Class: Legion::Extensions::Llm::AzureFoundry::Actor::AzureFoundryCallable

Inherits:
Object
  • Object
show all
Defined in:
lib/legion/extensions/llm/azure_foundry/actors/discovery_refresh.rb

Overview

Callable wrapper for an Azure Foundry provider instance. Implements the fleet dispatch operations by delegating to the per-instance AzureFoundry::Provider (errors propagate so normalize_dispatch_error can classify them), plus the disconnect and normalize_dispatch_error(error:) contracts required by Inventory::CallableHandle and Routing::ProviderOutcome.

The fleet passes model: as a raw string (the offering's model id). chat/stream_chat render paths call model.id (the stock OpenAICompatible#render_payload), so a Model::Info is required there; embed/count_tokens are string-tolerant (Provider#model_id accepts the value verbatim), so those pass through — wrapping them would serialize a Data object into the wire payload/response.

Instance Method Summary collapse

Constructor Details

#initialize(instance_cfg:, logger:, provider: nil) ⇒ AzureFoundryCallable

Returns a new instance of AzureFoundryCallable.



1005
1006
1007
1008
1009
1010
 
# File 'lib/legion/extensions/llm/azure_foundry/actors/discovery_refresh.rb', line 1005

def initialize(instance_cfg:, logger:, provider: nil)
  @instance_cfg = instance_cfg
  @logger = logger
  @provider = provider
  @disconnected = false
end

Instance Method Details

#chat(messages:, model:, **rest) ⇒ Object 



1026
1027
1028
 
# File 'lib/legion/extensions/llm/azure_foundry/actors/discovery_refresh.rb', line 1026

def chat(messages:, model:, **rest)
  provider.chat(messages: messages, model: to_model_info(model), **rest)
end

#count_tokens(messages:, model:, **rest) ⇒ Object 



1038
1039
1040
 
# File 'lib/legion/extensions/llm/azure_foundry/actors/discovery_refresh.rb', line 1038

def count_tokens(messages:, model:, **rest)
  provider.count_tokens(messages: messages, model: model, **rest)
end

#disconnect ⇒ Object 



1020
1021
1022
1023
1024
 
# File 'lib/legion/extensions/llm/azure_foundry/actors/discovery_refresh.rb', line 1020

def disconnect
  @disconnected = true
  @provider&.disconnect
  @logger.debug { '[azure_foundry][callable] disconnected' }
end

#disconnected? ⇒ Boolean

Returns:

  • (Boolean) 


1016
1017
1018
 
# File 'lib/legion/extensions/llm/azure_foundry/actors/discovery_refresh.rb', line 1016

def disconnected?
  @disconnected
end

#embed(text:, model:, **rest) ⇒ Object 



1034
1035
1036
 
# File 'lib/legion/extensions/llm/azure_foundry/actors/discovery_refresh.rb', line 1034

def embed(text:, model:, **rest)
  provider.embed(text: text, model: model, **rest)
end

#normalize_dispatch_error(error:) ⇒ Object

D17: production dispatch raises Legion::Extensions::Llm::*Error (ErrorMiddleware), not raw Faraday — the base Provider#normalize_dispatch_error classifies those. On top, Azure wire semantics supply two stronger signals: an explicit EndpointDeactivated body (the ONLY instance_unavailable signal) and a model-not-ready 503 body (request-local, not instance-down).



1048
1049
1050
1051
1052
1053
1054
1055
 
# File 'lib/legion/extensions/llm/azure_foundry/actors/discovery_refresh.rb', line 1048

def normalize_dispatch_error(error:)
  outcome = base_provider_outcome(error: error)
  return outcome_with_kind(outcome, :instance_unavailable) if explicit_endpoint_deactivated?(error: error)
  return outcome_with_kind(outcome, :model_not_ready) if model_not_ready_outcome?(outcome, error: error)
  return outcome_with_kind(outcome, :model_missing) if model_missing_outcome?(outcome, error: error)

  outcome
end

#provider ⇒ Object 



1012
1013
1014
 
# File 'lib/legion/extensions/llm/azure_foundry/actors/discovery_refresh.rb', line 1012

def provider
  @provider ||= Legion::Extensions::Llm::AzureFoundry::Provider.new(@instance_cfg)
end

#stream_chat(messages:, model:, **rest) ⇒ Object 



1030
1031
1032
 
# File 'lib/legion/extensions/llm/azure_foundry/actors/discovery_refresh.rb', line 1030

def stream_chat(messages:, model:, **rest, &)
  provider.stream(messages: messages, model: to_model_info(model), **rest, &)
end