queryparser.php

Bug tracker, and reporter.

     * @return string
     */
    public static function getDefaultEncoding()
    {
       return self::_getInstance()->_defaultEncoding;
    }

    /**
     * Set default boolean operator
     *
     * @param integer $operator
     */
    public static function setDefaultOperator($operator)
    {
        self::_getInstance()->_defaultOperator = $operator;
    }

    /**
     * Get default boolean operator
     *
     * @return integer
     */
    public static function getDefaultOperator()
    {
        return self::_getInstance()->_defaultOperator;
    }

    /**
     * Turn on 'suppress query parser exceptions' mode.
     */
    public static function suppressQueryParsingExceptions()
    {
        self::_getInstance()->_suppressQueryParsingExceptions = true;
    }
    /**
     * Turn off 'suppress query parser exceptions' mode.
     */
    public static function dontSuppressQueryParsingExceptions()
    {
        self::_getInstance()->_suppressQueryParsingExceptions = false;
    }
    /**
     * Check 'suppress query parser exceptions' mode.
     * @return boolean
     */
    public static function queryParsingExceptionsSuppressed()
    {
        return self::_getInstance()->_suppressQueryParsingExceptions;
    }
    

    
    /**
     * Parses a query string
     *
     * @param string $strQuery
     * @param string $encoding
     * @return Zend_Search_Lucene_Search_Query
     * @throws Zend_Search_Lucene_Search_QueryParserException
     */
    public static function parse($strQuery, $encoding = null)
    {
        self::_getInstance();
        
        // Reset FSM if previous parse operation didn't return it into a correct state 
        self::$_instance->reset();
        
        try {
            self::$_instance->_encoding     = ($encoding !== null) ? $encoding : self::$_instance->_defaultEncoding;
            self::$_instance->_lastToken    = null;
            self::$_instance->_context      = new Zend_Search_Lucene_Search_QueryParserContext(self::$_instance->_encoding);
            self::$_instance->_contextStack = array();
            self::$_instance->_tokens       = self::$_instance->_lexer->tokenize($strQuery, self::$_instance->_encoding);
    
            // Empty query
            if (count(self::$_instance->_tokens) == 0) {
                return new Zend_Search_Lucene_Search_Query_Insignificant();
            }
    
    
            foreach (self::$_instance->_tokens as $token) {
                try {
                    self::$_instance->_currentToken = $token;
                    self::$_instance->process($token->type);
    
                    self::$_instance->_lastToken = $token;
                } catch (Exception $e) {
                    if (strpos($e->getMessage(), 'There is no any rule for') !== false) {
                        throw new Zend_Search_Lucene_Search_QueryParserException( 'Syntax error at char position ' . $token->position . '.' );
                    }
    
                    throw $e;
                }
            }
    
            if (count(self::$_instance->_contextStack) != 0) {
                throw new Zend_Search_Lucene_Search_QueryParserException('Syntax Error: mismatched parentheses, every opening must have closing.' );
            }
    
            return self::$_instance->_context->getQuery();
        } catch (Zend_Search_Lucene_Search_QueryParserException $e) {
            if (self::$_instance->_suppressQueryParsingExceptions) {
                $queryTokens = Zend_Search_Lucene_Analysis_Analyzer::getDefault()->tokenize($strQuery, self::$_instance->_encoding);

                $query = new Zend_Search_Lucene_Search_Query_MultiTerm();
                $termsSign = (self::$_instance->_defaultOperator == self::B_AND) ? true /* required term */ :
                                                                                   null /* optional term */;
                                                                                   
                foreach ($queryTokens as $token) {
                    $query->addTerm(new Zend_Search_Lucene_Index_Term($token->getTermText()), $termsSign);
                }
                
                
                return $query;
            } else {
                throw $e;
            }
        }
    }


    /*********************************************************************
     * Actions implementation
     *
     * Actions affect on recognized lexemes list
     *********************************************************************/

    /**
     * Add term to a query
     */
    public function addTermEntry()
    {
        $entry = new Zend_Search_Lucene_Search_QueryEntry_Term($this->_currentToken->text, $this->_context->getField());
        $this->_context->addEntry($entry);
    }

    /**
     * Add phrase to a query
     */
    public function addPhraseEntry()
    {
        $entry = new Zend_Search_Lucene_Search_QueryEntry_Phrase($this->_currentToken->text, $this->_context->getField());
        $this->_context->addEntry($entry);
    }

    /**
     * Set entry field
     */
    public function setField()
    {
        $this->_context->setNextEntryField($this->_currentToken->text);
    }

    /**
     * Set entry sign
     */
    public function setSign()
    {
        $this->_context->setNextEntrySign($this->_currentToken->type);
    }


    /**
     * Process fuzzy search/proximity modifier - '~'
     */
    public function processFuzzyProximityModifier()
    {
        $this->_context->processFuzzyProximityModifier();
    }

    /**
     * Process modifier parameter
     *
     * @throws Zend_Search_Lucene_Exception
     */
    public function processModifierParameter()
    {
        if ($this->_lastToken === null) {
            throw new Zend_Search_Lucene_Search_QueryParserException('Lexeme modifier parameter must follow lexeme modifier. Char position 0.' );
        }

        switch ($this->_lastToken->type) {
            case Zend_Search_Lucene_Search_QueryToken::TT_FUZZY_PROX_MARK:
                $this->_context->processFuzzyProximityModifier($this->_currentToken->text);
                break;

            case Zend_Search_Lucene_Search_QueryToken::TT_BOOSTING_MARK:
                $this->_context->boost($this->_currentToken->text);
                break;

            default:
                // It's not a user input exception
                throw new Zend_Search_Lucene_Exception('Lexeme modifier parameter must follow lexeme modifier. Char position 0.' );
        }
    }


    /**
     * Start subquery
     */
    public function subqueryStart()
    {
        $this->_contextStack[] = $this->_context;
        $this->_context        = new Zend_Search_Lucene_Search_QueryParserContext($this->_encoding, $this->_context->getField());
    }

    /**
     * End subquery
     */
    public function subqueryEnd()
    {
        if (count($this->_contextStack) == 0) {
            throw new Zend_Search_Lucene_Search_QueryParserException('Syntax Error: mismatched parentheses, every opening must have closing. Char position ' . $this->_currentToken->position . '.' );
        }

        $query          = $this->_context->getQuery();
        $this->_context = array_pop($this->_contextStack);

        $this->_context->addEntry(new Zend_Search_Lucene_Search_QueryEntry_Subquery($query));
    }

    /**
     * Process logical operator
     */
    public function logicalOperator()
    {
        $this->_context->addLogicalOperator($this->_currentToken->type);
    }

    /**
     * Process first range query term (opened interval)
     */
    public function openedRQFirstTerm()
    {
        $this->_rqFirstTerm = $this->_currentToken->text;
    }

    /**
     * Process last range query term (opened interval)
     *
     * @throws Zend_Search_Lucene_Search_QueryParserException
     */
    public function openedRQLastTerm()
    {
        $tokens = Zend_Search_Lucene_Analysis_Analyzer::getDefault()->tokenize($this->_rqFirstTerm, $this->_encoding);
        if (count($tokens) > 1) {
            throw new Zend_Search_Lucene_Search_QueryParserException('Range query boundary terms must be non-multiple word terms');
        } else if (count($tokens) == 1) {
            $from = new Zend_Search_Lucene_Index_Term(reset($tokens)->getTermText(), $this->_context->getField());
        } else {
            $from = null;
        }

        $tokens = Zend_Search_Lucene_Analysis_Analyzer::getDefault()->tokenize($this->_currentToken->text, $this->_encoding);
        if (count($tokens) > 1) {
            throw new Zend_Search_Lucene_Search_QueryParserException('Range query boundary terms must be non-multiple word terms');
        } else if (count($tokens) == 1) {
            $to = new Zend_Search_Lucene_Index_Term(reset($tokens)->getTermText(), $this->_context->getField());
        } else {
            $to = null;
        }

        if ($from === null  &&  $to === null) {
            throw new Zend_Search_Lucene_Search_QueryParserException('At least one range query boundary term must be non-empty term');
        }

        $rangeQuery = new Zend_Search_Lucene_Search_Query_Range($from, $to, false);
        $entry      = new Zend_Search_Lucene_Search_QueryEntry_Subquery($rangeQuery);
        $this->_context->addEntry($entry);
    }

    /**
     * Process first range query term (closed interval)
     */
    public function closedRQFirstTerm()
    {
        $this->_rqFirstTerm = $this->_currentToken->text;
    }

    /**
     * Process last range query term (closed interval)
     *
     * @throws Zend_Search_Lucene_Search_QueryParserException
     */
    public function closedRQLastTerm()
    {
        $tokens = Zend_Search_Lucene_Analysis_Analyzer::getDefault()->tokenize($this->_rqFirstTerm, $this->_encoding);
        if (count($tokens) > 1) {
            throw new Zend_Search_Lucene_Search_QueryParserException('Range query boundary terms must be non-multiple word terms');
        } else if (count($tokens) == 1) {
            $from = new Zend_Search_Lucene_Index_Term(reset($tokens)->getTermText(), $this->_context->getField());
        } else {
            $from = null;
        }

        $tokens = Zend_Search_Lucene_Analysis_Analyzer::getDefault()->tokenize($this->_currentToken->text, $this->_encoding);
        if (count($tokens) > 1) {
            throw new Zend_Search_Lucene_Search_QueryParserException('Range query boundary terms must be non-multiple word terms');
        } else if (count($tokens) == 1) {
            $to = new Zend_Search_Lucene_Index_Term(reset($tokens)->getTermText(), $this->_context->getField());
        } else {
            $to = null;
        }

        if ($from === null  &&  $to === null) {
            throw new Zend_Search_Lucene_Search_QueryParserException('At least one range query boundary term must be non-empty term');
        }

        $rangeQuery = new Zend_Search_Lucene_Search_Query_Range($from, $to, true);
        $entry      = new Zend_Search_Lucene_Search_QueryEntry_Subquery($rangeQuery);
        $this->_context->addEntry($entry);
    }
}