Per-node cache of OpenSearch index mappings, read by Dowser.OpenSearch.Codec.

The cache is dowser_client's Dowser.MappingCacher — see it for how entries are keyed, fetched and expired, and for the start_link/1 options. This module is OpenSearch's own instance of it: its own process and ETS table, configured from the :dowser_opensearch application environment.

This package doesn't start it. An application that casts documents against the cluster's mappings — the default mapping source — starts it in its own supervision tree, with its options (see Dowser.MappingCacher):

children = [MyApp.Search, {Dowser.OpenSearch.MappingCacher, ttl: :timer.minutes(10)}]

Asked without having been started, it reports {:not_started, cacher}, which Dowser.OpenSearch.Codec turns into a Dowser.MappingError by default. An application that only casts with :struct, or with mappings of its own (config :dowser_opensearch, mapping: MyApp.Mappings), doesn't need it.

Mappings that are never fetched

A mapping given in the application environment answers for its index without any request at all — and so can never fail to be fetched:

config :dowser_opensearch,
  mappings: %{
    "posts" => %{
      "properties" => %{"published_at" => %{"type" => "date", "format" => "strict_date_time"}}
    }
  }

It wins over the cache. Use it for an index whose mapping is pinned, and in test suites, where it makes casting behave as it does in production without a cluster to ask (put/3 does the same through a running cacher).