scybwdf/jieba-php

The Jieba Chinese Word Segmentation Implemented in PHP

Jieba PHP

结巴分词 PHP 实现 - The Jieba Chinese Word Segmentation Implemented in PHP
使用 PHP 7.4 中新增的 FFI 对 jieba-rs 进行了包装。

Requirement

PHP >= 7.4，并开启 FFI 扩展

Installation

You can install the package via composer:

composer require binaryoung/jieba-php

Usage

use Binaryoung\Jieba\Jieba;

var_dump(Jieba::cut('PHP是世界上最好的语言！'));

API

array cut(string $sentence, bool $hmm = true)
array cutAll(string $sentence)
array cutForSearch(string $sentence, bool $hmm = true)
array TFIDFExtract(string $sentence, int $topK = 20, array $allowedPOS = [])
array textRankExtract(string $sentence, int $topK = 20, array $allowedPOS = [])
array tokenize(string $sentence, string $mode = 'default', bool $hmm = true)
array tag(string $sentence, bool $hmm = true)
int   suggestFrequency(string $segment)
self  addWord(string $word, ?int $frequency = null, ?string $tag = null)
self  useDictionary(string $path)

Examples

see examples/example.php

composer example

Testing

composer test

Benchmark

composer bench

对比 jukuball/jieba-php，循环 50 次对围城每行文字作为一个句子进行分词，分词算法都采用 HMM 模式。

名称	耗时	单次耗时	内存占用	内存峰值
jukuball/jieba-php	51.593	1.032	493.00MB	515.03MB
binaryoung/jieba-php	8.408	0.16816	10.00MB	22.01MB
差值	↓513.59%	↓513.59%	↓4830.00%	↓2240.20%

Changelog

Please see CHANGELOG for more information on what has changed recently.

Contributing

Please see CONTRIBUTING for details.

Security

If you discover any security related issues, please email me instead of using the issue tracker.

Credits

• young

License

The MIT License (MIT). Please see License File for more information.