Feature: LLM provider selection
  As a content system
  I want to route LLM calls to different providers
  So that I can use multiple AI services

  Scenario Outline: Different providers can generate blog content
    Given provider is "<provider>" and model is "<model>"
    And the Anthropic API key is configured in Settings
    When the LLM router is called with a prompt
    Then returns a non-empty string response
    And the generation log records the correct LLM provider
    And the generation log records the correct LLM model

    Examples:
      | provider  | model                      |
      | anthropic | claude-sonnet-4-20250514   |
      | openai    | gpt-4o                     |
      | gemini    | gemini-2.0-flash           |

  Scenario: Ollama provider generates blog content using local model
    Given the Ollama base URL is "http://localhost:11434"
    And provider is "ollama" and model is "mistral"
    When the LLM router is called with a prompt
    Then returns a non-empty string response
    And the router calls http://localhost:11434/api/chat

  Scenario: Unknown provider raises configuration error
    Given provider is "unknown_provider" and model is "some-model"
    When the LLM router is called with a prompt
    Then a UserError is raised with message containing "not configured"