Elasticsearch should query without computing relevance (_score) - performance

Question

I'm creating filtering queries which operates on two fields. I would like to avoid computing relevance by Elasticsearch. How to achieve OR statement without moving to query context.
My simplified model has two boolean fields:
{
is_opened,
is_send
}
I'd like to prepare query with logic:
(is_opened == true AND is_send == true) OR (is_opened == false)
In other words I want to exclude documents with fields:
is_opened == true AND is_send == false
My query looks like that:
GET documents/default/_search
{
"query": {
"bool": {
"should": [
{
"bool":{
"must":[
{"term": {"is_opened":true}},
{"term": {"is_send":true}}
]
}
},
{
"bool":{
"must":[
{"term": {"is_opened":false}}
]
}
}
]
}
}
}
Logically it works as I expected but Elasticsearch computes relevance.
I don't need it because at the end I sort results by another field so it's a place to optimize queries.
I ask about it because Frequently used filters will be cached automatically by Elasticsearch, to speed up performance.
My results have _score field computed so I think that above query is executed in query context so Elasticsearch won't cache it automatically.
In the future I would like to create queries which operates on status fields, where logic would be more complicated. Still I need to know how to block computing _score.
I noticed that changing should to filter block computing _score but works as must operator. Is it possible to change filter behavior?
Is it possible to use another query than should?
How to force Elasticserach to stop computing _score?

Answer 1

Simply wrap your query inside the constant_score query:
GET documents/default/_search
{
"query": {
"constant_score": {
"filter": {
"bool": {
"should": [
{
"bool": {
"must": [
{
"term": {
"is_opened": true
}
},
{
"term": {
"is_send": true
}
}
]
}
},
{
"bool": {
"must": [
{
"term": {
"is_opened": false
}
}
]
}
}
]
}
}
}
}
}