49 lines
2.8 KiB
Gherkin
49 lines
2.8 KiB
Gherkin
Feature: Plan Service Coverage Boost
|
|
Target uncovered lines in plan_service.py to increase branch and line coverage.
|
|
|
|
Scenario: Build actor context populates initial_context from config_blob dict
|
|
Given I have a lightweight plan service for coverage boost testing
|
|
And I have an actor with initial_context dict in config_blob
|
|
When I build the actor context for the actor with initial_context
|
|
Then the actor invocation context should contain the initial_context entries
|
|
|
|
Scenario: Build actor context populates graph_descriptor from actor dict
|
|
Given I have a lightweight plan service for coverage boost testing
|
|
And I have an actor with a dict graph_descriptor
|
|
When I build the actor context for the actor with graph_descriptor
|
|
Then the actor invocation context should contain the graph_descriptor
|
|
|
|
Scenario: Resolve AI provider raises PlanError when actor has no provider
|
|
Given I have a lightweight plan service for coverage boost testing
|
|
And the actor lookup returns an actor with empty provider
|
|
When I try to resolve the AI provider for the providerless actor
|
|
Then the providerless actor PlanError should contain "Actor is missing provider configuration"
|
|
|
|
Scenario: Strip code fences removes markdown python fences
|
|
Given I have a lightweight plan service for coverage boost testing
|
|
When I strip code fences from python-fenced content
|
|
Then the stripped content should contain only the inner code
|
|
|
|
Scenario: Strip code fences removes generic markdown fences
|
|
Given I have a lightweight plan service for coverage boost testing
|
|
When I strip code fences from generic-fenced content
|
|
Then the stripped content should contain only the inner text
|
|
|
|
Scenario: LangSmith config returns empty dict when builder returns None
|
|
Given I have a lightweight plan service for coverage boost testing
|
|
And LangSmith is enabled but build_langsmith_config returns None
|
|
When I build the LangSmith config for a project
|
|
Then the plan service LangSmith config result should be an empty dict
|
|
|
|
Scenario: Build actor context uses context_variables fallback for initial_context
|
|
Given I have a lightweight plan service for coverage boost testing
|
|
And I have an actor with context_variables fallback in config_blob
|
|
When I build the actor context for the actor with context_variables
|
|
Then the actor invocation context should contain the context_variables entries
|
|
|
|
Scenario: Build actor context falls back to config_blob graph_descriptor key
|
|
Given I have a lightweight plan service for coverage boost testing
|
|
And I have an actor with graph_descriptor in config_blob but not on the actor
|
|
When I build the actor context for the config_blob graph_descriptor actor
|
|
Then the actor invocation context should contain the config_blob graph_descriptor
|