forked from HAL9000/cleveragents-core
freemo
051ee7c290
Added 52 new .feature files and corresponding _steps.py files targeting previously uncovered code paths in the following areas: - TUI layer: app, commands, persona (state/schema/registry), widgets, input (shell_exec, reference_parser) - Application services: plan lifecycle/service/executor, session, project, repo indexing, correction, checkpoint, actor, llm_actors, strategy coordinator, resource file watcher, service retry wiring - CLI commands: session, resource, repl, plan, db, automation_profile - Domain models: retry_policy, resource_type, cost_budget, docker_compose_analyzer, detail_level, _sql_string_aware, _postgresql_helpers - Core: circuit_breaker, retry_service_patterns - Infrastructure: repositories, transaction_sandbox, strategy_registry, plugins/loader, container - Config: settings - Agents: plan_generation, context_analysis, auto_debug - A2A: facade All new tests follow the Behave/Gherkin BDD standard. Resolved step definition collisions with unique prefixes. Fixed Alembic fileConfig logger disabling issue (disable_existing_loggers=False). ISSUES CLOSED: #1068
83 lines
4.2 KiB
Gherkin
83 lines
4.2 KiB
Gherkin
Feature: Strategy Coordinator Coverage
|
|
Additional scenarios that exercise previously uncovered code paths
|
|
in the strategy_coordinator module (lines 147, 152, 157, 162,
|
|
192, 198, 200, 291-294, 328, 331).
|
|
|
|
Background:
|
|
Given the strategy coordinator module is imported
|
|
|
|
# -----------------------------------------------------------------------
|
|
# Property accessors (lines 147, 152, 157, 162)
|
|
# -----------------------------------------------------------------------
|
|
|
|
Scenario: Accessing the config property returns the coordinator configuration
|
|
Given a strategy coordinator with default configuration
|
|
When I access the config property
|
|
Then the config property returns a CoordinatorConfig instance
|
|
|
|
Scenario: Accessing the selector property returns the strategy selector
|
|
Given a strategy coordinator with default configuration
|
|
When I access the selector property
|
|
Then the selector property returns a ConfidenceWeightedSelector instance
|
|
|
|
Scenario: Accessing the allocator property returns the budget allocator
|
|
Given a strategy coordinator with default configuration
|
|
When I access the allocator property
|
|
Then the allocator property returns a ProportionalBudgetAllocator instance
|
|
|
|
Scenario: Accessing the executor property returns the strategy executor
|
|
Given a strategy coordinator with default configuration
|
|
When I access the executor property
|
|
Then the executor property returns a ParallelStrategyExecutor instance
|
|
|
|
# -----------------------------------------------------------------------
|
|
# coordinate() with fragments=None default (line 192)
|
|
# -----------------------------------------------------------------------
|
|
|
|
Scenario: Coordinate with no fragments defaults to empty list
|
|
Given a strategy coordinator with mock pipeline components
|
|
And a mock strategy that returns fragments without strategy_source
|
|
When I call coordinate without passing fragments
|
|
Then the coordination should succeed
|
|
And the result should contain fragments from the mock strategy
|
|
|
|
# -----------------------------------------------------------------------
|
|
# coordinate() with backends parameter (line 198)
|
|
# -----------------------------------------------------------------------
|
|
|
|
Scenario: Coordinate enriches request with backends parameter
|
|
Given a strategy coordinator with a request-capturing selector
|
|
And a mock strategy that returns fragments without strategy_source
|
|
When I call coordinate with backends parameter
|
|
Then the enriched request should contain a backends key
|
|
|
|
# -----------------------------------------------------------------------
|
|
# coordinate() with plan_context parameter (line 200)
|
|
# -----------------------------------------------------------------------
|
|
|
|
Scenario: Coordinate enriches request with plan_context parameter
|
|
Given a strategy coordinator with a request-capturing selector
|
|
And a mock strategy that returns fragments without strategy_source
|
|
When I call coordinate with plan_context parameter
|
|
Then the enriched request should contain a plan_context key
|
|
|
|
# -----------------------------------------------------------------------
|
|
# _apply_max_caps zero-confidence redistribution (lines 291-294)
|
|
# -----------------------------------------------------------------------
|
|
|
|
Scenario: Max cap redistribution with zero-confidence uncapped strategies
|
|
Given a strategy coordinator with per_strategy_max_cap of 500
|
|
And allocations with one capped strategy and two zero-confidence uncapped strategies
|
|
When I apply max caps to the allocations
|
|
Then the excess tokens should be equally redistributed among uncapped strategies
|
|
|
|
# -----------------------------------------------------------------------
|
|
# _extract_strategies_used with strategy_source (lines 328, 331)
|
|
# -----------------------------------------------------------------------
|
|
|
|
Scenario: Strategies used extracted from fragment strategy_source metadata
|
|
Given a strategy coordinator with mock pipeline components
|
|
And a mock strategy that returns fragments with strategy_source set
|
|
When I call coordinate without passing fragments
|
|
Then the strategies_used should be derived from fragment strategy_source
|