Ancient Tagger?

dsheroh has asked for the wisdom of the Perl Monks concerning the following question:

I'm currently working on a project for which it appears that Lingua::EN::Tagger would be quite useful. Unfortunately, this will have to run on a server which is still using perl 5.6.1, which appears to be incompatible with this module, instead complaining

Unrecognized escape \p in character class passed through before HERE m
+ark in regex m/(?<![\p << HERE {IsWord}\s'])'(?=.*\p{IsWord})/ at Lin
+gua/EN/Tagger.pm line XXX
[download]

for several different values of XXX. CPAN only shows two versions of this module, both dated 2005 and only a couple months apart.

Does anyone know of a module which provides similar functionality (i.e., extracting noun phrases from text) and will work on 5.6.1?

Comment on Ancient Tagger? Download Code

Replies are listed 'Best First'.
Re: Ancient Tagger? by Corion (Patriarch) on Nov 06, 2007 at 22:01 UTC
If you're searching for old versions of modules, http://backpan.cpan.org keeps everything that once was on CPAN. perlre tells me that `\p{PROP}` is a fancy name for a character class, and I guess that `\p{IsWord}` is identical to `\w`, so you could try to patch the module to use `\w` wherever `\p{IsWord}` is used and then see whether the tests still pass. Maybe `\p{IsWord}` only means `[a-z]` plus all those diacritics, but English won't have them, so maybe you're lucky.	[reply] [d/l] [select]
Re: Ancient Tagger? by diotalevi (Canon) on Nov 07, 2007 at 02:38 UTC
You're writing new software so you can get a new version of perl. If you install it to some place like /opt/perl-5.8.8 or similar then you won't ever conflict with /usr/bin/perl. This general strategy lets you have many perls so your legacy code won't break by changing your perl underneath it and your new code and can use recent perls. ⠤⠤ ⠙⠊⠕⠞⠁⠇⠑⠧⠊	[reply]