AScriptEngine is a backend for implementing a scripting language in Elasticsearch.

Custom script engines integrate with Elasticsearch scripting framework through theScriptEngine interface. To register theScriptEngine, your plugin should implement theScriptPlugin interface and override thegetScriptEngine(Settings settings) method during plugin initialization.

Consider implementing a custom script engine when you need to use advanced internals of scripting, such as scripts that require term frequencies while scoring, or when implementing specialized scripting languages with custom syntax beyond standard Painless capabilities.

The plugindocumentation has more information on how to write a plugin so Elasticsearch will properly load it. For the complete ScriptEngine interface reference, refer to theofficial implementation.

This code creates a custom script engine that allows you to useexpert_scripts as the language name andpure_df as the script source in your Elasticsearch queries. The script calculates document scores using term frequency data instead of Elasticsearch standard scoring algorithm.

The following example shows the essential parts of implementing a customScriptEngine:

private static class MyExpertScriptEngine implements ScriptEngine {    // 1. Define your custom language name    @Override    public String getType() {        return "expert_scripts";  // This becomes your "lang" value    }    // 2. Define your script source and compilation    @Override    public <T> T compile(String scriptName, String scriptSource,                         ScriptContext<T> context, Map<String, String> params) {        // This recognizes "pure_df" as your script source        if ("pure_df".equals(scriptSource)) {            ScoreScript.Factory factory = new PureDfFactory();            return context.factoryClazz.cast(factory);        }        throw new IllegalArgumentException("Unknown script: " + scriptSource);    }    // ... (additional required methods)}// 3. Where the actual score calculation happensprivate static class ScoreScriptImpl extends ScoreScript {    @Override    public double execute(ExplanationHolder explanation) {        // This is where you define your custom scoring logic        // In this example: return term frequency as the score        try {            return postings.freq();        } catch (IOException e) {            return 0.0d;        }    }}

1. Custom score calculation

• Language Definition: ThegetType() method returnsexpert_scripts, which becomes the value you use for thelang parameter in your scripts.
• Script Recognition: Thecompile() method identifiespure_df as a valid script source, which becomes the value you use for thesource parameter.
• Custom Scoring: Theexecute() method replaces Elasticsearch standard scoring with your custom logic. In this case, using term frequency as the document score.

For the complete implementation, refer to theofficial script engine example.

This example shows how to use your custom script engine in a search query:

POST /_search{  "query": {    "function_score": {      "query": {        "match": {          "body": "foo"        }      },      "functions": [        {          "script_score": {            "script": {                "source": "pure_df",                "lang" : "expert_scripts",                "params": {                    "field": "body",                    "term": "foo"                }            }          }        }      ]    }  }}