elkarte / Elkarte

sources/ElkArte/XmlArray.php

<?php

/**
 * The XmlArray class is an xml parser.
 *
 * @package   ElkArte Forum
 * @copyright ElkArte Forum contributors
 * @license   BSD http://opensource.org/licenses/BSD-3-Clause (see accompanying LICENSE.txt file)
 *
 * This file contains code covered by:
 * copyright: 2011 Simple Machines (http://www.simplemachines.org)
 *
 * @version 2.0 dev
 *
 */

namespace ElkArte;

/**
 * Class representing an xml array.
 *
 * Reads in xml, allows you to access it simply.
 * Version 2.0 dev.
 */
class XmlArray
{
	/** @var array Holds xml parsed results */
	public $array;

	/** @var int|null Holds debugging level */
	public $debug_level;

	/** @var bool Holds trim level textual data */
	public $trim;

	/**
	 * Constructor for the xml parser.
	 *
	 * Example use:
	 *   $xml = new \ElkArte\XmlArray(file('data.xml'));
	 *
	 * @param string $data the xml data or an array of, unless is_clone is true.
	 * @param bool $auto_trim default false, used to automatically trim textual data.
	 * @param int|null $level default null, the debug level, specifies whether notices should be generated for missing elements and attributes.
	 * @param bool $is_clone default false. If is_clone is true, the  \ElkArte\XmlArray is cloned from another - used internally only.
	 */
	public function __construct($data, $auto_trim = false, $level = null, $is_clone = false)
	{
		// If we're using this try to get some more memory.
		detectServer()->setMemoryLimit('128M');

		// Set the debug level.
		$this->debug_level = $level ?? error_reporting();
		$this->trim = $auto_trim;

		// Is the data already parsed?
		if ($is_clone)
		{
			$this->array = $data;

It seems like $data of type string is incompatible with the declared type array of property $array.

			return;
		}

		// Is the input an array? (ie. passed from file()?)
		if (is_array($data))

The condition is_array($data) is always false.

		{
			$data = implode('', $data);
		}

		// Remove any xml declaration or doctype, and parse out comments and CDATA.
		$data = preg_replace('/<!--.*?-->/s', '', $this->_to_cdata(preg_replace(['/^<\?xml.+?\?' . '>/is', '/<!DOCTYPE[^>]+?' . '>/'], '', $data)));

		// Now parse the xml!
		$this->array = $this->_parse($data);
	}

	/**
	 * Parse out CDATA tags. (htmlspecialchars them...)
	 *
	 * @param string $data The data with CDATA tags
	 *
	 * @return string
	 */
	protected function _to_cdata($data): string
	{
		$inCdata = false;
		$inComment = false;
		$output = '';

		$parts = preg_split('~(<!\[CDATA\[|\]\]>|<!--|-->)~', $data, -1, PREG_SPLIT_DELIM_CAPTURE);
		foreach ($parts as $part)
		{
			// Handle XML comments.
			if (!$inCdata && $part === '<!--')
			{
				$inComment = true;
			}
			if ($inComment && $part === '-->')
			{
				$inComment = false;
			}
			elseif ($inComment)
			{
				continue;
			}

			// Handle Cdata blocks.
			elseif (!$inComment && $part === '<![CDATA[')
			{
				$inCdata = true;
			}
			elseif ($inCdata && $part === ']]>')
			{
				$inCdata = false;
			}
			elseif ($inCdata)
			{
				$output .= htmlentities($part, ENT_QUOTES);
			}

			// Everything else is kept as is.
			else
			{
				$output .= $part;
			}
		}

		return $output;
	}

	/**
	 * Parse data into an array. (privately used...)
	 *
	 * @param string $data to parse
	 *
	 * @return array
	 */
	protected function _parse($data): array
	{
		// Start with an 'empty' array with no data.
		$current = [];

		// Loop until we're out of data.
		while ($data !== '')
		{
			// Find and remove the next tag.
			preg_match('/\A<([\w\-:]+)((?:\s+.+?)?)([\s]?\/)?' . '>/', $data, $match);
			if (isset($match[0]))
			{
				$data = preg_replace('/' . preg_quote($match[0], '/') . '/s', '', $data, 1);
			}

			// Didn't find a tag?  Keep looping....
			if (!isset($match[1]) || $match[1] === '')
			{
				// If there's no <, the rest is data.
				$data_strpos = strpos($data, '<');
				if ($data_strpos === false)
				{
					$text_value = $this->_from_cdata($data);
					$data = '';

					if ($text_value !== '')
					{
						$current[] = [
							'name' => '!',
							'value' => $text_value
						];
					}
				}
				// If the < isn't immediately next to the current position... more data.
				elseif ($data_strpos > 0)
				{
					$text_value = $this->_from_cdata(substr($data, 0, $data_strpos));
					$data = substr($data, $data_strpos);

					if ($text_value !== '')
					{
						$current[] = [
							'name' => '!',
							'value' => $text_value
						];
					}
				}
				// If we're looking at a </something> with no start, kill it.
				elseif ($data_strpos !== false && $data_strpos === 0)
				{
					$data_strpos = strpos($data, '<', 1);
					if ($data_strpos !== false)
					{
						$text_value = $this->_from_cdata(substr($data, 0, $data_strpos));
						$data = substr($data, $data_strpos);

						if ($text_value !== '')
						{
							$current[] = [
								'name' => '!',
								'value' => $text_value
							];
						}
					}
					else
					{
						$text_value = $this->_from_cdata($data);
						$data = '';

						if ($text_value !== '')
						{
							$current[] = [
								'name' => '!',
								'value' => $text_value
							];
						}
					}
				}

				// Wait for an actual occurrence of an element.
				continue;
			}

			// Create a new element in the array.
			$el = &$current[];
			$el['name'] = $match[1];

			// If this ISN'T empty, remove the close tag and parse the inner data.
			if ((!isset($match[3]) || trim($match[3]) !== '/') && (!isset($match[2]) || trim($match[2]) !== '/'))
			{
				// Because PHP 5.2.0+ seems to croak using regex, we'll have to do this the less fun way.
				$last_tag_end = strpos($data, '</' . $match[1] . '>');
				if ($last_tag_end === false)
				{
					continue;
				}

				$offset = 0;
				while (true)
				{
					// Where is the next start tag?
					$next_tag_start = strpos($data, '<' . $match[1], $offset);

					// If the next start tag is after the last end tag then we've found the right close.
					if ($next_tag_start === false || $next_tag_start > $last_tag_end)
					{
						break;
					}

					// If not then find the next ending tag.
					$next_tag_end = strpos($data, '</' . $match[1] . '>', $offset);

					// Didn't find one? Then just use the last and sod it.
					if ($next_tag_end === false)
					{
						break;
					}
					else
					{
						$last_tag_end = $next_tag_end;
						$offset = $next_tag_start + 1;
					}
				}

				// Parse the insides.
				$inner_match = substr($data, 0, $last_tag_end);

				// Data now starts from where this section ends.
				$data = substr($data, $last_tag_end + strlen('</' . $match[1] . '>'));

				if (!empty($inner_match))
				{
					// Parse the inner data.
					if (strpos($inner_match, '<') !== false)
					{
						$el += $this->_parse($inner_match);
					}
					elseif (trim($inner_match) !== '')
					{
						$text_value = $this->_from_cdata($inner_match);
						if (trim($text_value) !== '')
						{
							$el[] = [
								'name' => '!',
								'value' => $text_value
							];
						}
					}
				}
			}

			// If we're dealing with attributes as well, parse them out.
			if (isset($match[2]) && $match[2] !== '')
			{
				// Find all the attribute pairs in the string.
				preg_match_all('/([\w:]+)="(.+?)"/', $match[2], $attr, PREG_SET_ORDER);

				// Set them as @attribute-name.
				foreach ($attr as $match_attr)
				{
					$el['@' . $match_attr[1]] = $match_attr[2];
				}
			}
		}

		// Return the parsed array.
		return $current;
	}

	/**
	 * Turn the CDATAs back to normal text.
	 *
	 * @param string $data The data with CDATA tags
	 *
	 * @return string
	 */
	protected function _from_cdata($data): string
	{
		// Get the HTML translation table and reverse it
		$trans_tbl = array_flip(get_html_translation_table(HTML_ENTITIES, ENT_QUOTES));

		// Translate all the entities out.
		$data = preg_replace_callback('~&#(\d{1,4});~', fn($match) => $this->_from_cdata_callback($match), $data);
		$data = strtr($data, $trans_tbl);

		return $this->trim ? trim($data) : $data;
	}

	/**
	 * Callback for the preg_replace in _from_cdata
	 *
	 * @param array $match An array of data
	 *
	 * @return string
	 */
	protected function _from_cdata_callback($match): string
	{
		return chr($match[1]);
	}

	/**
	 * Get the root element's name.
	 *
	 * Example use:
	 *   echo $element->name();
	 */
	public function name()
	{
		return $this->array['name'] ?? '';
	}

	/**
	 * Get a specified element's value or attribute by path.
	 *
	 * - Children are parsed for text, but only textual data is returned
	 * unless get_elements is true.
	 *
	 * Example use:
	 *    $data = $xml->fetch('html/head/title');
	 *
	 * @param string $path - the path to the element to fetch
	 * @param bool $get_elements - whether to include elements
	 *
	 * @return bool|string
	 */
	public function fetch($path, $get_elements = false)
	{
		// Get the element, in array form.
		$array = $this->path($path);

		if ($array === false)
		{
			return false;
		}

		// Getting elements into this is a bit complicated...
		if ($get_elements && !is_string($array))
		{
			$temp = '';

			// Use the _xml() function to get the xml data.
			foreach ($array->array as $val)
			{
				// Skip the name and any attributes.
				if (is_array($val))
				{
					$temp .= $this->_xml($val, null);
				}
			}

			// Just get the XML data and then take out the CDATAs.
			return $this->_to_cdata($temp);
		}

		// Return the value - taking care to pick out all the text values.
		return is_string($array) ? $array : $this->_fetch($array->array);

The condition is_string($array) is always false.

	}

	/**
	 * Get an element, returns a new \ElkArte\XmlArray.
	 *
	 * - It finds any elements that match the path specified.
	 * - It will always return a set if there is more than one of the element
	 * or return_set is true.
	 *
	 * Example use:
	 *   $element = $xml->path('html/body');
	 *
	 * @param string $path - the path to the element to get
	 * @param bool $return_full - always return full result set
	 * @return \ElkArte\XmlArray|bool a new \ElkArte\XmlArray.
	 */
	public function path($path, $return_full = false)
	{
		// Split up the path.
		$path = explode('/', $path);

		// Start with a base array.
		$array = $this->array;

		// For each element in the path.
		foreach ($path as $el)
		{
			// Deal with sets....
			if (strpos($el, '[') !== false)
			{
				$lvl = (int) substr($el, strpos($el, '[') + 1);
				$el = substr($el, 0, strpos($el, '['));
			}
			// Find an attribute.
			elseif (substr($el, 0, 1) === '@')
			{
				// It simplifies things if the attribute is already there ;).
				if (isset($array[$el]))
				{
					return $array[$el];
				}

				$trace = debug_backtrace(DEBUG_BACKTRACE_IGNORE_ARGS);
				$i = 0;
				while ($i < count($trace) && isset($trace[$i]['class']) && $trace[$i]['class'] === get_class($this))
				{
					$i++;
				}
				$debug = ' from ' . $trace[$i - 1]['file'] . ' on line ' . $trace[$i - 1]['line'];

				// Cause an error.
				if (($this->debug_level & E_NOTICE) !== 0)
				{
					trigger_error('Undefined XML attribute: ' . substr($el, 1) . $debug, E_USER_NOTICE);
				}

				return false;
			}
			else
			{
				$lvl = null;
			}

			// Find this element.
			$array = $this->_path($array, $el, $lvl);
		}

		// Clean up after $lvl, for $return_full.
		if ($return_full && (!isset($array['name']) || substr($array['name'], -1) !== ']'))
		{
			$array = ['name' => $el . '[]', $array];

The variable $el seems to be defined by a foreach iteration on line 419. Are you sure the iterator is never empty, otherwise this variable is not defined?

		}

		// Create the right type of class...
		$newClass = get_class($this);

		// Return a new \ElkArte\XmlArray for the result.
		return $array === false ? false : new $newClass($array, $this->trim, $this->debug_level, true);
	}

	/**
	 * Get a specific array by path, one level down. (privately used...)
	 *
	 * @param array $array An array of data
	 * @param string $path The path
	 * @param int $level How far deep into the array we should go
	 * @param bool $no_error Whether or not to ignore errors
	 *
	 * @return array|bool
	 */
	protected function _path($array, $path, $level, $no_error = false)
	{
		// Is $array even an array?  It might be false!
		if (!is_array($array))

The condition is_array($array) is always true.

		{
			return false;
		}

		// Asking for *no* path?
		if ($path === '' || $path === '.')
		{
			return $array;
		}
		$paths = explode('|', $path);

		// A * means all elements of any name.
		$show_all = in_array('*', $paths);

		$results = [];

		// Check each element.
		foreach ($array as $value)
		{
			if (!is_array($value) || $value['name'] === '!')
			{
				continue;
			}

			if ($show_all || in_array($value['name'], $paths, true))
			{
				// Skip elements before "the one".
				if ($level !== null && $level > 0)
				{
					$level--;
				}
				else
				{
					$results[] = $value;
				}
			}
		}

		// No results found...
		if (empty($results))
		{
			$trace = debug_backtrace(DEBUG_BACKTRACE_IGNORE_ARGS);
			$i = 0;
			while ($i < count($trace) && isset($trace[$i]['class']) && $trace[$i]['class'] === get_class($this))
			{
				$i++;
			}

			$debug = ' from ' . $trace[$i - 1]['file'] . ' on line ' . $trace[$i - 1]['line'];

			// Cause an error.
			if ($this->debug_level & E_NOTICE && !$no_error)
			{
				trigger_error('Undefined XML element: ' . $path . $debug, E_USER_NOTICE);
			}

			return false;
		}

		// Only one result.
		if ($level !== null || count($results) === 1)

The condition $level !== null is always true.

		{
			return $results[0];
		}

		// Return the result set.
		return $results + ['name' => $path . '[]'];
	}

	/**
	 * Get a specific element's xml. (privately used...)
	 *
	 * @param array $array
	 * @param null|int $indent
	 *
	 * @return string
	 */
	protected function _xml($array, $indent): string
	{
		$indentation = $indent !== null ? '
' . str_repeat('	', $indent) : '';

		// This is a set of elements, with no name...
		if (is_array($array) && !isset($array['name']))
		{
			$temp = '';
			foreach ($array as $val)
			{
				$temp .= $this->_xml($val, $indent);
			}

			return $temp;
		}

		// This is just text!
		if ($array['name'] === '!')
		{
			return $indentation . '<![CDATA[' . $array['value'] . ']]>';
		}

		if (substr($array['name'], -2) === '[]')
		{
			$array['name'] = substr($array['name'], 0, -2);
		}

		// Start the element.
		$output = $indentation . '<' . $array['name'];

		$inside_elements = false;
		$output_el = '';

		// Run through and recursively output all the elements or attributes inside this.
		foreach ($array as $k => $v)
		{
			if (substr($k, 0, 1) === '@')
			{
				$output .= ' ' . substr($k, 1) . '="' . $v . '"';
			}
			elseif (is_array($v))
			{
				$output_el .= $this->_xml($v, $indent === null ? null : $indent + 1);
				$inside_elements = true;
			}
		}

		// Indent, if necessary.... then close the tag.
		if ($inside_elements)
		{
			$output .= '>' . $output_el . $indentation . '</' . $array['name'] . '>';
		}
		else
		{
			$output .= ' />';
		}

		return $output;
	}

	/**
	 * Given an array, return the text from that array. (recursive and privately used.)
	 *
	 * @param string[]|string $array An array of data
	 *
	 * @return string
	 */
	protected function _fetch($array): string
	{
		// Don't return anything if this is just a string.
		if (is_string($array))
		{
			return '';
		}

		$temp = '';
		foreach ($array as $text)
		{
			// This means it's most likely an attribute or the name itself.
			if (!isset($text['name']))
			{
				continue;
			}

			// This is text!
			if ($text['name'] === '!')
			{
				$temp .= $text['value'];
			}
			// Another element - dive in ;).
			else
			{
				$temp .= $this->_fetch($text);
			}
		}

		// Return all the bits and pieces we've put together.
		return $temp;
	}

	/**
	 * Check if an element exists.
	 *
	 * Example use,
	 *   echo $xml->exists('html/body') ? 'y' : 'n';
	 *
	 * @param string $path - the path to the element to get.
	 * @return bool
	 */
	public function exists($path): bool
	{
		// Split up the path.
		$path = explode('/', $path);

		// Start with a base array.
		$array = $this->array;

		// For each element in the path.
		foreach ($path as $el)
		{
			// Deal with sets....
			$el_strpos = strpos($el, '[');

			if ($el_strpos !== false)
			{
				$lvl = (int) substr($el, $el_strpos + 1);
				$el = substr($el, 0, $el_strpos);
			}
			// Find an attribute.
			elseif (substr($el, 0, 1) === '@')
			{
				return isset($array[$el]);
			}
			else
			{
				$lvl = null;
			}

			// Find this element.
			$array = $this->_path($array, $el, $lvl, true);

It seems like $array can also be of type boolean; however, parameter $array of ElkArte\XmlArray::_path() does only seem to accept array, maybe add an additional type check?

		}

		return $array !== false;
	}

	/**
	 * Count the number of occurrences of a path.
	 *
	 * Example use:
	 *   echo $xml->count('html/head/meta');
	 *
	 * @param string $path - the path to search for.
	 * @return int the number of elements the path matches.
	 */
	public function count($path): int
	{
		// Get the element, always returning a full set.
		$temp = $this->path($path, true);

		// Start at zero, then count up all the numeric keys.
		$i = 0;
		foreach ($temp->array as $item)
		{
			if (is_array($item))
			{
				$i++;
			}
		}

		return $i;
	}

	/**
	 * Get an array of \ElkArte\XmlArray's matching the specified path.
	 *
	 * - This differs from ->path(path, true) in that instead of an \ElkArte\XmlArray
	 * of elements, an array of \ElkArte\XmlArray's is returned for use with foreach.
	 *
	 * Example use:
	 *   foreach ($xml->set('html/body/p') as $p)
	 *
	 * @param string $path - the path to search for.
	 * @return array an array of \ElkArte\XmlArray objects
	 */
	public function set($path): array
	{
		// None as yet, just get the path.
		$array = [];
		$xml = $this->path($path, true);

		foreach ($xml->array as $val)
		{
			// Skip these, they aren't elements.
			if (!is_array($val) || $val['name'] === '!')
			{
				continue;
			}

			// Create the right type of class...
			$newClass = get_class($this);

			// Create a new \ElkArte\XmlArray and stick it in the array.
			$array[] = new $newClass($val, $this->trim, $this->debug_level, true);
		}

		return $array;
	}

	/**
	 * Create an xml file from an \ElkArte\XmlArray, the specified path if any.
	 *
	 * Example use:
	 *   echo $this->create_xml();
	 *
	 * @param string|null $path - the path to the element. (optional)
	 * @return string xml-formatted string.
	 */
	public function create_xml($path = null)
	{
		// Was a path specified?  If so, use that array.
		if ($path !== null)
		{
			$path = $this->path($path);

			// The path was not found
			if ($path === false)
			{
				return false;

The expression return false returns the type false which is incompatible with the documented return type string.

			}

			$path = $path->array;
		}
		// Just use the current array.
		else
		{
			$path = $this->array;
		}

		// Add the xml declaration to the front.
		return '<?xml version="1.0"?' . '>' . $this->_xml($path, 0);
	}

	/**
	 * Output the xml in an array form.
	 *
	 * Example use:
	 *   print_r($xml->to_array());
	 *
	 * @param string|null $path the path to output.
	 *
	 * @return array|bool|string
	 */
	public function to_array($path = null)
	{
		// Are we doing a specific path?
		if ($path !== null)
		{
			$path = $this->path($path);

			// The path was not found
			if ($path === false)
			{
				return false;
			}

			$path = $path->array;
		}
		// No, so just use the current array.
		else
		{
			$path = $this->array;
		}

		return $this->_array($path);
	}

	/**
	 * Return an element as an array
	 *
	 * @param array $array An array of data
	 *
	 * @return array|string
	 */
	protected function _array($array)
	{
		$return = [];
		$text = '';
		foreach ($array as $value)
		{
			if (!is_array($value) || !isset($value['name']))
			{
				continue;
			}

			if ($value['name'] === '!')
			{
				$text .= $value['value'];
			}
			else
			{
				$return[$value['name']] = $this->_array($value);
			}
		}

		if (empty($return))
		{
			return $text;
		}

		return $return;
	}
}