ElasticSearch painless scripts - Way to output variable values besides the final score?

Question

I am using a painless script to implement a custom scoring function while querying the ES index, that's being used as a basis for our recommendation engine. While calculating the final score in the painless script, I use a product of intermediate variables such as recency and uniqueness, calculated within the painless script.

Now, it is trivial to get the final scores of the top documents as they are returned within the query response. However, for detailed analysis, I'm trying to find a way to also get the intermediate variables' values (recency and uniqueness as in the above example). I understand these painless variables only exist within the context of the painless script, which does not have standard REPL setup. So is there really no way to access these painless variables? Has anyone found a workaround to do this? Thanks!

E.g. If I have the following simplified painless script:

def recency = 1/doc['date'].value
def uniqueness = doc['ctr].value

return recency * uniqueness

In the final ES response, I get the scores i.e. recency * uniqueness. However, I also want to know what the intermediate variables are i.e. recency and uniqueness

score 1 · Answer 1 · answered Jul 15 '19 at 18:42

1

You can try using a modular approach with multiple scripted fields like:

recency -- get the recency field
uniqueness -- get the uniqueness field
access the fields like a normal ES field in your final painless script

if(doc.containsKey('recency.keyword') && doc.containsKey('uniqueness.keyword'))
{
  def val1 = doc['recency.keyword'].value;
  def val2 = doc['uniqueness.keyword'].value;
}

Hope it helps

answered Jul 15 '19 at 18:42

Aritra Nayak

63
8

thanks for response aritra! It doesn't solve the issue so I think I was unclear in my question - I've added a clear example so hopefully it will help! Thanks – peterkim95 Jul 16 '19 at 01:57
well my suggestion still holds true if you do not mind having 3 scripted fields instead of one. However, if you really want to have just a single field - you may try to return a map ... ``` def recency = 1/doc['date'].value def uniqueness = doc['ctr].value def result = recency * uniqueness def returnMap = ['recency': recency, 'uniqueness': uniqueness, 'result': result] ``` – Aritra Nayak Jul 16 '19 at 04:56
so I tried using script fields and obtain the following fields in the result `fields': {'recency': [5669752], 'uniqueness': [8504628.0]}` But now I can't access them in my final painless script. I tried using your method but it doesn't work for me. Do I to create separate mappings for recency and uniqueness in my index prior? – peterkim95 Jul 18 '19 at 02:40
Also, referring to your returning map solution, how can you return a map in a painless script since the script is expecting a single score to sort on? – peterkim95 Jul 18 '19 at 02:49

score 0 · Answer 2 · edited Jul 01 '22 at 04:50

There is not direct way of printing it anywhere i suppose. But here is what you can give it a try to check the intermediate output of any variable.

create another scripted field which will return only the value of that variable.

For Ex: in your case,

"script_fields": {
"derivedRecency": {
      "script": {
        "lang": "painless",
        "source": """ 
            return doc['recency'].value;
            
          """
      }
    }
}

ElasticSearch painless scripts - Way to output variable values besides the final score?

2 Answers2