Arcana.Agent.Answerer behaviour (Arcana v1.3.3)

View Source

Behaviour for answer generation in the Agent pipeline.

The answerer generates the final response based on the question and retrieved context chunks.

Built-in Implementations

Implementing a Custom Answerer

defmodule MyApp.TemplateAnswerer do
  @behaviour Arcana.Agent.Answerer

  @impl true
  def answer(question, chunks, _opts) do
    context = Enum.map_join(chunks, "

", & &1.text)

    answer = "Based on " <> Integer.to_string(length(chunks)) <> " sources:

" <> context

    {:ok, answer}
  end
end

Using a Custom Answerer

Agent.new(question, repo: repo, llm: llm)
|> Agent.search()
|> Agent.answer(answerer: MyApp.TemplateAnswerer)

Using an Inline Function

Agent.answer(ctx,
  answerer: fn question, chunks, opts ->
    llm = Keyword.fetch!(opts, :llm)
    prompt = build_my_prompt(question, chunks)
    Arcana.LLM.complete(llm, prompt, [], [])
  end
)

Summary

Callbacks

answer(question, chunks, opts)

Generates an answer based on the question and context chunks.

Callbacks

answer(question, chunks, opts)

@callback answer(
  question :: String.t(),
  chunks :: [map()],
  opts :: keyword()
) :: {:ok, String.t()} | {:error, term()}

Generates an answer based on the question and context chunks.

Parameters

  • question - The user's original question
  • chunks - List of context chunks retrieved by search
  • opts - Options passed to Agent.answer/2, including:
    • :llm - The LLM function (for LLM-based answerers)
    • :prompt - Custom prompt function fn question, chunks -> prompt end
    • Any other options passed to Agent.answer/2

Returns

  • {:ok, answer} - The generated answer string
  • {:error, reason} - On failure