支持primary_keys属性建立search索引进行分词搜索

### Expected behavior 期望表现
能够使用分词搜索primary_keys属性。

### Actual behavior 实际表现
primary_keys无法建立search索引，无法分词搜索。

### Steps to reproduce the problem 复现步骤
1、查看domain顶点
http://127.0.0.1:9000/graphs/hugegraph/schema/vertexlabels/domain
{
"id": 3,
"name": "domain",
"id_strategy": "PRIMARY_KEY",
"primary_keys":[
"domain_name"
],
properties":[
"domain_name"]
}

2、在studio中执行 g.V().hasLabel('domain').has('domain_name',Text.contains('baidu')) 时报错：Error! Method userpropValues(List) only used for secondary index, relation must be EQ, but got TEXT_CONTAINS

3、建立search索引失败：
http://127.0.0.1:9000/graphs/hugegraph/schema/indexlabels
{
"name": "dns_domain_name",
"base_type": "VERTEX_LABEL",
"base_value": "domain",
"index_type": "SEARCH",
"fields": [
"domain_name"
]
}

{
"exception": "class java.lang.IllegalArgumentException",
"message": "No need to build index on properties [domain_name], because they contains all primary keys [domain_name] for vertex label 'domain'",
"cause": ""
}

### Specifications of environment 环境信息
- hugegraph version: "versions":{
"version": "v1",
"core": "0.9.2.0",
"gremlin": "3.2.5",
"api": "0.38.0.0"
}
- operating system: {like centos 7 x64
- hugegraph backend: rocksdb