source: trunk/lib/SpellCheck.inc.php @ 503

<?php
/**
 * The Strangecode Codebase - a general application development framework for PHP
 * For details visit the project site: <http://trac.strangecode.com/codebase/>
 * Copyright 2001-2012 Strangecode, LLC
 *
 * This file is part of The Strangecode Codebase.
 *
 * The Strangecode Codebase 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 3 of the License, or (at your option)
 * any later version.
 *
 * The Strangecode Codebase 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.
 *
 * You should have received a copy of the GNU General Public License along with
 * The Strangecode Codebase. If not, see <http://www.gnu.org/licenses/>.
 */

/**
 * SpellCheck.inc.php
 *
 * Interface to PHP's pspell functions.
 *
 * @author  Quinn Comendant <quinn@strangecode.com>
 * @version 1.1
 */

/* Implementation example:
--------------------------------------------------------------------------------
include_once dirname(__FILE__) . '/_config.inc.php';
include 'codebase/lib/SpellCheck.inc.php';

// Instantiate with parameters. In this example we'll set the language and the path to the personal wordlist file.
$spell = new SpellCheck(array(
    'language' => 'en',
    'personal_wordlist' => '/tmp/my_custom_dict'
));

// Just for the heck of it add a new word to persistent personal wordlist file.
$spell->add('mealworm');

$text_to_check = 'donky rinds taste like mealworm paste';

if (!$spell->checkString($text_to_check)) {
    $suggestions = $spell->getStringSuggestions($text_to_check);
    echo 'Spelling errors! Here are suggested alternatives:';
    print_r($suggestions);
} else {
    echo 'No spelling errors';
}

// Save added words to persistent custom wordlist file.
$spell->save();
--------------------------------------------------------------------------------
*/

class SpellCheck
{

    protected $_params = array(
        'language' => 'en',
        'personal_wordlist' => '', // Text file to save custom words to.
        'skip_length' => 3, // Words with this many chars or less will not be checked.
        'mode' => PSPELL_NORMAL, // PSPELL_FAST, PSPELL_NORMAL, or PSPELL_BAD_SPELLERS.
        'highlight_start' => '<strong style="color:red;">',
        'highlight_end' => '</strong>',
    );

    protected $_pspell_cfg_handle;
    protected $_pspell_handle;
    protected $_use_personal_wordlist = false;
    protected $_errors = array();

    /**
     * Constructor.
     *
     * @param  array    $params     Array of parameters (key => val pairs).
     */
    public function __construct($params)
    {
        $app =& App::getInstance();

        if (!extension_loaded('pspell')) {
            trigger_error('Pspell module not installed', E_USER_ERROR);
        }

        if (!is_array($params) || empty($params)) {
            trigger_error('SpellCheck parameters not set properly', E_USER_ERROR);
        }

        $this->setParam($params);

        $this->_pspell_cfg_handle = pspell_config_create($this->getParam('language'));

        pspell_config_ignore($this->_pspell_cfg_handle, $this->getParam('skip_length'));
        pspell_config_mode($this->_pspell_cfg_handle, $this->getParam('mode'));

        if ('' != $this->getParam('personal_wordlist')) {
            if (!is_writable(dirname($this->getParam('personal_wordlist'))) || !is_writable($this->getParam('personal_wordlist'))) {
                $app->logMsg(sprintf('Personal wordlist file not writable: %s', $this->getParam('personal_wordlist')), LOG_WARNING, __FILE__, __LINE__);
            } else {
                pspell_config_personal($this->_pspell_cfg_handle, $this->getParam('personal_wordlist'));
                $this->_use_personal_wordlist = true;
                $app->logMsg(sprintf('Using personal wordlist: %s', $this->getParam('personal_wordlist')), LOG_DEBUG, __FILE__, __LINE__);
            }
        }

        $this->_pspell_handle = pspell_new_config($this->_pspell_cfg_handle);
    }

    /**
     * Set (or overwrite existing) parameters by passing an array of new parameters.
     *
     * @access public
     * @param  array    $params     Array of parameters (key => val pairs).
     */
    public function setParam($params)
    {
        $app =& App::getInstance();

        if (isset($params) && is_array($params)) {
            // Merge new parameters with old overriding only those passed.
            $this->_params = array_merge($this->_params, $params);
        } else {
            $app->logMsg(sprintf('Parameters are not an array: %s', $params), LOG_ERR, __FILE__, __LINE__);
        }
    }

    /**
     * Return the value of a parameter, if it exists.
     *
     * @access public
     * @param string $param        Which parameter to return.
     * @return mixed               Configured parameter value.
     */
    public function getParam($param)
    {
        $app =& App::getInstance();

        if (array_key_exists($param, $this->_params)) {
            return $this->_params[$param];
        } else {
            $app->logMsg(sprintf('Parameter is not set: %s', $param), LOG_DEBUG, __FILE__, __LINE__);
            return null;
        }
    }

    /**
     * Check whether any errors have been triggered.
     *
     * @return bool   True if any errors were found, false otherwise.
     */
    public function anyErrors()
    {
        return (sizeof($this->_errors) > 0);
    }

    /**
     * Reset the error list.
     */
    public function resetErrorList()
    {
        $this->_errors = array();
    }

    /**
     * Check one word.
     *
     * @access  public
     * @param   string  $word
     * @return  bool    True if word is correct.
     * @author  Quinn Comendant <quinn@strangecode.com>
     * @version 1.0
     * @since   09 Jun 2005 18:23:51
     */
    public function check($word)
    {
        if (pspell_check($this->_pspell_handle, $word)) {
            return true;
        } else {
            $this->_errors[] = $word;
            return false;
        }
    }

    /**
     * Suggest the correct spelling for one misspelled word.
     *
     * @access  public
     * @param   string  $word
     * @return  array   Word suggestions.
     * @author  Quinn Comendant <quinn@strangecode.com>
     * @version 1.0
     * @since   09 Jun 2005 18:23:51
     */
    public function suggest($word)
    {
        return pspell_suggest($this->_pspell_handle, $word);
    }

    /**
     * Add a word to a personal list.
     *
     * @access  public
     * @param   string  $word
     * @return  array   Word suggestions.
     * @author  Quinn Comendant <quinn@strangecode.com>
     * @version 1.0
     * @since   09 Jun 2005 18:23:51
     */
    public function add($word)
    {
        $app =& App::getInstance();

        if ($this->_use_personal_wordlist) {
            if (pspell_add_to_personal($this->_pspell_handle, $word)) {
                $app->logMsg(sprintf('Added "%s" to personal wordlist: %s', $word, $this->getParam('personal_wordlist')), LOG_DEBUG, __FILE__, __LINE__);
                return true;
            } else {
                $app->logMsg(sprintf('Failed adding "%s" to personal wordlist: %s', $word, $this->getParam('personal_wordlist')), LOG_WARNING, __FILE__, __LINE__);
                return false;
            }
        }
    }

    /**
     * Save personal list to file.
     *
     * @access  public
     * @param   string  $word
     * @return  array   Word suggestions.
     * @author  Quinn Comendant <quinn@strangecode.com>
     * @version 1.0
     * @since   09 Jun 2005 18:23:51
     */
    public function save()
    {
        $app =& App::getInstance();

        if ($this->_use_personal_wordlist) {
            if (pspell_save_wordlist($this->_pspell_handle)) {
                $app->logMsg(sprintf('Saved personal wordlist: %s', $this->getParam('personal_wordlist')), LOG_DEBUG, __FILE__, __LINE__);
                return true;
            } else {
                $app->logMsg(sprintf('Failed saving personal wordlist: %s', $this->getParam('personal_wordlist')), LOG_ERR, __FILE__, __LINE__);
                return false;
            }
        }
    }

    /**
     * Returns an array of suggested words for each misspelled word in the given text.
     * The first word of the returned array is the (possibly) misspelled word.
     *
     * @access  public
     * @param   string  $string String to get suggestions for.
     * @return  mixed   Array of suggested words or false if none.
     * @author  Quinn Comendant <quinn@strangecode.com>
     * @version 1.0
     * @since   09 Jun 2005 21:29:49
     */
    public function getStringSuggestions($string)
    {
        $corrections = array();
        // Split words on punctuation except apostrophes (this regex is used in several places in this class).
        // http://stackoverflow.com/questions/790596/split-a-text-into-single-words
        $words = preg_split("/((?:^\p{P}+)|(?:\p{P}*\s+\p{P}*)|[\p{Pd}—–-]+|(?:\p{P}+$))/", $string, -1, PREG_SPLIT_DELIM_CAPTURE);
        if (is_array($words) && !empty($words)) {
            // Remove non-word elements.
            $words = preg_grep('/\w+/', $words);
            $words = array_map('strip_tags', $words);
            foreach ($words as $i => $word) {
                if (!$this->check($word)) {
                    $corrections[$i] = $this->suggest($word);
                    // Keep the original spelling as one of the suggestions.
                    array_unshift($corrections[$i], $word);
                    array_unique($corrections[$i]);
                }
            }
        }
        if (is_array($corrections) && !empty($corrections)) {
            return $corrections;
        } else {
            return false;
        }
    }

    /**
     * Checks all words in a given string.
     *
     * @access  public
     * @param   string  $string     String to check.
     * @return  void
     * @author  Quinn Comendant <quinn@strangecode.com>
     * @version 1.0
     * @since   09 Jun 2005 22:11:27
     */
    public function checkString($string)
    {
        $errors = array();
        // Split words on punctuation except apostrophes (this regex is used in several places in this class).
        // http://stackoverflow.com/questions/790596/split-a-text-into-single-words
        $words = preg_split("/((?:^\p{P}+)|(?:\p{P}*\s+\p{P}*)|[\p{Pd}—–-]+|(?:\p{P}+$))/", $string, -1, PREG_SPLIT_DELIM_CAPTURE);
        if (is_array($words) && !empty($words)) {
            // Remove non-word elements.
            $words = preg_grep('/\w+/', $words);
            $words = array_map('strip_tags', $words);
            foreach ($words as $i => $word) {
                if (!$this->check($word)) {
                    $errors[] = $word;
                }
            }
        }
        if (empty($errors)) {
            return true;
        } else {
            $this->_errors = $errors + $this->_errors;
            return false;
        }
    }

    /**
     * Returns a given string with misspelled words highlighted.
     *
     * @access  public
     * @param   string  $string     Text to highlight.
     * @return  string  Highlighted text.
     * @author  Quinn Comendant <quinn@strangecode.com>
     * @version 1.0
     * @since   09 Jun 2005 21:29:49
     */
    public function getStringHighlighted($string, $show_footnote=false)
    {
        // Split words on punctuation except apostrophes (this regex is used in several places in this class).
        // http://stackoverflow.com/questions/790596/split-a-text-into-single-words
        $words = preg_split("/((?:^\p{P}+)|(?:\p{P}*\s+\p{P}*)|[\p{Pd}—–-]+|(?:\p{P}+$))/", $string, -1, PREG_SPLIT_DELIM_CAPTURE);
        $cnt = 0;
        if (is_array($words) && !empty($words)) {
            $words = preg_grep('/\w+/', $words);
            $words = array_map('strip_tags', $words);
            foreach ($words as $i => $word) {
                if (!$this->check($word)) {
                    $footnote = $show_footnote ? '<sup style="color:#999;">' . ++$cnt . '</sup>' : '';
                    $words[$i] = $this->getParam('highlight_start') . $word . $this->getParam('highlight_end') . $footnote;
                    $string = preg_replace(sprintf('/\b%s\b/', preg_quote($word, '/')), $words[$i], $string);
                }
            }
        }
        return $string;
    }

    /**
     * Prints the HTML for correcting all misspellings found in the text of one $_FORM element.
     *
     * @access  public
     * @param   string  $form_name  Name of the form to check.
     * @return  void
     * @author  Quinn Comendant <quinn@strangecode.com>
     * @version 1.0
     * @since   09 Jun 2005 21:29:49
     */
    public function printCorrectionForm($form_name)
    {
        ?>
        <input name="<?php echo $form_name ?>" type="hidden" value="<?php echo oTxt(getFormData($form_name)) ?>" />
        <?php

        $form_words = $this->getStringSuggestions(getFormData($form_name));
        if (is_array($form_words) && !empty($form_words)) {
            ?><ol><?php
            foreach ($form_words as $i => $words) {
                ?>
                <li>
                <label style="color:#999;"><sub style="vertical-align:text-top;"><?php echo $j++; ?></sub></label>
                <select name="spelling_suggestions[<?php echo $form_name ?>][<?php echo $i ?>]" onchange="document.forms[0].elements['spelling_corrections[<?php echo $form_name ?>][<?php echo $i ?>]'].value = this.value;">
                <?php $original_word = array_shift($words); ?>
                <option value="<?php echo $original_word ?>">(<?php echo $original_word ?>)</option>
                <?php

                foreach ($words as $suggestion) {
                    ?>
                    <option value="<?php echo $suggestion ?>"><?php echo $suggestion ?></option>
                    <?php
                }

                ?>
                </select>
                <input type="text" name="spelling_corrections[<?php echo $form_name ?>][<?php echo $i ?>]" value="<?php echo oTxt($original_word) ?>" size="20">
                <?php if ($this->_use_personal_wordlist) { ?>
                <input name="save_to_personal_wordlist[]" type="checkbox" value="<?php echo $i ?>" /><?php echo _("Learn spelling") ?>
                <?php } ?>
                </li>
                <?php
            }
            ?></ol><?php
        }
    }

    /**
     * Tests if any form spelling corrections have been submitted.
     *
     * @access  public
     * @return  bool    True if form spelling has been checked.
     * @author  Quinn Comendant <quinn@strangecode.com>
     * @version 1.0
     * @since   09 Jun 2005 23:15:35
     */
    public function anyFormCorrections()
    {
        return (false !== getFormData('spelling_suggestions', false)) || (false !== getFormData('spelling_corrections', false));
    }

    /**
     * Replace the misspelled words in the text of a specified form with the corrections.
     *
     * @access  public
     * @param   string  $form_name      Name of form to apply corrections to.
     * @return  string  Corrected form text.
     * @author  Quinn Comendant <quinn@strangecode.com>
     * @version 1.0
     * @since   09 Jun 2005 23:18:51
     */
    public function applyFormCorrections($form_name)
    {
        // Split words on punctuation except apostrophes (this regex is used in several places in this class).
        // http://stackoverflow.com/questions/790596/split-a-text-into-single-words
        $form_words = preg_split("/((?:^\p{P}+)|(?:\p{P}*\s+\p{P}*)|[\p{Pd}—–-]+|(?:\p{P}+$))/", getFormData($form_name), -1, PREG_SPLIT_DELIM_CAPTURE);
        $suggestions = getFormData('spelling_suggestions');
        $corrections = getFormData('spelling_corrections');

        $form_words = array_diff($corrections[$form_name], array('')) + $suggestions[$form_name] + $form_words;
        ksort($form_words);

        if ($this->_use_personal_wordlist) {
            $save_to_personal_wordlist = getFormData('save_to_personal_wordlist');
            if (is_array($save_to_personal_wordlist) && !empty($save_to_personal_wordlist)) {
                foreach ($save_to_personal_wordlist as $cust) {
                    $this->add($form_words[$cust]);
                }
            }
            $this->save();
        }

        if (is_array($form_words) && !empty($form_words)) {
            return join('', $form_words);
        } else {
            return getFormData($form_name);
        }
    }

} // End.