elasticsearch自定义插件

人生最遗憾的,莫过于,轻易地放弃了不该放弃的,固执地坚持了不该坚持的

Posted by yishuifengxiao on 2022-01-20

文档分析器(Analyzer)由字符过滤器( Character Filter)、分词器(Tokenizer )和分词过滤器(Token Filter)组成。

字符过滤器( Character Filter)可以有多个。在 Tokenizer 之前对原始文本进行处理,比如增加、删除或替换字符;会影响后续 Tokenizer 解析的 postion 和 offset 的位置。

自带的如下:

  • HTML strip 取出 html 标签和转换 html 实体

  • Mapping 进行字符替换操作

  • Pattern Replace 进行正则匹配替换操作

分词器(Tokenizer )有且只能有一个。将原始文本按照一定的规则切割成单词 (term or token)

  • 标准分词器:standard tokenizer

  • 关键词分词器:keyword tokenizer

  • 字母分词器:letter tokenizer

  • 小写分词器:lowercase tokenizer

  • 空白分词器:whitespace tokenizer

  • 模式分词器:pattern tokenizer

  • UAX URL 电子邮件分词器:UAX RUL email tokenizer

  • 路径层次分词器:path hierarchy tokenizer

分词过滤器(Token Filter)可以有多个。对 Tokenizer 输出的单词(term)进行增加、修改、删除等操作

自带的如下:

  • lowercase 将所有的 term 都转换为 小写
  • stop 删除所有的 stop word
  • NGram 和 Edge NGram 连词分割
  • Synon 添加近义词的 term

除了分析器之外,ES还提供了一种称为规整器(Normalizer)的文档标准化工具,该工具与文档分析器(Analyzer)最大的区别在于没有分词器,只有字符过滤器和分词过滤器,所以他能保证分析后的结果只有一个分词

一 创建并配置文件

1.1 pom.xml配置

创建一个maven项目,项目的pom的大概结构如下

<project xmlns="http://maven.apache.org/POM/4.0.0"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
<modelVersion>4.0.0</modelVersion>

<groupId>com.demo.aidoc</groupId>
<artifactId>es-plugin</artifactId>
<version>0.0.1-SNAPSHOT</version>
<packaging>jar</packaging>

<name>es-plugin</name>
<url>http://maven.apache.org</url>
<repositories>

<repository>
<id>spring</id>
<url>https://maven.aliyun.com/repository/spring</url>
<releases>
<enabled>true</enabled>
</releases>
<snapshots>
<enabled>true</enabled>
</snapshots>
</repository>

</repositories>

<properties>
<project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
<maven.compiler.target>1.8</maven.compiler.target>
<!--elasticsearch 版本要与目标版本对应上-->
<elasticsearch.version>7.16.0</elasticsearch.version>
<hanlp.version>portable-1.7.3</hanlp.version>
<maven.compiler.target>1.8</maven.compiler.target>
</properties>

<dependencies>
<dependency>
<groupId>com.hankcs</groupId>
<artifactId>hanlp</artifactId>
<version>${hanlp.version}</version>
</dependency>

<dependency>
<groupId>org.elasticsearch</groupId>
<artifactId>elasticsearch</artifactId>
<version>${elasticsearch.version}</version>
<scope>compile</scope>
</dependency>
<dependency>
<groupId>org.apache.httpcomponents</groupId>
<artifactId>httpclient</artifactId>
<version>4.5.2</version>
</dependency>
<dependency>
<groupId>junit</groupId>
<artifactId>junit</artifactId>
<version>4.12</version>
<scope>test</scope>
</dependency>
<dependency>
<groupId>com.alibaba</groupId>
<artifactId>fastjson</artifactId>
<version>1.2.79</version>
</dependency>
</dependencies>
<build>
<finalName>${project.name}</finalName>
<resources>
<resource>
<directory>src/main/resources</directory>
<filtering>false</filtering>
<excludes>
<!--<exclude>*.properties</exclude>-->
<!-- 这个文件是es插件描述文件,不用打包到jar包里面 -->
<exclude>plugin-descriptor.properties</exclude>
</excludes>
</resource>
</resources>
<plugins>
<!-- 把package的jar包在打包成zip文件,es插件是zip文件格式 -->
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-assembly-plugin</artifactId>
<version>2.6</version>
<configuration>
<appendAssemblyId>false</appendAssemblyId>
<outputDirectory>${project.build.directory}/releases/</outputDirectory>
<descriptors>
<!-- 该插件的配置文件 -->
<descriptor>${basedir}/src/main/resources/plugin.xml</descriptor>
</descriptors>
</configuration>
<executions>
<execution>
<phase>package</phase>
<goals>
<goal>single</goal>
</goals>
</execution>
</executions>
</plugin>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-compiler-plugin</artifactId>
<version>3.5.1</version>
<configuration>
<source>${maven.compiler.target}</source>
<target>${maven.compiler.target}</target>
</configuration>
</plugin>
<!-- 配置文件复制,用于把 config目录下的文件复制到target目录下 -->
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-resources-plugin</artifactId>
<version>3.2.0</version>
<configuration>
<encoding>${project.build.sourceEncoding}</encoding>
</configuration>
<executions>
<execution>
<id>copy-spring-boot-resources</id>
<!-- here the phase you need -->
<phase>validate</phase>
<goals>
<!-- 资源文件配置 -->
<goal>copy-resources</goal>
</goals>
<configuration>
<encoding>utf-8</encoding>
<outputDirectory>${basedir}/target/config</outputDirectory>
<resources>
<resource>
<directory>${basedir}/config</directory>
<includes>
<include>*</include>
</includes>
</resource>
</resources>
</configuration>
</execution>
</executions>
</plugin>
</plugins>
</build>
</project>

注意其中依赖的elasticsearch的版本要与目标elasticsearch的版本保持一致。

1.2 assembly 插件配置

文件路径src/main/assembly/plugin.xml

<?xml version="1.0"?>
<assembly>
<id>analysis-gridsum</id>
<formats>
<format>zip</format>
</formats>
<includeBaseDirectory>false</includeBaseDirectory>
<fileSets>
<fileSet>
<directory>${project.basedir}/config</directory>
<outputDirectory>config</outputDirectory>
</fileSet>
</fileSets>

<files>
<file>
<source>${project.basedir}/src/main/resources/plugin-descriptor.properties</source>
<outputDirectory/>
<filtered>true</filtered>
</file>
<file>
<source>${project.basedir}/src/main/resources/plugin-security.policy</source>
<outputDirectory/>
<filtered>true</filtered>
</file>
</files>

<dependencySets>
<dependencySet>
<outputDirectory />
<useProjectArtifact>true</useProjectArtifact>
<useTransitiveFiltering>true</useTransitiveFiltering>
<excludes>
<exclude>org.elasticsearch:elasticsearch</exclude>
</excludes>
</dependencySet>
<dependencySet>
<outputDirectory/>
<useProjectArtifact>true</useProjectArtifact>
<useTransitiveFiltering>true</useTransitiveFiltering>
<includes>
<include>org.apache.httpcomponents:httpclient</include>
</includes>
</dependencySet>
</dependencySets>
</assembly>

1.3 插件描述文件

resources目录下,plugin-descriptor.properties文件。

所有插件都必须包含一个名为 plugin-descriptor.properties 的文件。 此文件的格式在此示例中详细描述:https://www.elastic.co/guide/en/elasticsearch/plugins/master/plugin-authors.html#_plugin_descriptor_file

官方示例如下

# Elasticsearch plugin descriptor file
# This file must exist as 'plugin-descriptor.properties' inside a plugin.
#
### example plugin for "foo"
#
# foo.zip <-- zip file for the plugin, with this structure:
# |____ <arbitrary name1>.jar <-- classes, resources, dependencies
# |____ <arbitrary nameN>.jar <-- any number of jars
# |____ plugin-descriptor.properties <-- example contents below:
#
# classname=foo.bar.BazPlugin
# description=My cool plugin
# version=6.0
# elasticsearch.version=6.0
# java.version=1.8
#
### mandatory elements for all plugins:
#
# 'type': the type of this plugin. 'isolated' indicated a typical sandboxed plugin,
# whereas 'bootstrap' indicates a plugin whose jars are added to the JVM's boot
# classpath.
type=${type}
#
# 'description': simple summary of the plugin
description=${description}
#
# 'version': plugin's version
version=${version}
#
# 'name': the plugin name
name=${name}
<% if (type != "bootstrap") { %>
#
# 'classname': the name of the class to load, fully-qualified. Only applies to
# "isolated" plugins
classname=${classname}
<% } %>
#
# 'java.version': version of java the code is built against
# use the system property java.specification.version
# version string must be a sequence of nonnegative decimal integers
# separated by "."'s and may have leading zeros
java.version=${javaVersion}
#
# 'elasticsearch.version': version of elasticsearch compiled against
elasticsearch.version=${elasticsearchVersion}
### optional elements for plugins:
#
# 'extended.plugins': other plugins this plugin extends through SPI
extended.plugins=${extendedPlugins}
#
# 'has.native.controller': whether or not the plugin has a native controller
has.native.controller=${hasNativeController}
<% if (type == "bootstrap") { %>
#
# 'java.opts': any additional command line parameters to pass to the JVM when
# Elasticsearch starts. Only applies to "bootstrap" plugins.
java.opts=${javaOpts}
<% } %>
<% if (licensed) { %>
# This plugin requires that a license agreement be accepted before installation
licensed=${licensed}
<% } %>

解释如下

Element Type Description
description String simple summary of the plugin
version String plugin’s version
name String the plugin name
classname String 要加载的类的名称,完全限定。
java.version String 构建代码的java版本。 使用系统属性 java.specification.version。 版本字符串必须是由“.”分隔的非负十进制整数序列,并且可以有前导零。

请注意,只有插件根目录下的 jar 文件才会添加到插件的类路径中! 如果您需要其他资源,请将它们打包到资源 jar 中。

插件发布生命周期:您必须为每个新的 Elasticsearch 版本发布一个新版本的插件。 加载插件时会检查此版本,因此 Elasticsearch 将拒绝在存在具有不正确 elasticsearch.version 的插件时启动。

本项目中实际的描述文件如下:

description=${project.description}
version=${project.version}
name=${project.name}
classname=com.demo.aidoc.es_plugin.DemoPlugin
java.version=${maven.compiler.target}
elasticsearch.version=${elasticsearch.version}

**注意:**上述配置中的 classname 的值一定要按照实际情况修改为启动类的路径

plugin-descriptor.properties的配置是从pom.xml读取的properties配置,这样维护更方便,在打包时会替换掉占位符

1.4 权限声明文件

resources目录下,文件名为plugin-security.policy

jdk的安全策略限制,必须声明项目使用的权限

一些插件可能需要额外的安全权限。 插件可以包含可选的 plugin-security.policy 文件,其中包含用于附加权限的授权语句。 任何额外的权限都会向用户显示一个大警告,并且他们必须在交互式安装插件时确认它们。 因此,如果可能,最好避免请求任何虚假权限!

如果您使用的是 Elasticsearch Gradle 构建系统,请将这个文件放在 src/main/plugin-metadata 中,它也会在单元测试期间应用。

请记住,Java 安全模型是基于堆栈的,额外的权限只会授予插件中的 jar,因此您将围绕需要提升权限的操作编写适当的安全代码。 建议添加检查以防止非特权代码(例如脚本)获得升级的权限。 例如:

// ES permission you should check before doPrivileged() blocks
import org.elasticsearch.SpecialPermission;

SecurityManager sm = System.getSecurityManager();
if (sm != null) {
// unprivileged code such as scripts do not have SpecialPermission
sm.checkPermission(new SpecialPermission());
}
AccessController.doPrivileged(
// sensitive operation
);

以ik分词器中权限检查部分的代码为

SpecialPermission.check();
AccessController.doPrivileged((PrivilegedAction<Void>) () -> {

// 这里执行需要权限的逻辑

return null;
});

本例中的plugin-security.policy的内容如下

grant {
permission java.util.PropertyPermission "*", "read,write";
permission java.net.SocketPermission "*", "connect,resolve";// okhttp
permission java.lang.RuntimePermission "getClassLoader"; // okhttp
permission java.net.NetPermission "getProxySelector"; // okhttp
permission java.lang.reflect.ReflectPermission "suppressAccessChecks";
permission java.lang.RuntimePermission "accessDeclaredMembers";
permission java.lang.RuntimePermission "setContextClassLoader";
permission java.lang.RuntimePermission "createClassLoader";
};

二 编写插件核心代码

  • 开发插件只需要继承Plugin实现AnalysisPlugin就可以了

  • DemoTokenizer是分词器,继承Tokenizer,通过重写incrementToken方法来实现自己的分词程序

  • DemoTokenizerFactory是分词器工厂,继承AbstractTokenizerFactory,通过重写create方法返回自定义的分词器

  • DemoAnalyzer是分析器,继承Analyzer,里面需要塞一个分词器

  • DemoAnalyzerProvider是分析器提供程序,继承AbstractIndexAnalyzerProvider,通过重写get方法返回自定义分析器

  • DemoPlugin自定义插件的主要实现,继承Plugin实现AnalysisPlugin,通过重写getTokenizers将分词器工厂放入map,通过重写getAnalyzers将分析器放入map(这里的key后面会用到)

2.1 自定义Tokenizer

先实现一个自定义Tokenizer,该类是项目中重要的代码部门,主要是在这里实现自己的分词逻辑。

本例中的代码如下

package com.demo.aidoc.es_plugin.tokenizer;

import com.alibaba.fastjson.JSONArray;
import com.alibaba.fastjson.JSONObject;
import com.hankcs.hanlp.corpus.tag.Nature;
import com.hankcs.hanlp.seg.common.Term;
import org.apache.lucene.analysis.Tokenizer;
import org.apache.lucene.analysis.tokenattributes.CharTermAttribute;
import org.apache.lucene.analysis.tokenattributes.OffsetAttribute;
import org.apache.lucene.analysis.tokenattributes.PositionIncrementAttribute;
import org.apache.lucene.analysis.tokenattributes.TypeAttribute;

import java.io.*;
import java.net.URL;
import java.util.List;
import java.util.Properties;
import java.util.concurrent.atomic.AtomicInteger;
import java.util.stream.Collectors;

/**
* GridsumTokenizer是分词器,继承Tokenizer,通过重写incrementToken方法来实现自己的分词程序
*
* @author yishui
* @version 1.0.0
* @since 1.0.0
*/
public class DemoTokenizer extends Tokenizer {

private final static String PUNCTION = " -()/";
private final StringBuilder buffer = new StringBuilder();
private int suffixOffset;
private int tokenStart = 0, tokenEnd = 0;
/**
* 当前词
*/
private final CharTermAttribute termAtt = addAttribute(CharTermAttribute.class);
/**
* 偏移量
*/
private final OffsetAttribute offsetAtt = addAttribute(OffsetAttribute.class);

/**
* 距离
*/
private final PositionIncrementAttribute positionAttr = addAttribute(PositionIncrementAttribute.class);
/**
* 词性
*/
private TypeAttribute typeAtt = addAttribute(TypeAttribute.class);


private List<Term> list = null;


private final AtomicInteger offset = new AtomicInteger(0);

@Override
public final boolean incrementToken() throws IOException {
clearAttributes();
buffer.setLength(0);

if (null == list || offset.get() >= list.size()) {
termAtt.setEmpty();
return false;
}
Term term = list.get(offset.get());
tokenStart = term.offset;
termAtt.setEmpty().append(term.word);
offsetAtt.setOffset(term.offset, term.offset + term.length());
positionAttr.setPositionIncrement(term.length());
typeAtt.setType(term.nature.toString());
offset.incrementAndGet();
return true;
}

@Override
public final void end() {
final int finalOffset = correctOffset(suffixOffset);
this.offsetAtt.setOffset(finalOffset, finalOffset);
}

@Override
public void reset() throws IOException {
super.reset();
tokenStart = tokenEnd = 0;
this.offset.set(0);
// 这里要实现初始化逻辑

}
}
}

2.2 自定义Tokenizer工厂

本例中的代码如下

package com.demo.aidoc.es_plugin.factory;

import org.apache.lucene.analysis.Tokenizer;
import org.elasticsearch.common.settings.Settings;
import org.elasticsearch.env.Environment;
import org.elasticsearch.index.IndexSettings;
import org.elasticsearch.index.analysis.AbstractTokenizerFactory;

import com.demo.aidoc.es_plugin.tokenizer.DemoTokenizer;

/**
* GridsumTokenizerFactory是分词器工厂,继承AbstractTokenizerFactory,通过重写create方法返回自定义的分词器
*
* @author yishui
* @version 1.0.0
* @since 1.0.0
*/
public class DemoTokenizerFactory extends AbstractTokenizerFactory {


public DemoTokenizerFactory(IndexSettings indexSettings, Environment env, String name, Settings settings) {
super(indexSettings, settings, name);
}

@Override
public Tokenizer create() {
// TODO Auto-generated method stub
return new DemoTokenizer();
}

}

2.3 自定义Analyzer

本例中的代码如下

package com.demo.aidoc.es_plugin.analyzer;

import org.apache.lucene.analysis.Analyzer;

import com.demo.aidoc.es_plugin.tokenizer.DemoTokenizer;

/**
* GridsumAnalyzer是分析器,继承Analyzer,里面需要塞一个分词器
* @author yishui
* @version 1.0.0
* @since 1.0.0
*/
public class DemoAnalyzer extends Analyzer {

@Override
protected TokenStreamComponents createComponents(String fieldName) {
// TODO Auto-generated method stub
return new TokenStreamComponents(new DemoTokenizer());
}

}

2.4 自定义Plugin

本例中的代码如下

/**
*
*/
package com.demo.aidoc.es_plugin;

import java.util.HashMap;
import java.util.Map;

import org.apache.lucene.analysis.Analyzer;
import org.elasticsearch.index.analysis.AnalyzerProvider;
import org.elasticsearch.index.analysis.TokenizerFactory;
import org.elasticsearch.indices.analysis.AnalysisModule;
import org.elasticsearch.indices.analysis.AnalysisModule.AnalysisProvider;
import org.elasticsearch.plugins.AnalysisPlugin;
import org.elasticsearch.plugins.Plugin;

import com.demo.aidoc.es_plugin.factory.DemoTokenizerFactory;
import com.demo.aidoc.es_plugin.provider.DemoAnalyzerProvider;

/**
* GridsumPlugin自定义插件的主要实现,继承Plugin实现AnalysisPlugin,通过重写getTokenizers将分词器工厂放入map,通过重写getAnalyzers将分析器放入map(这里的key后面会用到)
* @author yishui
* @version 1.0.0
* @since 1.0.0
*/
public class DemoPlugin extends Plugin implements AnalysisPlugin {

private final static String PLUGIN_NAME="demo";

@Override
public Map<String, AnalysisProvider<TokenizerFactory>> getTokenizers() {
Map<String, AnalysisModule.AnalysisProvider<TokenizerFactory>> extra = new HashMap<>();
extra.put(PLUGIN_NAME, DemoTokenizerFactory::new);

return extra;
}

@Override
public Map<String, AnalysisProvider<AnalyzerProvider<? extends Analyzer>>> getAnalyzers() {
Map<String, AnalysisModule.AnalysisProvider<AnalyzerProvider<? extends Analyzer>>> extra = new HashMap<>();
extra.put(PLUGIN_NAME, DemoAnalyzerProvider::new);
return extra;
}

}

注意上述代码中插件的名字demo很重要,在使用自定义分词器时插件的名字就是这个。

2.5 测试自定义插件

public class DemoMain {

public static void main(String[] args) throws Exception {
DemoAnalyzer analyzer = new DemoAnalyzer();
TokenStream ts = analyzer.tokenStream("text", "中华人民共和国车船税法");
CharTermAttribute term = ts.addAttribute(CharTermAttribute.class);
ts.reset();
while (ts.incrementToken()) {
System.out.println(term.toString());
}
ts.end();
ts.close();
}

}

三 部署测试

测试 Java 插件时,只有在plugins/目录下才会自动加载。 使用 bin/elasticsearch-plugin install file:///path/to/your/plugin安装插件进行测试。

你也可以在集成测试的测试框架中加载你的插件。

测试请求如下

POST /_analyze
{
"analyzer": "demo",
"text": "我爱北京 天安门"
}