This library should be updated to support PHP 7.4:

PHP Deprecated: Array and string offset access syntax with curly braces is deprecated in src/Behat/Transliterator/Transliterator.php on line 357.

This pull request updates the bank files to mirror the files in the original implementation (http://cpansearch.perl.org/src/SBURKE/Text-Unidecode-0.04/lib/Text/Unidecode/)

Is there a reason the data in the bankfiles isn't the same as in the original code? http://cpansearch.perl.org/src/SBURKE/Text-Unidecode-0.04/lib/Text/Unidecode/

Only the files x00 - x07 seem to be the same. Because of some empty files and some missing items, Chinese and Japanese characters can only be partially transliterated.

Could you release a new version with the latest commits?

We would like to use this package, but we need the following commit:

• Updated to version 1.27 of Perl Unidecode (8732e225c7a450f9ed657bd158e6487463800867)

Thank you in advance!

As addressed on #4 issue, there is a problem with Russian 'e' letter conversion to latin.

Examples would be with - тест (word), Брест (city in Belarus).

In order to fix this issue, I had to modify current test examples with Russian words.

To prove that my assertions were correct I've used online transliterator tools such as http://translit-online.ru/ & http://translate.google.com.

So phrase това е тестово заглавие with online transliterator would be tova e testovo zaglavie.

I would presume that older russian test cases were written no by russian guy :)

Thank you for reviewing my PR & if you have any questions or should do anything else - please write.

Currently the transliteration is not based on the source language of the text, but this makes a big difference in some cases. Following some examples:

'Ä' (U+00C4) transliterates to 'A' but when the source text is in German should transliterate to 'Ae'

There are many more examples: https://github.com/pimcore/pimcore/blob/master/pimcore/lib/Pimcore/Tool/Transliteration/Data/x00.php#L25

Would be cool to have this integrated.

@everzet please perform the Packagist migration to the oauth apps, or add other github admins (me and @ciaranmcnulty) as maintainers of the various (non-abandoned) Behat packages on Packagist so that we can perform it.

This class can be quite useful for standalone usage. For example:

$stringWithAccent = 'accès';
$stringWithoutAccent = \Behat\Transliterator\Transliterator::unaccent($stringWithAccent);

It can be used as a string tool and solve many tricks like this: http://stackoverflow.com/q/27680624/1731473

The current dev version fixes (at least) Japanese hiragana & katakana romanisation, the current release version (1.0.1) has issues with them.

It's been quite a while now any change hasn't been done. Would it be possible to create a new release out of master?

Cheers for that!

The current data are based on on the 0.04 release (see #2 ensuring that everything matches), which is from 2001. We should update it to the 1.23 version from 2014: http://cpansearch.perl.org/src/SBURKE/Text-Unidecode-1.23/

While updating an old project, I realized that urlize wasn't working the way it used to before.

before (long time ago): "c'est qu'il m'énerve" -> "c-est-qu-il-m-enerve"

now : "c'est qu'il m'énerve" -> "cest-quil-menerve"

// Remove apostrophes which are not used as quotes around a string
$text = preg_replace('/(\\w)\'(\w)/', '${1}${2}', $text);

it's not really a problem. but as I don't want to change the old urls of the site I'm going to turn the "'" into " " before using urlize.

But maybe there could be an option to decide if we want to make the apostrophes disappear or turn them into a dash. :wink:

If I understand the code correctly, there is unneccessary code in Transliterator::postProcessText (https://github.com/Behat/Transliterator/blob/master/src/Behat/Transliterator/Transliterator.php#L568)

The first thing the function does is lowercase the input. After that, in line 580 there is a lot of preg_replacegoing on, checking uppercase chars and finally lowercasing the whole string again.

I was wondering why that is there. I think, the code could be simplified

before

$text = strtolower(preg_replace('/[^A-Za-z0-9\/]+/', $separator,
            preg_replace('/([a-z\d])([A-Z])/', '\1_\2',
                preg_replace('/([A-Z]+)([A-Z][a-z])/', '\1_\2',
                    preg_replace('/::/', '/', $text)))));

after

$text = preg_replace('/[^a-z0-9\/]+/', $separator, preg_replace('/::/', '/', $text));

If it's not already supported it would be nice to restrict transliteration to nonlatin scripts, that is keep all accents from french or hungarian text, but transliterate russian text in cyrillyc or chineese. Essentially have ISO-8859-* as target, and not strictly ASCII.

I would like to package the Behat stack for Fedora/EPEL but Artistic 1.0 licensed software is not allowed in Fedora/EPEL. Please consider changing this library's license to the newer Artistic 2.0 license or perhaps even the MIT license like most other Behat libraries.

For reference, here are the Fedora software license lists:

• Good licenses
• Bad licenses (note this is where "Artistic 1.0" is listed for Fedora)

From http://www.gnu.org/licenses/license-list.html#ArtisticLicense :

Artistic License 1.0

We cannot say that this is a free software license because it is too vague; some passages are too clever for their own good, and their meaning is not clear. We urge you to avoid using it, except as part of the disjunctive license of Perl.