Accessing number of matching document inside a script

Elastic Stack Elasticsearch
1

Hi,
I developped a custom plugin to modify the scores, according to some internal criteria, of documents matching a user query.
My problem is : the function that modifies the score is very expensive and I do not want to run it when the user query matches a very large number of documents.
I am wondering how to directly get this number of matching documents without putting a counter.

private static class MyExpertScriptEngine implements ScriptEngine {
   
    //choosen arbitrarily
    public static final double THRESHOLD = 10000;
   
    @Override
    public <T> T compile(String scriptName, String scriptSource, ScriptContext<T> context, Map<String, String> params) {
        // we use the script "source" as the script identifier
        if ("pure_df".equals(scriptSource)) {
            SearchScript.Factory factory = (p, lookup) -> new SearchScript.LeafFactory() {
		if (numberOfMatchingDocuments > THRESHOLD ){
                    throw InternalException("Cannot apply the expensive cost function because the number of matching documents is greater than " + THRESHOLD );	
                }     

                @Override
                public SearchScript newInstance(LeafReaderContext context) throws IOException {
                    return new SearchScript(p, lookup, context) {
                        
                        @Override
                        public double runAsDouble() {
                           // run our expensive function to modify the score of matching documents. 
                            
                    	};              
            };
            return context.factoryClazz.cast(factory);
        }
    }
}


POST /_search
{
  "query": {
    "function_score": {
      "query": {
        user_query
      },
      "functions": [
        {
          "script_score": {
            "script": {
                "source": "pure_df",
                "lang" : "expert_scripts",
                "params": {
                    "param1": "value1",
                    "params2": "value2"
                }
            }
          }
        }
      ]
    }
  }
}

Thanks for your help

2

If your "expensive" function is packaged as a query you could looking at using rescoring to limit the number of docs you apply this logic to

3

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.