

    /**
     * set an attribute on the current tag
     * @author Glen Solsberry
     */
    public function attr($key, $value) {
        $this->tags[$this->current_tag]['attrs'][$key] = $value;

        return $this;
    }

    /**
     * close the current tag, if it's open.  if the tag is already closed, we work back up the chain to determine which tag needs to be closed.
     * @author Glen Solsberry
     */
    public function close() {
        if ($this->isOpen($this->current_tag)) {
            $this->tags[$this->current_tag]['open'] = false;
        } else {
            // work backwards till we find the first open tag, and close it
            for($previous_id = $this->current_tag - 1; $previous_id >= 0; $previous_id--) {
                if ($this->isOpen($previous_id)) {
                    $this->tags[$previous_id]['open'] = false;
                    break;
                }
            }
        }

        return $this;
    }

    /**
     * Create a text node
     * @author Glen Solsberry
     */
    public function text($text) {
        $this->updateChildrenAndParent();

        $this->tags[$this->current_tag]['text'] = $text;
        $this->tags[$this->current_tag]['open'] = false;

        return $this;
    }

    /**
     * Updates children and parent information, so that all tags link properly.
     * Duplicates will be handled later.
     * @author Glen Solsberry
     */
    private function updateChildrenAndParent() {
        // if the current tag is still open, then this needs to be a child of it
        if ($this->isOpen($this->current_tag)) {
            $this->tags[$this->current_tag]['children'][] = count($this->tags);
        }
        $this->current_tag = count($this->tags);
        // default value for this tags parent
        $this->tags[$this->current_tag]['parent'] = $this->current_tag - 1;

        for($parent_id = $this->current_tag - 1; $parent_id >= 0; $parent_id--) {
            // is the parent still open?
            if ($this->isOpen($parent_id)) {
                $this->tags[$this->current_tag]['parent'] = $parent_id;
                $this->tags[$parent_id]['children'][] = $this->current_tag;
                break;
            }
        }
    }

    /**
     * Determines whether the passed tag_id is open
     * @author Glen Solsberry
     */
    private function isOpen($tag_id) {
        return (bool)$this->tags[$tag_id]['open'];
    }

    /**
     * Determines whether the passed tag_id is closed
     * @author Glen Solsberry
     */
    private function isClosed($tag_id) {
        return (bool)!$this->tags[$tag_id]['open'];
    }

    /**
     * Generates a single tag's html represenation.  Called recursively if there are children
     * @author Glen Solsberry
     */
    private function generateTag($tag_id) {
        $current_tag = $this->tags[$tag_id];

        if (isset($current_tag['tag'])) {
            $this->output .= "<" . $current_tag['tag'];
            if (isset($current_tag['attrs'])) {
                ksort($current_tag['attrs']);
                foreach($current_tag['attrs'] as $key => $value) {
                    $this->output .= " " . $key . "=\"" . htmlspecialchars($value) . "\"";
                }
            }
            $this->output .= $this->pretty_print(">");

            if (isset($current_tag['children']) && count($current_tag['children']) > 0) {
                $children = array_unique($current_tag['children']);
                foreach($children as $position => $child_id) {
                    $this->depth++;
                    $this->generateTag($child_id);
                    $this->depth--;
                }
                $this->output .= $this->pretty_print("</" . $current_tag['tag'] . ">");
            } else {
                $this->output .= $this->pretty_print("</" . $current_tag['tag'] . ">");
                $this->depth--;
            }
        } else if (isset($current_tag['text'])) {
            $this->output .= $current_tag['text'];
        }
    }

    /**
     * Pretty prints the output.  Uses newlines and line starters.
     * @author Glen Solsberry
     */
    private function pretty_print($string) {
        $output = "";
        if ($this->pretty_print == true && $this->depth > 0) {
            $output .= str_repeat($this->separator, $this->depth);
        }
        $output .= $string;
        if ($this->pretty_print == true) {
            $output .= "\n";
        }
        return $output;
    }

    function __toString() {
        $this->generateTag(0);
        return $this->output;
    }

}

?>

I built some tests to make sure everything comes out properly (and to be a good PHP dev)...

<?php
require_once 'PHPUnit/Framework.php';
require_once 'PHPUnit/Extensions/OutputTestCase.php';
require_once '../html.class.inc';

class HTMLTest extends PHPUnit_Extensions_OutputTestCase
{
    public function testOutputLooksCorrect1() {
        $html = $this->sharedFixture;

        $html->tag("html");
        $html->close();

        $this->expectOutputString('<html></html>');
        print $html;
    }

    public function testOutputLooksCorrect2() {
        $html = $this->sharedFixture;

        $html->tag("html");
            $html->tag("head");
            $html->close();
        $html->close();

        $this->expectOutputString('<html><head></head></html>');
        print $html;
    }

    public function testOutputLooksCorrect3() {
        $html = $this->sharedFixture;

        $html->tag("html");
            $html->tag("head");
                $html->tag("link")->attr("rel", "stylesheet")->attr("href", "testOutputLooksCorrect.css")->attr("type", "text/css")->close();
            $html->close();
        $html->close();

        $this->expectOutputString('<html><head><link href="testOutputLooksCorrect.css" rel="stylesheet" type="text/css"></link></head></html>');
        print $html;
    }

    public function testOutputLooksCorrect4() {
        $html = $this->sharedFixture;

        $html->tag("html");
            $html->tag("head");
                $html->tag("link")->attr("href", "testOutputLooksCorrect.css")->attr("rel", "stylesheet")->attr("type", "text/css")->close();
            $html->close();
        $html->close();

        $this->expectOutputString('<html><head><link href="testOutputLooksCorrect.css" rel="stylesheet" type="text/css"></link></head></html>');
        print $html;
    }

    public function testOutputLooksCorrect5() {
        $html = $this->sharedFixture;

        $html->tag("html");
            $html->tag("head");
                $html->tag("link")->attr("type", "text/css")->attr("href", "testOutputLooksCorrect.css")->attr("rel", "stylesheet")->close();
            $html->close();
        $html->close();

        $this->expectOutputString('<html><head><link href="testOutputLooksCorrect.css" rel="stylesheet" type="text/css"></link></head></html>');
        print $html;
    }

    public function testOutputLooksCorrect10() {
        $html = $this->sharedFixture;

        $html->tag("html");
            $html->tag("head");
                $html->tag("link")->attr("rel", "stylesheet")->attr("href", "testOutputLooksCorrect.css")->attr("type", "text/css")->close();
            $html->close();
            $html->tag("body");
            $html->close();
        $html->close();

        $this->expectOutputString('<html><head><link href="testOutputLooksCorrect.css" rel="stylesheet" type="text/css"></link></head><body></body></html>');
        print $html;
    }

    public function testOutputLooksCorrect11() {
        $html = $this->sharedFixture;

        $html->tag("html");
            $html->tag("head");
                $html->tag("link")->attr("rel", "stylesheet")->attr("href", "testOutputLooksCorrect.css")->attr("type", "text/css")->close();
            $html->close();
            $html->tag("body");
                $html->text("This is a testOutputLooksCorrect.");
            $html->close();
        $html->close();

        $this->expectOutputString('<html><head><link href="testOutputLooksCorrect.css" rel="stylesheet" type="text/css"></link></head><body>This is a testOutputLooksCorrect.</body></html>');
        print $html;
    }

    public function testOutputLooksCorrect12() {
        $html = $this->sharedFixture;

        $html->text("This is a testOutputLooksCorrect.");

        $this->expectOutputString('This is a testOutputLooksCorrect.');
        print $html;
    }

    public function testOutputLooksCorrect13() {
        $html = $this->sharedFixture;

        $html->tag("head")->close();

        $this->expectOutputString('<head></head>');
        print $html;
    }

    public function testOutputLooksCorrect14() {
        $html = $this->sharedFixture;

        $html->tag("head")->tag("title")->text("This is the title")->close()->close();

        $this->expectOutputString('<head><title>This is the title</title></head>');
        print $html;
    }

    public function testOutputLooksCorrect15() {
        $html = $this->sharedFixture;

        $html->tag("html");
            $html->tag("head");
                $html->tag("link")->attr("rel", "stylesheet")->attr("href", "testOutputLooksCorrect.css")->attr("type", "text/css")->close();
                $html->tag("meta")->attr("http-equiv", "Content-Type")->attr("content", "text/html; charset=utf-8")->close();
            $html->close();
            $html->tag("body");
                $html->text("This is a testOutputLooksCorrect.");
            $html->close();
        $html->close();
        $this->expectOutputString('<html><head><link href="testOutputLooksCorrect.css" rel="stylesheet" type="text/css"></link><meta content="text/html; charset=utf-8" http-equiv="Content-Type"></meta></head><body>This is a testOutputLooksCorrect.</body></html>');
        print $html;
    }

    public function testOutputLooksCorrect16() {
        $html = $this->sharedFixture;

        $html->tag("html");
            $html->tag("head");
                $html->tag("link")->attr("rel", "stylesheet")->attr("href", "testOutputLooksCorrect.css")->attr("type", "text/css")->close();
                $html->tag("meta")->attr("http-equiv", "Content-Type")->attr("content", "text/html; charset=utf-8")->close();
                $html->tag("script")->attr("type", "javascript")->attr("src", "blah.js")->close();
            $html->close();
            $html->tag("body");
                $html->text("This is a testOutputLooksCorrect.");
            $html->close();
        $html->close();

        $this->expectOutputString('<html><head><link href="testOutputLooksCorrect.css" rel="stylesheet" type="text/css"></link><meta content="text/html; charset=utf-8" http-equiv="Content-Type"></meta><script src="blah.js" type="javascript"></script></head><body>This is a testOutputLooksCorrect.</body></html>');
        print $html;
    }

    protected function setUp() {
        $html = new HTML;
        $html->setPretty(0);
        $this->sharedFixture = $html;
    }

    protected function tearDown() {
        $this->sharedFixture = NULL;
    }
}
?>

Everything works properly. However, the time needed is much higher than I'd like. On the order of ~3 seconds for 14k worth of HTML.

After profiling the code, it seems that most of the time is spent in the updateChildrenAndParent. Almost 75% of the entire execution is spent there. Can anyone suggest ways to speed this up, since I don't want to slow down the site noticably.

Answer 1

You should be using real dom methods to parse HTML. They are written in C and will be orders of magnitude faster than anything you can code up in naively.