Text::Aspell - Perl interface to the GNU Aspell library

    use Text::Aspell;
    my $speller = Text::Aspell->new;

    die unless $speller;

    # Set some options
    $speller->set_option('lang','en_US');
    $speller->set_option('sug-mode','fast');

    # check a word
    print $speller->check( $word )
          ? "$word found\n"
          : "$word not found!\n";

    # lookup up words
    my @suggestions = $speller->suggest( $misspelled );

    # lookup config options
    my $language = $speller->get_option('lang');
    print $speller->errstr unless defined $language;

    # fetch a config item that is a list
    my @sgml_extensions = $speller->get_option_as_list('sgml-extension');

    # fetch the configuration keys and their default settings
    my $options = $speller->fetch_option_keys;

    # or dump config settings to STDOUT
    $speller->print_config || $speller->errstr;

    # What dictionaries are installed as simple strings
    my @dicts = $speller->list_dictionaries;

    # or as an array of hashes
    @dicts = $speller->dictionary_info;
    print Data::Dumper::Dumper( \@dicts );

Here's an example how to create and use your own word list

Create a dictionary:

    $ aspell --lang=en create master ./dictionary.local < space_separated_word_list

Then in your code:

    use Text::Aspell;
    my $speller = Text::Aspell->new;
    die unless $speller;
    $speller->set_option('master','./dictionary.local');
    # check a word
    print $speller->check( $word )
          ? "$word found\n"
          : "$word not found!\n";

This module provides a Perl interface to the GNU Aspell library. The GNU Aspell library provides access to system spelling libraries, including a spell checker. This module is to meet the need of looking up many words, one at a time, in a single session.

This is a Perl xs interface which should provide good performance compared to forking the aspell program for every word.

For example, a tiny run of about 400 word lookups resulted in:

    aspell-fast: 18.44 seconds
    pspell-fast:  0.63 seconds

Where aspell-fast was forking (about 4 words per fork) and pspell-fast was using this module[1] (fast refers to the spelling mode used).

The GNU C interface is described at:

  http://savannah.gnu.org/download/aspell/manual/user/6_Writing.html

It's worth looking over the way config and speller (manager) objects are created when using the Aspell C API as some of that is hidden in the Text::Aspell module. In this discussion the term speller object means the internal Aspell object, not the value returned from Text::Aspell->new.

For example, with Text::Aspell you do not have to explicitly create a speller object. The speller (manager) object is created automatically the first time you call suggest() or check().

Note also that once the speller object is created some (all?) config options cannot be changed. For example, setting configuration options such as lang are what determine what dictionary Aspell will use. Once the speller object is created that dictionary will be used. I.e. setting lang after the speller object is created will have no effect.

[1] Actually, running Text::Pspell which was the predecessor to this module.

You must have installed GNU Aspell version 0.50.1 or higher on your system. Download from:

    Aspell: http://aspell.net

The following methods are available:

$speller = Text::Aspell->new;

Creates a new speller object. New does not take any parameters (future version may allow options set by passing in a hash reference of options and value pairs). Returns CWundef if the object could not be created, which is unlikely. Internally, new() creates an object to store Aspell structures (AspellConfig, AspellSpeller, and a space for an error string and then calls new_aspell_config(); Sets the configuration option CW$option_name to the value of CW$value. Returns CWundef on error, and the error message can be printed with CW$speller->errstr. You should set configuration options before calling the CW$speller->create_speller method. See the GNU Aspell documentation for the available configuration settings and how (and when) they may be used.

$speller->remove_option($option_name);

Removes (sets to the default value) the configuration option specified by CW$option_name. Returns CWundef on error, and the error message can be printed with CW$speller->errstr. You may only set configuration options before calling the CW$speller->create_speller method. Returns the current setting for the given configuration option. The values are strings. For configuration options that are lists used the CWget_option_as_list() method. Returns CWundef on error, and the error message can be printed with CW$speller->errstr. Note that this may return different results depending on if it's called before or after CW$speller->create_speller is called. Returns an array of list items for the given option. Use this method to fetch configuration values that are of type list. Returns CWundef on error, and the error message can be printed with CW$speller->errstr. Note that this may return different results depending on if it's called before or after CW$speller->create_speller is called. Returns a hash of hashes. The keys are the possible configuration options and the values is a hash with keys of:

    desc    : A short description of the option
    default : The default value for this option
    type    : The data type of option (see aspell.h)

$speller->print_config;

Prints the current configuration to STDOUT. Useful for debugging. Note that this will return different results depending on if it's called before or after CW$speller->create_speller is called.

$speller->errstr;

Returns the error string from the last error. Check the previous call for an CWundef return value before calling this method Returns the error number from the last error. Some errors may only set the error string ($speller->errstr) on errors, so it's best to check use the errstr method over this method. Checks if a word is found in the dictionary. Returns true if the word is found in the dictionary, false but defined if the word is not in the dictionary. Returns CWundef on error, and the error message can be printed with CW$speller->errstr. This calls CW$speller->create_speller if the speller has not been created by an explicit call to CW$speller->create_speller. Returns an array of word suggestions for the specified word. The words are returned with the best guesses at the start of the list.

$speller->create_speller;

This method is normally not called by your program. It is called automatically the first time CW$speller->check() or CW$speller->suggest() is called to create a spelling speller. You might want to call this when your program first starts up to make the first access a bit faster, or if you need to read back configuration settings before looking up words. The creation of the speller builds a configuration profile in the speller structure. Results from calling print_config() and get_option() will change after calling create_speller(). In general, it's best to read config settings back after calling create_speller() or after calling spell() or suggest(). Returns CWundef on error, and the error message can be printed with CW$speller->errstr.

$speller->add_to_session($word)

$speller->add_to_personal($word)

Adds a word to the session or personal word lists. Words added will be offered as suggestions. This method can be used to instruct the speller which word you used as a replacement for a misspelled word. This allows the speller to offer up the replacement next time the word is misspelled. See section 6.3 of the GNU Aspell documentation for a better description.

$speller->save_all_word_lists;

Writes any pending word lists to disk.

$speller->clear_session;

Clears the current session word list. This returns an array of installed dictionary files. Each is a single string formatted as:

    [name]:[code]:[jargon]:[size]:[module]

Name and code will often be the same, but name is the complete name of the dictionary which can be used to directly select a dictionary, and code is the language/region code only. Like the CWlist_dictionaries() method, this method returns an array of hash references. For example, an entry for a dictionary might have the following hash reference:

    {
        'module' => 'default',
        'code' => 'en_US',
        'size' => 60,
        'jargon' => 'w-accents',
        'name' => 'en_US-w-accents'
    },

Not all hash keys will be available for every dictionary (e.g. the dictionary may not have a jargon key).

Text::Aspell works with GNU Aspell and is a replacement for the module Text::Pspell. Text::Pspell is no longer supported.

Upgrading should be a simple process. Only one method name has changed: CWcreate_manager is now called CWcreate_speller. Code designed to use the old Text::Pspell module may not even call the CWcreate_manager method so this may not be an issue.

The CWlanguage_tag configuration setting is now called CWlang.

Diffs for code that uses Text::Pspell might look like:

    -    use Text::Pspell;
    +    use Text::Aspell;

    -    $speller = Text::Pspell->new;
    +    $speller = Text::Aspell->new;

    -    $speller->create_manager || die "Failed to create speller: " . $speller->errstr;
    +    $speller->create_speller || die "Failed to create speller: " . $speller->errstr;

If you used a custom dictionary installed in non-standard location and indexed the dictionary with Aspell/Pspell .pwli files you will need to change how you access your dictionary (e.g. by setting the master configuration setting with the path to the dictionary). See the GNU Aspell documentation for details.

Probably.

This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.

Bill Moseley moseley@hank.org.

This module is based on a perl module written by Doru Theodor Petrescu <pdoru@kappa.ro>.

Aspell is written and maintained by Kevin Atkinson.

Please see:

    http://aspell.net