Mock implementation for the Claude Code CLI for testing purposes.

This module provides a GenServer-based mock system that allows testing and development without making actual API calls to the Claude service. It can be configured with predefined responses for different prompt patterns.

Features

• Pattern-based responses: Configure responses for specific prompt patterns
• Default fallback: Provides realistic default responses for unmatched prompts
• Integration testing: Seamlessly integrates with the main SDK for testing
• Cost-free development: Enables development without incurring API costs

Usage

Start the mock server and configure responses:

{:ok, _pid} = ClaudeCodeSDK.Mock.start_link()

# Set up a specific response
ClaudeCodeSDK.Mock.set_response("hello", [
  %{"type" => "system", "subtype" => "init", "session_id" => "mock-123"},
  %{"type" => "assistant", "message" => %{"content" => "Hello from mock!"}},
  %{"type" => "result", "subtype" => "success", "total_cost_usd" => 0.001}
])

Enable mocking in your application configuration:

Application.put_env(:claude_code_sdk, :use_mock, true)

Response Format

Mock responses should follow the same format as the actual Claude CLI output:

• Each response is a list of message maps
• Each message has a "type" field (system, assistant, user, result)
• Messages may include optional "subtype" fields for categorization
• Result messages should include cost and timing information for realistic testing

Testing Integration

The mock system is designed to work seamlessly with testing frameworks:

test "queries return expected responses" do
  ClaudeCodeSDK.Mock.set_response("test prompt", expected_messages)

  result = ClaudeCodeSDK.query("test prompt", options)

  assert length(Enum.to_list(result)) == length(expected_messages)
end