<?php
/**
 * Sequence matcher for Diff
 *
 * PHP version 5
 *
 * Copyright (c) 2009 Chris Boulton <chris.boulton@interspire.com>
 *
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions are met:
 *
 *  - Redistributions of source code must retain the above copyright notice,
 *    this list of conditions and the following disclaimer.
 *  - Redistributions in binary form must reproduce the above copyright notice,
 *    this list of conditions and the following disclaimer in the documentation
 *    and/or other materials provided with the distribution.
 *  - Neither the name of the Chris Boulton nor the names of its contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
 * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
 * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
 * POSSIBILITY OF SUCH DAMAGE.
 *
 * @package Diff
 * @author Chris Boulton <chris.boulton@interspire.com>
 * @copyright (c) 2009 Chris Boulton
 * @license New BSD License http://www.opensource.org/licenses/bsd-license.php
 * @version 1.1
 * @link http://github.com/chrisboulton/php-diff
 */

class Diff_SequenceMatcher
{
	/**
	 * @var string|array Either a string or an array containing a callback function to determine if a line is "junk" or not.
	 */
	private $junkCallback = null;

	/**
	 * @var array The first sequence to compare against.
	 */
	private $a = null;

	/**
	 * @var array The second sequence.
	 */
	private $b = null;

	/**
	 * @var array Array of characters that are considered junk from the second sequence. Characters are the array key.
	 */
	private $junkDict = array();

	/**
	 * @var array Array of indices that do not contain junk elements.
	 */
	private $b2j = array();

	private $options = array();

	private $defaultOptions = array(
		'ignoreNewLines' => false,
		'ignoreWhitespace' => false,
		'ignoreCase' => false
	);
	
	private $matchingBlocks = null;
	private $opCodes = null;
	private $fullBCount = null;

	/**
	 * The constructor. With the sequences being passed, they'll be set for the
	 * sequence matcher and it will perform a basic cleanup & calculate junk
	 * elements.
	 *
	 * @param string|array $a A string or array containing the lines to compare against.
	 * @param string|array $b A string or array containing the lines to compare.
	 * @param string|array $junkCallback Either an array or string that references a callback function (if there is one) to determine 'junk' characters.
	 */
	public function __construct($a, $b, $junkCallback=null, $options=array())
	{
		$this->a = null;
		$this->b = null;
		$this->junkCallback = $junkCallback;
		$this->setOptions($options);
		$this->setSequences($a, $b);
	}

	public function setOptions($options)
	{
		$this->options = array_merge($this->defaultOptions, $options);
	}

	/**
	 * Set the first and second sequences to use with the sequence matcher.
	 *
	 * @param string|array $a A string or array containing the lines to compare against.
	 * @param string|array $b A string or array containing the lines to compare.
	 */
	public function setSequences($a, $b)
	{
		$this->setSeq1($a);
		$this->setSeq2($b);
	}

	/**
	 * Set the first sequence ($a) and reset any internal caches to indicate that
	 * when calling the calculation methods, we need to recalculate them.
	 *
	 * @param string|array $a The sequence to set as the first sequence.
	 */
	public function setSeq1($a)
	{
		if(!is_array($a)) {
			$a = str_split($a);
		}
		if($a == $this->a) {
			return;
		}

		$this->a= $a;
		$this->matchingBlocks = null;
		$this->opCodes = null;
	}

	/**
	 * Set the second sequence ($b) and reset any internal caches to indicate that
	 * when calling the calculation methods, we need to recalculate them.
	 *
	 * @param string|array $b The sequence to set as the second sequence.
	 */
	public function setSeq2($b)
	{
		if(!is_array($b)) {
			$b = str_split($b);
		}
		if($b == $this->b) {
			return;
		}

		$this->b = $b;
		$this->matchingBlocks = null;
		$this->opCodes = null;
		$this->fullBCount = null;
		$this->chainB();
	}

	/**
	 * Generate the internal arrays containing the list of junk and non-junk
	 * characters for the second ($b) sequence.
	 */
	private function chainB()
	{
		$length = count ($this->b);
		$this->b2j = array();
		$popularDict = array();

		for($i = 0; $i < $length; ++$i) {
			$char = $this->b[$i];
			if(isset($this->b2j[$char])) {
				if($length >= 200 && count($this->b2j[$char]) * 100 > $length) {
					$popularDict[$char] = 1;
					unset($this->b2j[$char]);
				}
				else {
					$this->b2j[$char][] = $i;
				}
			}
			else {
				$this->b2j[$char] = array(
					$i
				);
			}
		}

		// Remove leftovers
		foreach(array_keys($popularDict) as $char) {
			unset($this->b2j[$char]);
		}

		$this->junkDict = array();
		if(is_callable($this->junkCallback)) {
			foreach(array_keys($popularDict) as $char) {
				if(call_user_func($this->junkCallback, $char)) {
					$this->junkDict[$char] = 1;
					unset($popularDict[$char]);
				}
			}

			foreach(array_keys($this->b2j) as $char) {
				if(call_user_func($this->junkCallback, $char)) {
					$this->junkDict[$char] = 1;
					unset($this->b2j[$char]);
				}
			}
		}
	}

	/**
	 * Checks if a particular character is in the junk dictionary
	 * for the list of junk characters.
	 *
	 * @return boolean $b True if the character is considered junk. False if not.
	 */
	private function isBJunk($b)
	{
		if(isset($this->juncDict[$b])) {
			return true;
		}

		return false;
	}

	/**
	 * Find the longest matching block in the two sequences, as defined by the
	 * lower and upper constraints for each sequence. (for the first sequence,
	 * $alo - $ahi and for the second sequence, $blo - $bhi)
	 *
	 * Essentially, of all of the maximal matching blocks, return the one that
	 * startest earliest in $a, and all of those maximal matching blocks that
	 * start earliest in $a, return the one that starts earliest in $b.
	 *
	 * If the junk callback is defined, do the above but with the restriction
	 * that the junk element appears in the block. Extend it as far as possible
	 * by matching only junk elements in both $a and $b.
	 *
	 * @param int $alo The lower constraint for the first sequence.
	 * @param int $ahi The upper constraint for the first sequence.
	 * @param int $blo The lower constraint for the second sequence.
	 * @param int $bhi The upper constraint for the second sequence.
	 * @return array Array containing the longest match that includes the starting position in $a, start in $b and the length/size.
	 */
	public function findLongestMatch($alo, $ahi, $blo, $bhi)
	{
		$a = $this->a;
		$b = $this->b;

		$bestI = $alo;
		$bestJ = $blo;
		$bestSize = 0;

		$j2Len = array();
		$nothing = array();

		for($i = $alo; $i < $ahi; ++$i) {
			$newJ2Len = array();
			$jDict = $this->arrayGetDefault($this->b2j, $a[$i], $nothing);
			foreach($jDict as $jKey => $j) {
				if($j < $blo) {
					continue;
				}
				else if($j >= $bhi) {
					break;
				}

				$k = $this->arrayGetDefault($j2Len, $j -1, 0) + 1;
				$newJ2Len[$j] = $k;
				if($k > $bestSize) {
					$bestI = $i - $k + 1;
					$bestJ = $j - $k + 1;
					$bestSize = $k;
				}
			}

			$j2Len = $newJ2Len;
		}

		while($bestI > $alo && $bestJ > $blo && !$this->isBJunk($b[$bestJ - 1]) &&
			!$this->linesAreDifferent($bestI - 1, $bestJ - 1)) {
				--$bestI;
				--$bestJ;
				++$bestSize;
		}

		while($bestI + $bestSize < $ahi && ($bestJ + $bestSize) < $bhi &&
			!$this->isBJunk($b[$bestJ + $bestSize]) && !$this->linesAreDifferent($bestI + $bestSize, $bestJ + $bestSize)) {
				++$bestSize;
		}

		while($bestI > $alo && $bestJ > $blo && $this->isBJunk($b[$bestJ - 1]) &&
			!$this->isLineDifferent($bestI - 1, $bestJ - 1)) {
				--$bestI;
				--$bestJ;
				++$bestSize;
		}

		while($bestI + $bestSize < $ahi && $bestJ + $bestSize < $bhi &&
			$this->isBJunk($b[$bestJ + $bestSize]) && !$this->linesAreDifferent($bestI + $bestSize, $bestJ + $bestSize)) {
					++$bestSize;
		}

		return array(
			$bestI,
			$bestJ,
			$bestSize
		);
	}

	/**
	 * Check if the two lines at the given indexes are different or not.
	 *
	 * @param int $aIndex Line number to check against in a.
	 * @param int $bIndex Line number to check against in b.
	 * @return boolean True if the lines are different and false if not.
	 */
	public function linesAreDifferent($aIndex, $bIndex)
	{
		$lineA = $this->a[$aIndex];
		$lineB = $this->b[$bIndex];

		if($this->options['ignoreWhitespace']) {
			$replace = array("\t", ' ');
			$lineA = str_replace($replace, '', $lineA);
			$lineB = str_replace($replace, '', $lineB);
		}

		if($this->options['ignoreCase']) {
			$lineA = strtolower($lineA);
			$lineB = strtolower($lineB);
		}

		if($lineA != $lineB) {
			return true;
		}

		return false;
	}

	/**
	 * Return a nested set of arrays for all of the matching sub-sequences
	 * in the strings $a and $b.
	 *
	 * Each block contains the lower constraint of the block in $a, the lower
	 * constraint of the block in $b and finally the number of lines that the
	 * block continues for.
	 *
	 * @return array Nested array of the matching blocks, as described by the function.
	 */
	public function getMatchingBlocks()
	{
		if(!empty($this->matchingBlocks)) {
			return $this->matchingBlocks;
		}

		$aLength = count($this->a);
		$bLength = count($this->b);

		$queue = array(
			array(
				0,
				$aLength,
				0,
				$bLength
			)
		);

		$matchingBlocks = array();
		while(!empty($queue)) {
			list($alo, $ahi, $blo, $bhi) = array_pop($queue);
			$x = $this->findLongestMatch($alo, $ahi, $blo, $bhi);
			list($i, $j, $k) = $x;
			if($k) {
				$matchingBlocks[] = $x;
				if($alo < $i && $blo < $j) {
					$queue[] = array(
						$alo,
						$i,
						$blo,
						$j
					);
				}

				if($i + $k < $ahi && $j + $k < $bhi) {
					$queue[] = array(
						$i + $k,
						$ahi,
						$j + $k,
						$bhi
					);
				}
			}
		}

		usort($matchingBlocks, array($this, 'tupleSort'));

		$i1 = 0;
		$j1 = 0;
		$k1 = 0;
		$nonAdjacent = array();
		foreach($matchingBlocks as $block) {
			list($i2, $j2, $k2) = $block;
			if($i1 + $k1 == $i2 && $j1 + $k1 == $j2) {
				$k1 += $k2;
			}
			else {
				if($k1) {
					$nonAdjacent[] = array(
						$i1,
						$j1,
						$k1
					);
				}

				$i1 = $i2;
				$j1 = $j2;
				$k1 = $k2;
			}
		}

		if($k1) {
			$nonAdjacent[] = array(
				$i1,
				$j1,
				$k1
			);
		}

		$nonAdjacent[] = array(
			$aLength,
			$bLength,
			0
		);

		$this->matchingBlocks = $nonAdjacent;
		return $this->matchingBlocks;
	}

	/**
	 * Return a list of all of the opcodes for the differences between the
	 * two strings.
	 *
	 * The nested array returned contains an array describing the opcode
	 * which includes:
	 * 0 - The type of tag (as described below) for the opcode.
	 * 1 - The beginning line in the first sequence.
	 * 2 - The end line in the first sequence.
	 * 3 - The beginning line in the second sequence.
	 * 4 - The end line in the second sequence.
	 *
	 * The different types of tags include:
	 * replace - The string from $i1 to $i2 in $a should be replaced by
	 *           the string in $b from $j1 to $j2.
	 * delete -  The string in $a from $i1 to $j2 should be deleted.
	 * insert -  The string in $b from $j1 to $j2 should be inserted at
	 *           $i1 in $a.
	 * equal  -  The two strings with the specified ranges are equal.
	 *
	 * @return array Array of the opcodes describing the differences between the strings.
	 */
	public function getOpCodes()
	{
		if(!empty($this->opCodes)) {
			return $this->opCodes;
		}

		$i = 0;
		$j = 0;
		$this->opCodes = array();

		$blocks = $this->getMatchingBlocks();
		foreach($blocks as $block) {
			list($ai, $bj, $size) = $block;
			$tag = '';
			if($i < $ai && $j < $bj) {
				$tag = 'replace';
			}
			else if($i < $ai) {
				$tag = 'delete';
			}
			else if($j < $bj) {
				$tag = 'insert';
			}

			if($tag) {
				$this->opCodes[] = array(
					$tag,
					$i,
					$ai,
					$j,
					$bj
				);
			}

			$i = $ai + $size;
			$j = $bj + $size;

			if($size) {
				$this->opCodes[] = array(
					'equal',
					$ai,
					$i,
					$bj,
					$j
				);
			}
		}
		return $this->opCodes;
	}

	/**
	 * Return a series of nested arrays containing different groups of generated
	 * opcodes for the differences between the strings with up to $context lines
	 * of surrounding content.
	 *
	 * Essentially what happens here is any big equal blocks of strings are stripped
	 * out, the smaller subsets of changes are then arranged in to their groups.
	 * This means that the sequence matcher and diffs do not need to include the full
	 * content of the different files but can still provide context as to where the
	 * changes are.
	 *
	 * @param int $context The number of lines of context to provide around the groups.
	 * @return array Nested array of all of the grouped opcodes.
	 */
	public function getGroupedOpcodes($context=3)
	{
		$opCodes = $this->getOpCodes();
		if(empty($opCodes)) {
			$opCodes = array(
				array(
					'equal',
					0,
					1,
					0,
					1
				)
			);
		}

		if($opCodes[0][0] == 'equal') {
			$opCodes[0] = array(
				$opCodes[0][0],
				max($opCodes[0][1], $opCodes[0][2] - $context),
				$opCodes[0][2],
				max($opCodes[0][3], $opCodes[0][4] - $context),
				$opCodes[0][4]
			);
		}

		$lastItem = count($opCodes) - 1;
		if($opCodes[$lastItem][0] == 'equal') {
			list($tag, $i1, $i2, $j1, $j2) = $opCodes[$lastItem];
			$opCodes[$lastItem] = array(
				$tag,
				$i1,
				min($i2, $i1 + $context),
				$j1,
				min($j2, $j1 + $context)
			);
		}

		$maxRange = $context * 2;
		$groups = array();
		$group = array();
		foreach($opCodes as $code) {
			list($tag, $i1, $i2, $j1, $j2) = $code;
			if($tag == 'equal' && $i2 - $i1 > $maxRange) {
				$group[] = array(
					$tag,
					$i1,
					min($i2, $i1 + $context),
					$j1,
					min($j2, $j1 + $context)
				);
				$groups[] = $group;
				$group = array();
				$i1 = max($i1, $i2 - $context);
				$j1 = max($j1, $j2 - $context);
			}
			$group[] = array(
				$tag,
				$i1,
				$i2,
				$j1,
				$j2
			);
		}

		if(!empty($group) && !(count($group) == 1 && $group[0][0] == 'equal')) {
			$groups[] = $group;
		}

		return $groups;
	}

	/**
	 * Return a measure of the similarity between the two sequences.
	 * This will be a float value between 0 and 1.
	 *
	 * Out of all of the ratio calculation functions, this is the most
	 * expensive to call if getMatchingBlocks or getOpCodes is yet to be
	 * called. The other calculation methods (quickRatio and realquickRatio)
	 * can be used to perform quicker calculations but may be less accurate.
	 *
	 * The ratio is calculated as (2 * number of matches) / total number of
	 * elements in both sequences.
	 *
	 * @return float The calculated ratio.
	 */
	public function Ratio()
	{
		$matches = array_reduce($this->getMatchingBlocks(), array($this, 'ratioReduce'), 0);
		return $this->calculateRatio($matches, count ($this->a) + count ($this->b));
	}

	/**
	 * Helper function to calculate the number of matches for Ratio().
	 *
	 * @param int $sum The running total for the number of matches.
	 * @param array $triple Array containing the matching block triple to add to the running total.
	 * @return int The new running total for the number of matches.
	 */
	private function ratioReduce($sum, $triple)
	{
		return $sum + ($triple[count($triple) - 1]);
	}


	/**
	 * Helper function for calculating the ratio to measure similarity for the strings.
	 * The ratio is defined as being 2 * (number of matches / total length)
	 *
	 * @param int $matches The number of matches in the two strings.
	 * @param int $length The length of the two strings.
	 * @return float The calculated ratio.
	 */
	private function calculateRatio($matches, $length=0)
	{
		if($length) {
			return 2 * ($matches / $length);
		}
		else {
			return 1;
		}
	}

	/**
	 * Helper function that provides the ability to return the value for a key
	 * in an array of it exists, or if it doesn't then return a default value.
	 * Essentially cleaner than doing a series of if(isset()) {} else {} calls.
	 *
	 * @param array $array The array to search.
	 * @param string $key The key to check that exists.
	 * @param mixed $default The value to return as the default value if the key doesn't exist.
	 * @return mixed The value from the array if the key exists or otherwise the default.
	 */
	private function arrayGetDefault($array, $key, $default)
	{
		if(isset($array[$key])) {
			return $array[$key];
		}
		else {
			return $default;
		}
	}

	/**
	 * Sort an array by the nested arrays it contains. Helper function for getMatchingBlocks
	 *
	 * @param array $a First array to compare.
	 * @param array $b Second array to compare.
	 * @return int -1, 0 or 1, as expected by the usort function.
	 */
	private function tupleSort($a, $b)
	{
		$max = max(count($a), count($b));
		for($i = 0; $i < $max; ++$i) {
			if($a[$i] < $b[$i]) {
				return -1;
			}
			else if($a[$i] > $b[$i]) {
				return 1;
			}
		}

		if(count($a) == count($b)) {
			return 0;
		}
		else if(count($a) < count($b)) {
			return -1;
		}
		else {
			return 1;
		}
	}
}