Re: [PECL-DEV] PHP Extension -- feedback request

[email protected] ("Cesar D. Rodas")
Newsgroups php.pecl.dev
Message-ID <[email protected]>
Hello Rasmus,


Thanks for the feedback :-)

2010/1/15 Rasmus Lerdorf <[email protected]>:
> Cesar D. Rodas wrote:
>> Hello everybody,
>>
>> In the fast few weeks I've been developing a brand new PHP Extension
>> and I'd like a feedback about it, in order to request a PECL account.
>>
>> The extension it's pretty simple, it implements a N-gram based text
>> categorization [1] class, useful to detect languages of texts with a
>> minimum supervision (it requires a simple learning from samples).
>>
>> Currently the Extension, provides two classes:
>>
>> - BaseTextCat: This class has few abstract methods, allowing
>> developers override the default behavior of the library from the PHP
>> userland.
>> - TextCat: A ready-to-use class with standard behavior (saving the
>> "knowledge" in flat files, treating texts as one character per byte).
>>
>> The Extension it's hosted currently at Github,
>> http://github.com/crodas/PHPTextCat
>
> Looks cool.  Fill out the svn account request form at
> http://us.php.net/svn-php.php and we will get you set up.
Thanks for the advice, I just did it.

>
> I did notice a few issues.  Most of your sample text has character
> encoding issues.  And your Swedish sample text is actually Swahili.
> Easy mistake to make.  Those Swahili-speaking swedes are everywhere.
The encoding issues are intentional, as I said I treat right now
one-byte per character (but it will change soon overriding the class
from PHP), I think it's useful as is, since it detect language and
encoding (i.e Spanish.UTF-8, Spanish.ISO-8859-1). And I wasn't aware
about my Swedish sample text (I speak Spanish, Portuguese and a few
English), but I'll get a sample text from Wikipedia.

>
> -Rasmus
>



-- 
Cesar D. Rodas
http://cesar.la/
Phone: +595-961-974165
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.