class.tslib_search.php

Typo3, 开源里边最强大的

<?php
/***************************************************************
*  Copyright notice
*
*  (c) 1999-2005 Kasper Skaarhoj ([e-mail])
*  All rights reserved
*
*  This script is part of the TYPO3 project. The TYPO3 project is
*  free software; you can redistribute it and/or modify
*  it under the terms of the GNU General Public License as published by
*  the Free Software Foundation; either version 2 of the License, or
*  (at your option) any later version.
*
*  The GNU General Public License can be found at
*  http://www.gnu.org/copyleft/gpl.html.
*  A copy is found in the textfile GPL.txt and important notices to the license
*  from the author is found in LICENSE.txt distributed with these scripts.
*
*
*  This script is distributed in the hope that it will be useful,
*  but WITHOUT ANY WARRANTY; without even the implied warranty of
*  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
*  GNU General Public License for more details.
*
*  This copyright notice MUST APPEAR in all copies of the script!
***************************************************************/
/**
 * Searching in database tables, typ. "pages" and "tt_content"
 * Used to generate search queries for TypoScript.
 * The class is included from "class.tslib_pagegen.php" based on whether there has been detected content in the GPvar "sword"
 *
 * $Id: class.tslib_search.php 593 2005-04-01 14:37:15Z typo3 $
 * Revised for TYPO3 3.6 June/2003 by Kasper Skaarhoj
 *
 * @author	Kasper Skaarhoj <[e-mail]>
 * @author	Rene Fritz	<[e-mail]>
 */
/**
 * [CLASS/FUNCTION INDEX of SCRIPT]
 *
 *
 *
 *   88: class tslib_search
 *  127:     function register_tables_and_columns($requestedCols,$allowedCols)
 *  168:     function explodeCols($in)
 *  193:     function register_and_explode_search_string($sword)
 *  226:     function split($origSword, $specchars='+-', $delchars='+.,-')
 *  269:     function quotemeta($str)
 *  285:     function build_search_query($endClause)
 *  371:     function build_search_query_for_searchwords()
 *  413:     function get_operator($operator)
 *  436:     function count_query()
 *  449:     function execute_query()
 *  462:     function get_searchwords()
 *  477:     function get_searchwordsArray()
 *
 * TOTAL FUNCTIONS: 12
 * (This index is automatically created/updated by the extension "extdeveval")
 *
 */



















/**
 * Search class used for the content object SEARCHRESULT
 *
 * @author	Kasper Skaarhoj <[e-mail]>
 * @package TYPO3
 * @subpackage tslib
 * @see	tslib_cObj::SEARCHRESULT()
 */
class tslib_search {
	var $tables = Array ();

	var $group_by = 'PRIMARY_KEY';							// Alternatively 'PRIMARY_KEY'; sorting by primary key
	var $default_operator = 'AND';							// Standard SQL-operator between words
	var $operator_translate_table_caseinsensitive = TRUE;
	var $operator_translate_table = Array (					// case-sensitiv. Defineres the words, which will be operators between words
		Array ('+' , 'AND'),
		Array ('|' , 'AND'),
		Array ('-' , 'AND NOT'),
			// english
		Array ('and' , 'AND'),
		Array ('or' , 'OR'),
		Array ('not' , 'AND NOT'),
	);

	// Internal
	var $sword_array;		// Contains the search-words and operators
	var $queryParts;		// Contains the query parts after processing.

	var $other_where_clauses;	// Addition to the whereclause. This could be used to limit search to a certain page or alike in the system.
	var $fTable;		// This is set with the foreign table that 'pages' are connected to.

	var $res_offset = 0;	// How many rows to offset from the beginning
	var $res_shows = 20;	// How many results to show (0 = no limit)
	var $res_count;			// Intern: How many results, there was last time (with the exact same searchstring.

	var $pageIdList='';		// List of pageIds.

	var $listOfSearchFields ='';

	/**
	 * Creates the $this->tables-array.
	 * The 'pages'-table is ALWAYS included as the search is page-based. Apart from this there may be one and only one table, joined with the pages-table. This table is the first table mentioned in the requested-list. If any more tables are set here, they are ignored.
	 *
	 * @param	string		is a list (-) of columns that we want to search. This could be input from the search-form (see TypoScript documentation)
	 * @param	string		$allowedCols: is the list of columns, that MAY be searched. All allowed cols are set as result-fields. All requested cols MUST be in the allowed-fields list.
	 * @return	void
	 */
	function register_tables_and_columns($requestedCols,$allowedCols)	{
		$rCols=$this->explodeCols($requestedCols);
		$aCols=$this->explodeCols($allowedCols);

		foreach ($rCols as $k => $v)	{
			$rCols[$k]=trim($v);
			if (in_array($rCols[$k], $aCols))	{
				$parts = explode('.',$rCols[$k]);
				$this->tables[$parts[0]]['searchfields'][] = $parts[1];
			}
		}
		$this->tables['pages']['primary_key'] = 'uid';
		$this->tables['pages']['resultfields'][] = 'uid';
		unset($this->tables['pages']['fkey']);

		foreach ($aCols as $k => $v)	{
			$aCols[$k]=trim($v);
			$parts = explode('.',$aCols[$k]);
			$this->tables[$parts[0]]['resultfields'][] = $parts[1].' AS '.str_replace('.','_',$aCols[$k]);
			$this->tables[$parts[0]]['fkey']='pid';
		}

		$this->fTable='';
		foreach ($this->tables as $t => $v)	{
			if ($t!='pages')	{
				if (!$this->fTable)	{
					$this->fTable = $t;
				} else {
					unset($this->tables[$t]);
				}
			}
		}
	}

	/**
	 * Function that can convert the syntax for entering which tables/fields the search should be conducted in.
	 *
	 * @param	string		This is the code-line defining the tables/fields to search. Syntax: '[table1].[field1]-[field2]-[field3] : [table2].[field1]-[field2]'
	 * @return	array		An array where the values is "[table].[field]" strings to search
	 * @see	register_tables_and_columns()
	 */
	function explodeCols($in)	{
		$theArray = explode(':',$in);
		$out = Array();
		while(list(,$val)=each($theArray))	{
			$val=trim($val);
			$parts = explode('.',$val);
			if ($parts[0] && $parts[1])	{
				$subparts = explode('-',$parts[1]);
				while(list(,$piece)=each($subparts))	{
					$piece=trim($piece);
					if ($piece)		$out[]=$parts[0].'.'.$piece;
				}
			}
		}
		return $out;
	}

	/**
	 * Takes a search-string (WITHOUT SLASHES or else it'll be a little sppooky , NOW REMEMBER to unslash!!)
	 * Sets up $this->sword_array op with operators.
	 * This function uses $this->operator_translate_table as well as $this->default_operator
	 *
	 * @param	string		The input search-word string.
	 * @return	void
	 */
	function register_and_explode_search_string($sword)	{
		$sword = trim($sword);
		if ($sword)	{
			$components = $this->split($sword);
			$s_sword = '';	 // the searchword is stored here during the loop
			if (is_array($components))	{
				$i=0;
				$lastoper = '';
				reset($components);
				while (list($key,$val) = each ($components))	{
					$operator=$this->get_operator($val);
					if ($operator)	{
						$lastoper = $operator;
					} elseif (strlen($val)>1) {		// A searchword MUST be at least two characters long!
						$this->sword_array[$i]['sword'] = $val;
						$this->sword_array[$i]['oper'] = ($lastoper) ? $lastoper : $this->default_operator;
						$lastoper = '';
						$i++;
					}
				}
			}
		}
	}

	/**
	 * Used to split a search-word line up into elements to search for. This function will detect boolean words like AND and OR, + and -, and even find sentences encapsulated in ""
	 * This function could be re-written to be more clean and effective - yet it's not that important.
	 *
	 * @param	string		The raw sword string from outside
	 * @param	string		Special chars which are used as operators (+- is default)
	 * @param	string		Special chars which are deleted if the append the searchword (+-., is default)
	 * @return	mixed		Returns an ARRAY if there were search words, othervise the return value may be unset.
	 */
	function split($origSword, $specchars='+-', $delchars='+.,-')	{
		$sword = $origSword;
		$specs = '['.$this->quotemeta($specchars).']';
		$delchars = '['.$this->quotemeta($delchars).']';

			// As long as $sword is true (that means $sword MUST be reduced little by little until its empty inside the loop!)
		while ($sword)	{
			if (ereg('^"',$sword))	{		// There was a double-quote and we will then look for the ending quote.
				$sword = ereg_replace('^"','',$sword);		// Removes first double-quote
				ereg('^[^"]*',$sword,$reg);  // Removes everything till next double-quote
				$value[] = $reg[0];  // reg[0] is the value, should not be trimmed
				$sword = ereg_replace('^'.$this->quotemeta($reg[0]),'',$sword);
				$sword = trim(ereg_replace('^"','',$sword));		// Removes last double-quote
			} elseif (ereg('^'.$specs,$sword,$reg)) {
				$value[] = $reg[0];
				$sword = trim(ereg_replace('^'.$specs,'',$sword));		// Removes = sign
			} elseif (ereg('[\+\-]',$sword)) {	// Check if $sword contains + or -
					// + and - shall only be interpreted as $specchars when there's whitespace before it
					// otherwise it's included in the searchword (e.g. "know-how")
				$a_sword = explode(' ',$sword);	// explode $sword to single words
				$word = array_shift($a_sword);	// get first word
				$word = ereg_replace($delchars.'$','',$word);		// Delete $delchars at end of string
				$value[] = $word;	// add searchword to values