Module generation

Expand description

AI-powered example generation for skill tools

This module provides infrastructure for generating synthetic usage examples using LLMs, with real-time streaming feedback via SSE events.

Structs§

AccuracyEvaluator: Evaluator for measuring accuracy of generated examples
AccuracyMetrics: Comprehensive accuracy metrics for a batch of generated examples
AgentStep: A step in the agent’s reasoning process (self-ask-with-search pattern)
ChatMessage: Chat message for multi-turn conversations
CompletionRequest: LLM completion request
DeterministicMockProvider: A deterministic mock LLM provider for testing
ExampleGenerator: AI-powered example generator
ExampleValidator: Validator for generated examples
FailingMockProvider: A mock provider that always fails
GeneratedExample: A generated usage example for a tool
GenerationStreamBuilder: Builder for creating generation event streams
GeneratorConfig: Configuration for the example generator
LlmChunk: A chunk from streaming completion
LlmResponse: Response from an LLM completion
ParsedCommand: Parsed command representation
PerformanceMetrics: Performance metrics for generation
SearchResultRef: A reference to a search result (lightweight for streaming)
TokenUsage: Token usage statistics
ToolMetrics: Metrics for a single tool’s generated examples
ValidationResult: Result of validating an example

aws_s3_tool: Create an AWS S3 tool
create_llm_provider: Create an LLM provider from configuration
docker_build_tool: Create a Docker build tool
kubernetes_apply_tool: Create a complex Kubernetes tool fixture
mock_response_for_tool: Get a mock JSON response for a tool
mock_response_with_errors: Get a mock response with some invalid examples for testing validation
simple_tool: Create a simple tool with minimal parameters
tool_with_constraints: Create a tool with enum constraints