Module: Omakase

Defined in:
lib/omakase.rb,
lib/omakase/doc.rb,
lib/omakase/mcp.rb,
lib/omakase/type.rb,
lib/omakase/agent.rb,
lib/omakase/trace.rb,
lib/omakase/memory.rb,
lib/omakase/schema.rb,
lib/omakase/skills.rb,
lib/omakase/request.rb,
lib/omakase/version.rb,
lib/omakase/executor.rb,
lib/omakase/fake_chat.rb,
lib/omakase/generation.rb,
lib/omakase/strategies.rb,
lib/omakase/tools/ruby.rb,
lib/omakase/capabilities.rb,
lib/omakase/strategies/predict.rb,
lib/omakase/strategies/code_act.rb

Defined Under Namespace

Modules: Capabilities, Doc, Executor, MCP, Skills, Strategies, Tools Classes: Agent, FakeChat, Generation, Memory, Request, Schema, Trace, Type

Constant Summary collapse

Error = 
Class.new(StandardError)
ContractError =

The answer did not match the declared return type.

Class.new(Error)
ProviderError =

The model or its provider failed. RubyLLM has already retried what it retries.

Class.new(Error)
CHAT_ARGUMENTS =

What RubyLLM.chat itself takes; every other option is a with_* call on the chat.

%i[model provider protocol assume_model_exists context].freeze
VERSION = 
"0.4.0"

Class Attribute Summary collapse

Class Method Summary collapse

Class Attribute Details

.listener ⇒ Object

Returns the value of attribute listener.



102
103
104
 
# File 'lib/omakase.rb', line 102

def listener
  @listener
end

Class Method Details

.build_chat(**options) ⇒ Object

caching: true, thinking: {effort: :high}, temperature: 0.2 become the chat's own with_* calls. Caching is on unless you say caching: false: the tool loop resends the whole chat on every step.



78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
 
# File 'lib/omakase.rb', line 78

def build_chat(**options)
  chat = RubyLLM.chat(**options.slice(*CHAT_ARGUMENTS))
  {caching: true, **options.except(*CHAT_ARGUMENTS)}.each do |name, value|
    setter = :"with_#{name}"
    raise Error, "unknown chat option #{name}: RubyLLM::Chat has no ##{setter}" unless chat.respond_to?(setter)

    keywords = chat.method(setter).parameters.any? { |kind, _| kind == :keyrest }
    case value
    when true then chat.public_send(setter)
    when Array then chat.public_send(setter, *value)
    when Hash then keywords ? chat.public_send(setter, **value) : chat.public_send(setter, value)
    else chat.public_send(setter, value)
    end
  end
  chat
end

.chat_factory ⇒ Object 



73
 
# File 'lib/omakase.rb', line 73

def chat_factory = @chat_factory ||= method(:build_chat)

.chat_factory=(factory) ⇒ Object

How an agent gets a chat when none was injected. Anything answering call(**options) will do — one line in test_helper.rb keeps a whole suite off the network, including the class-level calls a job makes.



69
70
71
 
# File 'lib/omakase.rb', line 69

def chat_factory=(factory)
  @chat_factory = callable!(factory, "chat_factory")
end

.configure ⇒ Object

Providers, keys, default model, timeouts, logging — all of it is RubyLLM's.



29
 
# File 'lib/omakase.rb', line 29

def configure(&) = RubyLLM.configure(&)

.configure_from_env(env = ENV) ⇒ Object

Reads every provider credential RubyLLM knows from the environment: ANTHROPIC_API_KEY, OPENROUTER_API_KEY, OLLAMA_API_BASE, VERTEXAI_PROJECT_ID, …



33
34
35
36
37
38
39
40
 
# File 'lib/omakase.rb', line 33

def configure_from_env(env = ENV)
  configure do |config|
    provider_options.each do |option|
      value = env[option.to_s.upcase]
      config.public_send(:"#{option}=", value) if value
    end
  end
end

.embedder ⇒ Object 



64
 
# File 'lib/omakase.rb', line 64

def embedder = @embedder ||= ->(text) { RubyLLM.embed(text).vectors }

.embedder=(embedder) ⇒ Object

How text becomes a vector, for Memory. Anything answering call(text) will do; the model and its provider are RubyLLM's to configure.



60
61
62
 
# File 'lib/omakase.rb', line 60

def embedder=(embedder)
  @embedder = callable!(embedder, "embedder")
end

.emit(event, **payload) ⇒ Object 



104
 
# File 'lib/omakase.rb', line 104

def emit(event, **payload) = @listener&.call(event, **payload)

.executor ⇒ Object 



56
 
# File 'lib/omakase.rb', line 56

def executor = @executor ||= Executor

.executor=(executor) ⇒ Object

Where generated code runs. Anything answering call(agent, code, timeout:) will do — Executor::Subprocess is the reference: a child process, so a timeout cannot take this one with it.



52
53
54
 
# File 'lib/omakase.rb', line 52

def executor=(executor)
  @executor = callable!(executor, "executor")
end

.provider_options ⇒ Object 



42
43
44
45
46
47
 
# File 'lib/omakase.rb', line 42

def provider_options
  slugs = RubyLLM::Provider.providers.keys
  RubyLLM::Configuration.options.select do |option|
    slugs.any? { |slug| option.to_s.start_with?("#{slug}_") }
  end
end