Pspell Functions

Introduction

These functions allow you to check the spelling of a word and offer suggestions.

Requirements

To compile PHP with pspell support, you need the aspell library, available from » http://aspell.sourceforge.net/.

Installation

If you have the libraries needed add the --with-pspell[=dir] option when compiling PHP.

Note to Win32 Users:

In order for this extension to work, there are DLL files that must be available to the Windows system PATH. See the FAQ titled "How do I add my PHP directory to the PATH on Windows" for information on how to do this. Although copying DLL files from the PHP folder into the Windows system directory also works (because the system directory is by default in the systems PATH), it is not recommended. This extension requires the following files to be in the PATH: aspell-15.dll from the bin folder of the aspell installation.

Win32 support is available only in PHP 4.3.3 and later versions. Also, at least aspell version 0.50 is required.

Runtime Configuration

This extension has no configuration directives defined in php.ini.

Resource Types

This extension has no resource types defined.

Predefined Constants

The constants below are defined by this extension, and will only be available when the extension has either been compiled into PHP or dynamically loaded at runtime.

PSPELL_FAST (integer)
PSPELL_NORMAL (integer)
PSPELL_BAD_SPELLERS (integer)
PSPELL_RUN_TOGETHER (integer)

Table of Contents

pspell_add_to_personal — Add the word to a personal wordlist
pspell_add_to_session — Add the word to the wordlist in the current session
pspell_check — Check a word
pspell_clear_session — Clear the current session
pspell_config_create — Create a config used to open a dictionary
pspell_config_data_dir — location of language data files
pspell_config_dict_dir — Location of the main word list
pspell_config_ignore — Ignore words less than N characters long
pspell_config_mode — Change the mode number of suggestions returned
pspell_config_personal — Set a file that contains personal wordlist
pspell_config_repl — Set a file that contains replacement pairs
pspell_config_runtogether — Consider run-together words as valid compounds
pspell_config_save_repl — Determine whether to save a replacement pairs list along with the wordlist
pspell_new_config — Load a new dictionary with settings based on a given config
pspell_new_personal — Load a new dictionary with personal wordlist
pspell_new — Load a new dictionary
pspell_save_wordlist — Save the personal wordlist to a file
pspell_store_replacement — Store a replacement pair for a word
pspell_suggest — Suggest spellings of a word

alan dunsmuir

There appears to be a problem currently affecting PSpell functionality on Windows machines. Currently, the compiled build of PHP5 for Windows does not by default include PSpell, and to activate these functions the Windows user has to (a) install GNU ASpell; (b) copy the file aspell-15.dll to a directory in the Windows "Path" statement (typically C:\Windows\System32); (c) install one or more ASpell dictionaries; (d) enable the PSpell extension to PHP by adding a reference to php_pspell.dll in the php.ini file, and moving the file itself to the main PHP directory. However, when this has all been done and PSpell appears to have been fully activated, there remains a problem. While all the PSpell functions work as specified when referenced from the "main" body of PHP code, if they are placed within a PHP function, the system reports that it is unable to find any word lists for the specified dictionary (whichever one is chosen). I have a sample of PSpell-using code (taken from the SAMS book PHP 5 In Practice) which works perfectly on a Unix-based Web Server, but which fails with the indicated error message on a range of Windows machines.

dylan

preg_replace_callback() is the best way I found to check words in a large text and change the HTML color of words that are not found. Here is a script I wrote which does just that, and optionally uses pspell_suggest() to auto-correct. I would love to write a better regex to ignore <text inside xml tags>, but that really isn't as easy as it sounds... <?php $pspell = pspell_new('en','canadian','','utf-8',PSPELL_FAST); function spellCheckWord($word) { global $pspell; $autocorrect = TRUE; // Take the string match from preg_replace_callback's array $word = $word[0]; // Ignore ALL CAPS if (preg_match('/^[A-Z]*$/',$word)) return $word; // Return dictionary words if (pspell_check($pspell,$word)) return $word; // Auto-correct with the first suggestion, color green if ($autocorrect && $suggestions = pspell_suggest($pspell,$word)) return ''.current($suggestions).''; // No suggestions, color red return ''.$word.''; } function spellCheck($string) { return preg_replace_callback('/\b\w+\b/','spellCheckWord',$string); } echo spellCheck('PHP is a reflecktive proegramming langwage origenaly dezigned for prodewcing dinamic waieb pagges.'); ?> OUTPUT: PHP is a reflective programming language originally designed for producing dynamic Web pages.

tidd

In response to csnyder's comment about calling aspell instead of using the aspell or pspell libs: If you are running a low traffic site this will work fine and it's a good idea. If, however, you have many customers using your spell checking script, this method does not scale well and you should consider spending the time to make the libs work properly.

andrew

If you use red hat or fedora, you may need to is use yum install pspell-devel (fixes: configure: error: Cannot find pspell)

gregory boshoff

If you are receiving the error message: PSPELL couldn't open the dictionary. reason: No word lists can be found for the language "en". Add following lines prior to calling pspell_check: $pspell_config = pspell_config_create("en"); $pspell_link = pspell_new_config($pspell_config);

jrweir _a_t_ yahoo

I had the same problems after following the detailed instructions for installing on windows but was able to solve them by doing the following: first, I copied all the files in the data directory of the aspell install (ie. C:\Program Files\Aspell\data) to a unix machine and ran dos2unix on all of the files and then copied them back. second, added this line before the pspell_new("en") call pspell_config_create("en"); hope this works for you too.

beau_scott

hehe, forgot to remove the echo $string line... whoops ;) Should be: <? /** * Checks spelling of $string. Whole phrases can be sent in, too, and each word will be checked. * Returns an associative array of mispellings and their suggested spellings * @param string $string Phrase to be checked * @return array */ function checkSpelling ( $string ) { // Make word list based word boundries $wordlist = preg_split('/\s/',$string); // Filter words $words = array(); for($i = 0; $i < count($wordlist); $i++) { $word = trim($wordlist[$i]); if(!preg_match('/[A-Za-z]/', $word)) continue; $word = preg_replace('/[^\w\']*(.+)/', '\1', $word); $word = preg_replace('/([^\W]*)[^\w\']*$/', '\1', $word); $word = trim($word); if(!in_array($word, $words, true)) $words[] = $word; } $misspelled = $return = array(); $int = pspell_new('en'); foreach ($words as $value) if (!pspell_check($int, $value)) $misspelled[] = $value; foreach ($misspelled as $value) $return[$value] = pspell_suggest($int, $value); return $return; } ?>

flint

For those will problems installing on Win32 with IIS: Symptom: browser just hangs when calling pspell function. Running PHP.exe from command line causes pspell to claim that there is a corrupt file. Running Aspell works just fine on the command line. You need new data files: You can download and get more information about it here: http://wiki.wordpress.org/SpellCheck Just replace the data directory in C:\Programs Files\Aspell\ with the new data directory you downloaded. No more problems.

10-dec-2004 09:52

For those of us in the Windows world, here is a great link on how to enable pspell for PHP on Win32: http://www.oblius.com/?.blogs.184

webmaster

Easy way to make a spellchecker: <? function spellcheck ( $string ) { $words = preg_split('/[\W]+?/',$string); $misspelled = $return = array(); $int = pspell_new('en'); foreach ($words as $value) { if (!pspell_check($int, $value)) { $misspelled[] = $value; } } foreach ($misspelled as $value) { $return[$value] = pspell_suggest($int, $value); } return $return; } ?>

jeremy hepler

compiling php with pspell on debian sarge: apt-get install aspell libaspell15 libaspell-dev libpspell-dev configure php: ./configure <your other options> --with-pspell=/usr make & install if previous tries have failed or it is not functioning properly, unpack the php source again and configure / compile in a fresh tree

sholland

Back in 10/2002 csnyder at chxo dot com wrote the first comment about the spell functions, and wrote a very sophisticated spell check with a direct call to aspell. In a similar vein, I wrote a simplified a PHP function that will spell check a string and insert tags before and after the misspelled words. Auto-correction is not offered. <? function spellcheck($string) { // Entry: $string: Text to spellcheck. // Returns: string with '' and '' inserted before and // after misspellings. $pre=''; // Inserted before each mis-spelling. $post=''; // Inserted after each mis-spelling. $string=strtr($string,"\n"," "); // Drop newlines in string. (It bothers aspell in this context.) $mistakes = `echo $string | /usr/local/bin/aspell list`; // Get list or errors. $offset=0; foreach (explode("\n",$mistakes) as $word) // Walk list, inserting $pre and $post strings. I move along and // do the insertions, keeping track of the location. A global replace // with str_replace($string,$pre.$work.$post) is problematic if the // same misspelling is found more than once. if ($word<>"") { $offset=strpos($string,$word,$offset); $string=substr_replace($string, $post, $offset+strlen($word), 0); $string=substr_replace($string, $pre, $offset, 0); $offset=$offset+strlen($word)+strlen("$pre $post"); }; return $string;}; ?> For this to work on your system, see if /usr/local/bin/aspell list runs from the shell. It needs to get input from standard input. You may not be able to run apell without the path for PHP because the PATH variable may be different in the PHP invocation from a shell invocation.

kevina

Aspell Author Here. Since Aspell 0.50, Pspell is no longer used. It is not necessary to download and install the Pspell package, only Aspell 0.50 (or better) is required. Aspell now provided a backwards compatibility interface for programs expecting the old Pspell interface (such as PHP). Even though Aspell now provided a new "Aspell" interface this, the PHP Pspell interface is the one you want to use. The Aspell PHP interface is the one used with a *very* old version of Aspell as the manual says. Sorry if this is confusing, but it is not really my fought as I have no control over PHP development. Eventually a new Aspell interface should be provided for PHP which uses the new Aspell interface provided with Aspell 0.50 but for now the Pspell interface will work just fine. Note: If you wish to use an older version of Aspell (_before_ 0.50) than both Aspell and Pspell are required.

csnyder

As an alternative to mucking about with the compiled-in aspell/pspell functions, which only check word by word, one could write a script that calls aspell as a shell exec to check a whole block of text: http://chxo.com/scripts/spellcheck.php (there is a view-source option) Not sure about Linux or Windows, but installing aspell on FreeBSD was trivially easy from the ports collection. No doubt such a script could be easily modified to use ispell instead if you'd rather.

abhishek dot ratani

http://www.zend.com/zend/spotlight/spellchecking.php#Heading2 Great tutorial on spell checking. Check it out!