search-adapter.md

Search Adapter

! Experimental ! - this API is highly experimental and subject to changes.

The goal behind this adapter is to provide a search abstraction that various storage technologies can implement.

If an actor provides a list of searchableBy fields, then the storage adapter should receive "eager" updates, not limited by snapshot behavior, by which it can update its search data for the model.

Without this, search results will only be as valid and recent as the last snapshot for the model.

API

create (actorType)

Returns a promise that resolves to a searchAdapter instance for a specific type of actor.

find( criteria )

Criteria is an array with one or more sets where each set is a hash of criteria which must be true (all elements in a set are AND'd). Each set in the array should be treated as an independent group of requirements - so that, in effect sets are OR'd together.

Specific operations are represented as unique key/value sets. Any adapter implementing this API should throw exceptions for any unsupported operations.

There are inherent limitations to this approach - a more robust DSL/approach may be added to a future API call. Many times, the solution is to design a model that contains criteria for easier selection.

update ( type, fieldList, state, original )

Sends the latest version of state generated from applying all events, including any returned from the latest command.

fieldlist contains the list of fields (specified on the actor by searchableBy) that the model should be individually searchable by.

operations

equal

{
  x: 100
}

contains

{
  x: { contains: 100 }
}

matching / like

{
  x: { match: "pattern" }
}

in

{
  x: { in: [ 100, 101, 102 ] }
}

not in

{
  x: { not: [ 100, 101, 102 ] }
}

greater than

{
  x: { gt: 100 }
}

less than

{
  x: { lt: 100 }
}

greater than or equal to

{
  x: { gte: 100 }
}

less than or equal to

{
  x: { lte: 100 }
}

between

{
  x: [ lower, upper ]
}