[ES|QL] Add Decay function

[timgrein]

Addresses ESQL: Add decay functions

This PR introduces the decay function for numeric, spatial and temporal data types. For better intuition on decay functions you can read up on the detailed example on the decay function in the ES docs.

Key differences to the existing decay function:

• ESQL decay also works with dates using nanosecond precision (type: date_nanos)
• ESQL decay also works on cartesian points (using euclidean distance)

The function supports the following data types for the main input (value): int, double, long, geo_point, cartesian_point, datetime, date_nanos.

Numeric (int) example:

Creating the index:

curl --location --request PUT 'http://localhost:9200/decay_integer_linear_index' \
--header 'Content-Type: application/json' \
--data '{
  "mappings": {
    "properties": {
      "value": {
        "type": "integer"
      }
    }
  }
}'

Indexing sample data:

curl --location 'http://localhost:9200/_bulk' \
--header 'Content-Type: application/json' \
--data '{ "index": { "_index": "decay_integer_linear_index", "_id": "1" } }
{ "value": "0" }
{ "index": { "_index": "decay_integer_linear_index", "_id": "2" } }
{ "value": "10" }
{ "index": { "_index": "decay_integer_linear_index", "_id": "3" } }
{ "value": "50" }
{ "index": { "_index": "decay_integer_linear_index", "_id": "4" } }
{ "value": "100" }
{ "index": { "_index": "decay_integer_linear_index", "_id": "5" } }
{ "value": "2500" }
'

Querying using the Query DSL:

curl --location 'http://localhost:9200/decay_integer_linear_index/_search' \
--header 'Content-Type: application/json' \
--data '{
  "query": {
    "function_score": {
      "linear": {
        "value": {
          "origin": "0",
          "scale": "10",
          "offset": "5",
          "decay": 0.5
        }
      }
    }
  }
}'

Querying using ES|QL:

curl --location 'http://localhost:9200/_query?format=txt' \
--header 'Content-Type: application/json' \
--data '{
    "query": "FROM decay_integer_linear_index | KEEP value | EVAL score = decay(value, 0, 10, {\"offset\": 5, \"decay\": 0.5, \"type\": \"linear\"})"
}'

Spatial (GeoPoint) example:

Creating the index:

curl --location --request PUT 'http://localhost:9200/decay_geo_point_linear_index' \
--header 'Content-Type: application/json' \
--data '{
  "mappings": {
    "properties": {
      "value": {
        "type": "geo_point"
      }
    }
  }
}'

Indexing sample data:

curl --location 'http://localhost:9200/_bulk' \
--header 'Content-Type: application/json' \
--data '{ "index": { "_index": "decay_geo_point_linear_index", "_id": "1" } }
{ "value": "POINT (0 0)" }
{ "index": { "_index": "decay_geo_point_linear_index", "_id": "2" } }
{ "value": "POINT (1 1)" }
{ "index": { "_index": "decay_geo_point_linear_index", "_id": "3" } }
{ "value": "POINT (180 90)" }
{ "index": { "_index": "decay_geo_point_linear_index", "_id": "4" } }
{ "value": "POINT (-180 -90)" }
{ "index": { "_index": "decay_geo_point_linear_index", "_id": "5" } }
{ "value": "POINT (12.3 45.6)" }
'

Querying using the Query DSL:

curl --location 'http://localhost:9200/decay_geo_point_linear_index/_search' \
--header 'Content-Type: application/json' \
--data '{
  "query": {
    "function_score": {
      "linear": {
        "value": {
          "origin": "POINT (1 1)",
          "scale": "10000km",
          "offset": "10km",
          "decay": 0.33
        }
      }
    }
  }
}'

Querying using ES|QL:

curl --location 'http://localhost:9200/_query?format=txt' \
--header 'Content-Type: application/json' \
--data '{
    "query": "FROM decay_geo_point_linear_index | KEEP value | EVAL score = decay(TO_GEOPOINT(value), TO_GEOPOINT(\"POINT (1 1)\"), \"10000km\", {\"offset\": \"10km\", \"decay\": 0.33, \"type\": \"linear\"})"
}'

Temporal (Datetime) example:

Creating the index:

curl --location --request PUT 'http://localhost:9200/decay_datetime_linear_index' \
--header 'Content-Type: application/json' \
--data '{
  "mappings": {
    "properties": {
      "value": {
        "type": "date"
      }
    }
  }
}'

Indexing sample data:

curl --location 'http://localhost:9200/_bulk' \
--header 'Content-Type: application/json' \
--data '{ "index": { "_index": "decay_datetime_linear_index", "_id": "1" } }
{ "value": "2025-08-20" }
{ "index": { "_index": "decay_datetime_linear_index", "_id": "2" } }
{ "value": "2020-08-20" }
{ "index": { "_index": "decay_datetime_linear_index", "_id": "3" } }
{ "value": "2000-01-01" }
{ "index": { "_index": "decay_datetime_linear_index", "_id": "4" } }
{ "value": "1970-08-20" }
{ "index": { "_index": "decay_datetime_linear_index", "_id": "5" } }
{ "value": "1900-12-12" }
'

Querying using the Query DSL:

curl --location 'http://localhost:9200/decay_datetime_linear_index/_search' \
--header 'Content-Type: application/json' \
--data '{
  "query": {
    "function_score": {
      "linear": {
        "value": {
          "origin": "2000-01-01",
          "scale": "10000d",
          "offset": "10d",
          "decay": 0.33
        }
      }
    }
  }
}'

Querying using ES|QL:

curl --location 'http://localhost:9200/_query?format=txt' \
--header 'Content-Type: application/json' \
--data '{
    "query": "FROM decay_datetime_linear_index | KEEP value | EVAL score = decay(TO_DATETIME(value), TO_DATETIME(\"2000-01-01T00:00:00Z\"), 240000 hours, {\"offset\": 240 hours, \"decay\": 0.33, \"type\": \"linear\"})"
}'

[ioanatia]

nice work!
I checked the last review from @carlosdelest and it looks like you addressed all comments.

[leemthompo]

Docs look good for this first iteration 👍