Counting Chinese Words

[eyrique]

Hi,
 
I had written a function to calculate total words in a string. For example:
 
祝你 Happy Birthday - considered as 4 words
祝你生日快樂 - considered as 6 words
Happy Birthday 帥哥 2013 - considered as 5 words
Happy Birthday 09/09/13 - considered as 3 words
 
The problem is that I'd tested it on localhost (using AppServ Windows 7) and it works perfectly. But when I upload it to the server, the chinese words are calculated wrongly, example:
 
祝你 Happy Birthday - become as 3 words
祝你生日快樂 - become as 1 word
Happy Birthday 帥哥 2013 - become as 4 words
 
Not sure what's wrong with it.
 
Can someone help on this?
 
Here's my code:
 

function count_total_word($txt){
		
		$total = count(preg_split('~[\p{Z}\p{P}]+~u', $txt, null, PREG_SPLIT_NO_EMPTY)) + 1; // Count Words
		$total -= count(preg_split('~[/]+~u', $txt, null, PREG_SPLIT_NO_EMPTY)); // Ignore "/"
		return $total;
	} 

 
Localhost PHP: 5.2.6

Server PHP: 5.2.17
 
Thanks

Edited October 3, 2013 by eyrique

[.josh]

Sounds like on your server PCRE is not compiled with "--enable-unicode-properties" enabled.

 

edit: and possibly also "--enable-utf8". Well, you aren't getting errors thrown at you so you prolly already have this one enabled.

 

Check out this article: http://chrisjean.com/2009/01/31/unicode-support-on-centos-52-with-php-and-pcre/

[jazzman1]

Are you sure, that this string "祝你生日快樂" contains itself 6 words? Is there letters in this language like in European languages?

Instead using PCRE (Perl Compatible Regular Expressions) you have to consider using multibyte character encoding schemes and some multibyte string functions in php.

Have a look this example:

$str = "祝你生日快樂";
echo strlen($str); // 18
echo '<br />';
echo mb_strlen($str, 'utf8'); // 6

What's wrong and right you should tell us

[eyrique]

Sounds like on your server PCRE is not compiled with "--enable-unicode-properties" enabled.

 

edit: and possibly also "--enable-utf8". Well, you aren't getting errors thrown at you so you prolly already have this one enabled.

 

Check out this article: http://chrisjean.com/2009/01/31/unicode-support-on-centos-52-with-php-and-pcre/

 

I check on my server

$ pcretest -C

 

and got this following output:

 

PCRE version 6.6 06-Feb-2006

Compiled with

  UTF-8 support

  Unicode properties support

  Newline character is LF

  Internal link size = 2

  POSIX malloc threshold = 10

  Default match limit = 10000000

  Default recursion depth limit = 10000000

  Match recursion uses stack

 

I suppose this is unicode-properties enabled?

[eyrique]

Are you sure, that this string "祝你生日快樂" contains itself 6 words? Is there letters in this language like in European languages?

Instead using PCRE (Perl Compatible Regular Expressions) you have to consider using multibyte character encoding schemes and some multibyte string functions in php.

Have a look this example:

$str = "祝你生日快樂";
echo strlen($str); // 18
echo '<br />';
echo mb_strlen($str, 'utf8'); // 6

What's wrong and right you should tell us

 

Technically this string should be considered as 6 character instead of 6 words, but the Chinese don't count it that way. 

 

I'd tried mb_strlen() but that work for English words as this function is calculating the characters.

 

Here's what we are looking for:

祝你生日快樂 - 6 words

Happy Birthday - 2 words

[jazzman1]

So, if I understand you correctly there is no letters in Chinese language, right?

 

Have you ever checked these multibyte string functions?

 

You can apply a rule if the string contains itself EN characters use str_word_count() if they are Chinese use mb_strlen(). 

<?php

$str_en = "Happy Birdday";

$str_ch = '祝你生日快樂';

echo mb_strlen($str_c,'utf8'); // 6 Chinese words (characters)

echo str_word_count($str_e); // two EN words

Edited October 4, 2013 by jazzman1

[eyrique]

 

So, if I understand you correctly there is no letters in Chinese language, right?

 

Have you ever checked these multibyte string functions? There is

 

You can apply a rule if the string contains itself EN characters use str_word_count() if they are Chinese use mb_strlen(). 

<?php

$str_en = "Happy Birdday";

$str_ch = '祝你生日快樂';

echo mb_strlen($str_c,'utf8'); // 6 Chinese words (characters)

echo str_word_count($str_e); // two EN words

 

The string might contain combination of chinese & english words, or just a single language.

 

Eg: 

祝你生日快樂 (6 words)

祝你 Happy Birthday (4 words)

Happy Birthday (2 words)

[jazzman1]

Then count (gather) the result of Chinese and EN words and display the output!

Edited October 4, 2013 by jazzman1

[eyrique]

Then count (gather) the result of Chinese and EN words and display the output!

 

how do you do that?

[jazzman1]

Both functions return integers!

<?php

$str_en = "Happy Birdday";

$str_ch = '祝你生日快樂';

var_dump(mb_strlen($str_ch,'utf8')); // int(6)

var_dump(str_word_count($str_en)); // int(2)

echo mb_strlen($str_ch,'utf8') + str_word_count($str_en); // 8

[eyrique]

Thanks~ got it