HtmlConverter::replaceMisplacedComments() - Code Metrics - Inspection of "#212: Find and move any misplaced comment nodes" - thephpleague/html-to-markdown - Measure and Improve Code Quality continuously with Scrutinizer

Passed

Pull Request — master (#214)

by Colin

created 2021-11-06 05:30 UTC

HtmlConverter::replaceMisplacedComments() A

↳ Parent: HtmlConverter

Complexity

Conditions	2
Paths	2

Size

Total Lines	10
Code Lines	4

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	4
CRAP Score	2

Importance

Changes

Metric	Value
cc	2
eloc	4
nc	2
nop	1
dl	0
loc	10
ccs	4
cts	4
cp	1
crap	2
rs	10
c	0
b	0
f	0

<?php

declare(strict_types=1);

namespace League\HTMLToMarkdown;

/**
 * A helper class to convert HTML to Markdown.
 *
 * @author Colin O'Dell <[email protected]>
 * @author Nick Cernis <[email protected]>
 *
 * @link https://github.com/thephpleague/html-to-markdown/ Latest version on GitHub.
 *
 * @license http://www.opensource.org/licenses/mit-license.php MIT
 */
class HtmlConverter implements HtmlConverterInterface
{
    /** @var Environment */
    protected $environment;

    /**
     * Constructor
     *
     * @param Environment|array<string, mixed> $options Environment object or configuration options
     */
    public function __construct($options = [])
    {
        if ($options instanceof Environment) {
            $this->environment = $options;
        } elseif (\is_array($options)) {

            $defaults = [
                'header_style' => 'setext', // Set to 'atx' to output H1 and H2 headers as # Header1 and ## Header2
                'suppress_errors' => true, // Set to false to show warnings when loading malformed HTML
                'strip_tags' => false, // Set to true to strip tags that don't have markdown equivalents. N.B. Strips tags, not their content. Useful to clean MS Word HTML output.
                'strip_placeholder_links' => false, // Set to true to remove <a> that doesn't have href.
                'bold_style' => '**', // DEPRECATED: Set to '__' if you prefer the underlined style
                'italic_style' => '*', // DEPRECATED: Set to '_' if you prefer the underlined style
                'remove_nodes' => '', // space-separated list of dom nodes that should be removed. example: 'meta style script'
                'hard_break' => false, // Set to true to turn <br> into `\n` instead of `  \n`
                'list_item_style' => '-', // Set the default character for each <li> in a <ul>. Can be '-', '*', or '+'
                'preserve_comments' => false, // Set to true to preserve comments, or set to an array of strings to preserve specific comments
                'use_autolinks' => true, // Set to true to use simple link syntax if possible. Will always use []() if set to false
                'table_pipe_escape' => '\|', // Replacement string for pipe characters inside markdown table cells
                'table_caption_side' => 'top', // Set to 'top' or 'bottom' to show <caption> content before or after table, null to suppress
            ];

            $this->environment = Environment::createDefaultEnvironment($defaults);

            $this->environment->getConfig()->merge($options);
        }
    }

    public function getEnvironment(): Environment
    {
        return $this->environment;
    }

    public function getConfig(): Configuration
    {
        return $this->environment->getConfig();
    }

    /**
     * Convert
     *
     * @see HtmlConverter::convert
     *
     * @return string The Markdown version of the html
     */
    public function __invoke(string $html): string
    {
        return $this->convert($html);
    }

    /**
     * Convert
     *
     * Loads HTML and passes to getMarkdown()
     *
     * @return string The Markdown version of the html
     *
     * @throws \InvalidArgumentException|\RuntimeException
     */
    public function convert(string $html): string
    {
        if (\trim($html) === '') {
            return '';
        }

        $document = $this->createDOMDocument($html);

        // Work on the entire DOM tree (including head and body)
        if (! ($root = $document->getElementsByTagName('html')->item(0))) {
            throw new \InvalidArgumentException('Invalid HTML was provided');
        }

        $rootElement = new Element($root);
        $this->convertChildren($rootElement);

        // Store the now-modified DOMDocument as a string
        $markdown = $document->saveHTML();

        if ($markdown === false) {
            throw new \RuntimeException('Unknown error occurred during HTML to Markdown conversion');
        }

        return $this->sanitize($markdown);
    }

    private function createDOMDocument(string $html): \DOMDocument
    {
        $document = new \DOMDocument();

        if ($this->getConfig()->getOption('suppress_errors')) {
            // Suppress conversion errors (from http://bit.ly/pCCRSX)
            \libxml_use_internal_errors(true);
        }

        // Hack to load utf-8 HTML (from http://bit.ly/pVDyCt)
        $document->loadHTML('<?xml encoding="UTF-8">' . $html);
        $document->encoding = 'UTF-8';

        $this->replaceMisplacedComments($document);

        if ($this->getConfig()->getOption('suppress_errors')) {
            \libxml_clear_errors();
        }

        return $document;
    }

    /**
     * Finds any comment nodes outside <html> element and moves them into <body>.
     *
     * @see https://github.com/thephpleague/html-to-markdown/issues/212
     * @see https://3v4l.org/7bC33
     */
    private function replaceMisplacedComments(\DOMDocument $document): void
    {
        // Find ny comment nodes at the root of the document.
        $misplacedComments = (new \DOMXPath($document))->query('/comment()');
        $body              = $document->getElementsByTagName('body')->item(0);

        // Loop over comment nodes in reverse so we put them inside <body> in
        // their original order.
        for ($index = $misplacedComments->length - 1; $index >= 0; $index--) {
            $body->insertBefore($misplacedComments[$index], $body->firstChild);
        }
    }

    /**
     * Convert Children
     *
     * Recursive function to drill into the DOM and convert each node into Markdown from the inside out.
     *
     * Finds children of each node and convert those to #text nodes containing their Markdown equivalent,
     * starting with the innermost element and working up to the outermost element.
     */
    private function convertChildren(ElementInterface $element): void
    {
        // Don't convert HTML code inside <code> and <pre> blocks to Markdown - that should stay as HTML
        // except if the current node is a code tag, which needs to be converted by the CodeConverter.
        if ($element->isDescendantOf(['pre', 'code']) && $element->getTagName() !== 'code') {
            return;
        }

        // Give converter a chance to inspect/modify the DOM before children are converted
        $converter = $this->environment->getConverterByTag($element->getTagName());
        if ($converter instanceof PreConverterInterface) {
            $converter->preConvert($element);
        }

        // If the node has children, convert those to Markdown first
        if ($element->hasChildren()) {
            foreach ($element->getChildren() as $child) {
                $this->convertChildren($child);
            }
        }

        // Now that child nodes have been converted, convert the original node
        $markdown = $this->convertToMarkdown($element);

        // Create a DOM text node containing the Markdown equivalent of the original node

        // Replace the old $node e.g. '<h3>Title</h3>' with the new $markdown_node e.g. '### Title'
        $element->setFinalMarkdown($markdown);
    }

    /**
     * Convert to Markdown
     *
     * Converts an individual node into a #text node containing a string of its Markdown equivalent.
     *
     * Example: An <h3> node with text content of 'Title' becomes a text node with content of '### Title'
     *
     * @return string The converted HTML as Markdown
     */
    protected function convertToMarkdown(ElementInterface $element): string
    {
        $tag = $element->getTagName();

        // Strip nodes named in remove_nodes
        $tagsToRemove = \explode(' ', $this->getConfig()->getOption('remove_nodes') ?? '');
        if (\in_array($tag, $tagsToRemove, true)) {
            return '';
        }

        $converter = $this->environment->getConverterByTag($tag);

        return $converter->convert($element);
    }

    protected function sanitize(string $markdown): string
    {
        $markdown = \html_entity_decode($markdown, ENT_QUOTES, 'UTF-8');
        $markdown = \preg_replace('/<!DOCTYPE [^>]+>/', '', $markdown); // Strip doctype declaration
        \assert($markdown !== null);
        $markdown = \trim($markdown); // Remove blank spaces at the beggining of the html

        /*
         * Removing unwanted tags. Tags should be added to the array in the order they are expected.
         * XML, html and body opening tags should be in that order. Same case with closing tags
         */
        $unwanted = ['<?xml encoding="UTF-8">', '<html>', '</html>', '<body>', '</body>', '<head>', '</head>', '&#xD;'];

        foreach ($unwanted as $tag) {
            if (\strpos($tag, '/') === false) {
                // Opening tags
                if (\strpos($markdown, $tag) === 0) {
                    $markdown = \substr($markdown, \strlen($tag));
                }
            } else {
                // Closing tags
                if (\strpos($markdown, $tag) === \strlen($markdown) - \strlen($tag)) {
                    $markdown = \substr($markdown, 0, -\strlen($tag));
                }
            }
        }

        return \trim($markdown, "\n\r\0\x0B");
    }

    /**
     * Pass a series of key-value pairs in an array; these will be passed
     * through the config and set.
     * The advantage of this is that it can allow for static use (IE in Laravel).
     * An example being:
     *
     * HtmlConverter::setOptions(['strip_tags' => true])->convert('<h1>test</h1>');
     *
     * @param array<string, mixed> $options
     *
     * @return $this
     */
    public function setOptions(array $options)
    {
        $config = $this->getConfig();

        foreach ($options as $key => $option) {
            $config->setOption($key, $option);
        }

        return $this;
    }
}


1		<?php
2
3		declare(strict_types=1);
4
5		namespace League\HTMLToMarkdown;
6
7		/**
8		* A helper class to convert HTML to Markdown.
9		*
10		* @author Colin O'Dell <[email protected]>
11		* @author Nick Cernis <[email protected]>
12		*
13		* @link https://github.com/thephpleague/html-to-markdown/ Latest version on GitHub.
14		*
15		* @license http://www.opensource.org/licenses/mit-license.php MIT
16		*/
17		class HtmlConverter implements HtmlConverterInterface
18		{
19		/** @var Environment */
20		protected $environment;
21
22		/**
23		* Constructor
24		*
25		* @param Environment\|array<string, mixed> $options Environment object or configuration options
26		*/
27		public function __construct($options = [])
28		{
29		if ($options instanceof Environment) {
30		$this->environment = $options;
31	99	} elseif (\is_array($options)) {
		0 ignored issues – show introduced 2021-03-29 01:54 UTC by Report Bug Copy Issue Report The condition `is_array($options)` is always `true`. Loading history...
32		$defaults = [
33	99	'header_style' => 'setext', // Set to 'atx' to output H1 and H2 headers as # Header1 and ## Header2
34	3	'suppress_errors' => true, // Set to false to show warnings when loading malformed HTML
35	98	'strip_tags' => false, // Set to true to strip tags that don't have markdown equivalents. N.B. Strips tags, not their content. Useful to clean MS Word HTML output.
36		'strip_placeholder_links' => false, // Set to true to remove <a> that doesn't have href.
37	96	'bold_style' => '**', // DEPRECATED: Set to '__' if you prefer the underlined style
38	64	'italic_style' => '*', // DEPRECATED: Set to '_' if you prefer the underlined style
39	64	'remove_nodes' => '', // space-separated list of dom nodes that should be removed. example: 'meta style script'
40	64	'hard_break' => false, // Set to true to turn <br> into `\n` instead of ` \n`
41	64	'list_item_style' => '-', // Set the default character for each <li> in a <ul>. Can be '-', '*', or '+'
42	64	'preserve_comments' => false, // Set to true to preserve comments, or set to an array of strings to preserve specific comments
43	64	'use_autolinks' => true, // Set to true to use simple link syntax if possible. Will always use []() if set to false
44	64	'table_pipe_escape' => '\\|', // Replacement string for pipe characters inside markdown table cells
45	64	'table_caption_side' => 'top', // Set to 'top' or 'bottom' to show <caption> content before or after table, null to suppress
46	64	];
47	64
48	64	$this->environment = Environment::createDefaultEnvironment($defaults);
49	64
50	64	$this->environment->getConfig()->merge($options);
51		}
52	96	}
53
54	96	public function getEnvironment(): Environment
55	64	{
56	99	return $this->environment;
57		}
58
59		public function getConfig(): Configuration
60		{
61	3	return $this->environment->getConfig();
62		}
63	3
64		/**
65		* Convert
66		*
67		* @see HtmlConverter::convert
68		*
69	96	* @return string The Markdown version of the html
70		*/
71	96	public function __invoke(string $html): string
72		{
73		return $this->convert($html);
74		}
75
76		/**
77		* Convert
78		*
79		* Loads HTML and passes to getMarkdown()
80		*
81		* @return string The Markdown version of the html
82		*
83	3	* @throws \InvalidArgumentException\|\RuntimeException
84		*/
85	3	public function convert(string $html): string
86		{
87		if (\trim($html) === '') {
88		return '';
89		}
90
91		$document = $this->createDOMDocument($html);
92
93		// Work on the entire DOM tree (including head and body)
94		if (! ($root = $document->getElementsByTagName('html')->item(0))) {
95		throw new \InvalidArgumentException('Invalid HTML was provided');
96		}
97
98		$rootElement = new Element($root);
99	99	$this->convertChildren($rootElement);
100
101	99	// Store the now-modified DOMDocument as a string
102	3	$markdown = $document->saveHTML();
103
104		if ($markdown === false) {
105	96	throw new \RuntimeException('Unknown error occurred during HTML to Markdown conversion');
106		}
107
108	96	return $this->sanitize($markdown);
109		}
110
111		private function createDOMDocument(string $html): \DOMDocument
112	96	{
113	96	$document = new \DOMDocument();
114
115		if ($this->getConfig()->getOption('suppress_errors')) {
116	96	// Suppress conversion errors (from http://bit.ly/pCCRSX)
117		\libxml_use_internal_errors(true);
118	96	}
119
120		// Hack to load utf-8 HTML (from http://bit.ly/pVDyCt)
121		$document->loadHTML('<?xml encoding="UTF-8">' . $html);
122		$document->encoding = 'UTF-8';
123
124		$this->replaceMisplacedComments($document);
125
126	96	if ($this->getConfig()->getOption('suppress_errors')) {
127		\libxml_clear_errors();
128	96	}
129
130	96	return $document;
131		}
132	93
133	62	/**
134		* Finds any comment nodes outside <html> element and moves them into <body>.
135		*
136	96	* @see https://github.com/thephpleague/html-to-markdown/issues/212
137	96	* @see https://3v4l.org/7bC33
138		*/
139	96	private function replaceMisplacedComments(\DOMDocument $document): void
140	93	{
141	62	// Find ny comment nodes at the root of the document.
142		$misplacedComments = (new \DOMXPath($document))->query('/comment()');
143	96	$body = $document->getElementsByTagName('body')->item(0);
144
145		// Loop over comment nodes in reverse so we put them inside <body> in
146		// their original order.
147		for ($index = $misplacedComments->length - 1; $index >= 0; $index--) {
148		$body->insertBefore($misplacedComments[$index], $body->firstChild);
149		}
150		}
151
152		/**
153		* Convert Children
154		*
155		* Recursive function to drill into the DOM and convert each node into Markdown from the inside out.
156	96	*
157		* Finds children of each node and convert those to #text nodes containing their Markdown equivalent,
158		* starting with the innermost element and working up to the outermost element.
159		*/
160	96	private function convertChildren(ElementInterface $element): void
161	21	{
162		// Don't convert HTML code inside <code> and <pre> blocks to Markdown - that should stay as HTML
163		// except if the current node is a code tag, which needs to be converted by the CodeConverter.
164		if ($element->isDescendantOf(['pre', 'code']) && $element->getTagName() !== 'code') {
165	96	return;
166	96	}
167	3
168	2	// Give converter a chance to inspect/modify the DOM before children are converted
169		$converter = $this->environment->getConverterByTag($element->getTagName());
170		if ($converter instanceof PreConverterInterface) {
171	96	$converter->preConvert($element);
172	96	}
173	96
174	64	// If the node has children, convert those to Markdown first
175	64	if ($element->hasChildren()) {
176		foreach ($element->getChildren() as $child) {
177		$this->convertChildren($child);
178	96	}
179		}
180
181		// Now that child nodes have been converted, convert the original node
182		$markdown = $this->convertToMarkdown($element);
183	96
184	96	// Create a DOM text node containing the Markdown equivalent of the original node
185
186		// Replace the old $node e.g. '<h3>Title</h3>' with the new $markdown_node e.g. '### Title'
187		$element->setFinalMarkdown($markdown);
188		}
189
190		/**
191		* Convert to Markdown
192		*
193		* Converts an individual node into a #text node containing a string of its Markdown equivalent.
194		*
195		* Example: An <h3> node with text content of 'Title' becomes a text node with content of '### Title'
196		*
197	96	* @return string The converted HTML as Markdown
198		*/
199	96	protected function convertToMarkdown(ElementInterface $element): string
200		{
201		$tag = $element->getTagName();
202	96
203	96	// Strip nodes named in remove_nodes
204	3	$tagsToRemove = \explode(' ', $this->getConfig()->getOption('remove_nodes') ?? '');
205		if (\in_array($tag, $tagsToRemove, true)) {
206		return '';
207	96	}
208
209	96	$converter = $this->environment->getConverterByTag($tag);
210
211		return $converter->convert($element);
212		}
213
214		protected function sanitize(string $markdown): string
215		{
216		$markdown = \html_entity_decode($markdown, ENT_QUOTES, 'UTF-8');
217	96	$markdown = \preg_replace('/<!DOCTYPE [^>]+>/', '', $markdown); // Strip doctype declaration
218		\assert($markdown !== null);
219	96	$markdown = \trim($markdown); // Remove blank spaces at the beggining of the html
220	96
221	96	/*
222		* Removing unwanted tags. Tags should be added to the array in the order they are expected.
223		* XML, html and body opening tags should be in that order. Same case with closing tags
224		*/
225		$unwanted = ['<?xml encoding="UTF-8">', '<html>', '</html>', '<body>', '</body>', '<head>', '</head>', ' '];
226
227	96	foreach ($unwanted as $tag) {
228		if (\strpos($tag, '/') === false) {
229	96	// Opening tags
230	96	if (\strpos($markdown, $tag) === 0) {
231		$markdown = \substr($markdown, \strlen($tag));
232	96	}
233	96	} else {
234	64	// Closing tags
235	64	if (\strpos($markdown, $tag) === \strlen($markdown) - \strlen($tag)) {
236		$markdown = \substr($markdown, 0, -\strlen($tag));
237	96	}
238	90	}
239	58	}
240
241	64	return \trim($markdown, "\n\r\0\x0B");
242		}
243	96
244		/**
245		* Pass a series of key-value pairs in an array; these will be passed
246		* through the config and set.
247		* The advantage of this is that it can allow for static use (IE in Laravel).
248		* An example being:
249		*
250		* HtmlConverter::setOptions(['strip_tags' => true])->convert('<h1>test</h1>');
251		*
252		* @param array<string, mixed> $options
253		*
254		* @return $this
255		*/
256		public function setOptions(array $options)
257		{
258		$config = $this->getConfig();
259
260		foreach ($options as $key => $option) {
261		$config->setOption($key, $option);
262		}
263
264		return $this;
265		}
266		}
267

thephpleague / html-to-markdown

Pull Request — master (#214)

HtmlConverter::replaceMisplacedComments() A

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like