分词接口：
post http://host:post/segment

参数：
text：要拆分的文本
multi：是否复合分词，默认不复合分词
top：返回TF-IDF权重最高的前N个词，默认返回所有

如：
curl -d"text=多发中国掌上中国移动是个混蛋&multi=1" http://172.16.33.238:9090/segment

返回：
{"error":null,
"words":[{"word":"中国","weight":12.52,"times":2},
{"word":"多发","weight":6.93,"times":1},
{"word":"混蛋","weight":6.44,"times":1},
{"word":"中国移动","weight":5.91,"times":1},
{"word":"掌上","weight":5.79,"times":1},
{"word":"移动","weight":4.76,"times":1}]}