TypesenseEx

(EXPERIMENTAL / WORK IN PROGRESS)

• Elixir bindings for the Typesense REST API
• A round-robbin load balancer for Typesense nodes

We will continue to add to this LiveBook as the API develops.

Installation and Configuration

Install Typesense

Either install Typesense locally or use the hosted version, Typesense Cloud.

See the docker-compose.yml file for a an example of a local development instance.

Install the Package

def deps do
  [
    {:typesense_ex, git: "https://github.com/MixTapeSoftware/typesense_ex"}
  ]
end

Add the Typesense to your supervision tree:

  @impl true
  def start(_type, _args) do
    children = [
      # Start the Telemetry supervisor
      YourAppWeb.Telemetry,
      # Start the Ecto repository
      YourApp.Repo,
      # Start the PubSub system
      {Phoenix.PubSub, name: YourApp.PubSub},
      # Start Finch
      {Finch, name: YourApp.Finch},
      # Start the Endpoint (http/https)
      YourAppWeb.Endpoint,
      {TypesenseEx, typesense_config()}, # <-- Add Me!

      # Start a worker by calling: YourApp.Worker.start_link(arg)
      # {YourApp.Worker, arg}
    ]

    # See https://hexdocs.pm/elixir/Supervisor.html
    # for other strategies and supported options
    opts = [strategy: :one_for_one, name: YourApp.Supervisor]
    Supervisor.start_link(children, opts)
  end


  defp typesense_config() do
    Application.get_env(:your_app, :typesense) |> Enum.into(%{})
    # Alternatively, you can just configure this here.
    # %{
    #   api_key: "MY_TYPESENSE_API_KEY",
    #   nodes: [%{host: "localhost", port: "8108", protocol: "http"}]
    # }
  end

Configuration

TypesenseEx is configurable at application start. Configuration may also be updated in a running service (e.g. to dynamically add/remove Typesense nodes as they come online or have been terminated).

Configuration Options

Name	Description
nodes	(optional) A list of typesense node configurations
nearest_node	(optional) A Typesense cloud Search Delivery Network feature.
api_key	(optional but recommended) A string representing the configured typesense api key (e.g. above it's "MY_TYPESENSE_API_KEY")
connection_timeout	[defaults to 10,000] Establishes how long in milliseconds Typesense should wait before retrying a request to a typesense node after timing out
num_retries	[defaults to 3] The number of retry attempts that should be made before marking a node unhealthy
retry_interval	[defaults to 100 / .10 second] The number of milliseconds to wait between retries
healthcheck_interval	[defaults to 15,000 / 15 seconds] The number of milliseconds to wait before resuming requests for a node after it has been marked unhealthy