A PHP extension converting Chinese characters to Pinyin.

一个来自百度的汉字转拼音PHP扩展,其他的汉字转拼音方案存在两个问题:

1. 可转的汉字数有限,几千个左右
2. 不能解决多音字问题

Currently you have two ways to use php-pinyin. One depends on PHP-CPP, while another one is plain php extenstion which works with php 7.x. (For php 5.x support, please checkout the branch legacy)

Main improvements:

• Depend PHP-CPP, an awesome library which wrapper Zend Engine with friendly api
• Support PHP 7
• This time we support UTF-8 and GBK encoding
• Add ini_setting (pinyin.dict_path and pinyin.dict_tone), you shoud not loadDict yourself.

1. Install PHP-CPP or its LEGACY Version. Before that, you need to change the Makefile,,, because PHP-CPP was written with C++11, but libpinyin was written with C++98,,, So you should build PHP-CPP with -D_GLIBCXX_USE_CXX11_ABI=0 option, which means "Do not use Cxx11's Application Binary Interface"
2. cd /path/to/php-pinyin/cpp-ext
3. make
4. make install

This is upgraded from old php-pinyin for php 5.x.

1. cd /path/to/php-pinyin/ext
2. /path/to/php/bin/phpize
3. ./configure --with-php-config=/path/to/php/bin/php-config --with-baidu-pinyin=/path/to/pinyin
4. make
5. make install

Here /path/to/pinyin is the directory where you copied libpinyin to.

$obj = new Pinyin();
// UTF-8
var_dump($obj->convert("重庆重量"));
var_dump($obj->multiConvert(array("重庆南京市长江大桥财务会议会计")));
// GBK
var_dump($obj->multiConvert(array(iconv("UTF-8", "GBK", "重庆"), iconv("UTF-8", "GBK", "重量"))));

Results will be:

string(22) "chong'qing'zhong'liang"
array(1) {
 [0] =>
 string(65) "chong'qing'nan'jing'shi'chang'jiang'da'qiao'cai'wu'hui'yi'kuai'ji"
}
array(2) {
 [0] =>
 string(10) "chong'qing"
 [1] =>
 string(11) "zhong'liang"
}
array(1) {
 [0] =>
 string(29) "zhong'hua'ren'min'gong'he'guo"
}

If you want to get the Abbr. of the whole pinyin-string, you can simply do this:

echo preg_replace("/\'([a-zA-Z])[0-9a-zA-Z]*/e", "strtoupper('1ドル')", "'".$py_string);

This lib only support Chinese characters and english letters, or else it will return false. So you can write a safeConvert function to avoid this.

$p = new Pinyin();
function safeConvert($word, $pyOnly = true) {
 global $p;
 // UTF-8 regex for Chinese
 $result = preg_match_all("/([\x{4e00}-\x{9fa5}]+)/iu", $word, $matches);
 if(!$result) {
 throw new \Exception("No Chinese characters in word");
 }
 $pys = $p->multiConvert($matches[1]);
 if($pyOnly == true) {
 return implode("'", $pys);
 } else {
 return str_replace($matches[1], $pys, $word);
 }
}

If you want to customize dict-files yourself and then convert them to binary-format again, do it like this:

$result = $obj->generateDict("/home/work/local/pinyin/dict/dict.txt", "/home/work/tmp/dict.dat");
if($result) echo "Generate complete";

Issues and contributions are welcome.

Thank you!