brent.edwards
e801eb1ee8
- Rewrite .forgejo/workflows/ci.yml to route all jobs through nox sessions - Fix coverage_report nox session: serial behave mode replaces broken parallel mode (22% -> 97% accuracy), raise fail-under from 85% to 97% - Pass posargs through format nox session for CI --check support - Add 11 CI workflow validation scenarios (Behave) + Robot smoke test + ASV bench - Add 108 new Behave scenarios covering 6 largest coverage gaps to reach 97%: yaml_template_engine, actor/config, actor/registry, message_router, context_analysis, context_service - Update docs/development/ci-cd.md with nox-based CI docs and 97% threshold - Restore implementation_plan.md verbose style, check off completed CI tasks Verified: 1673 scenarios pass, 97% coverage, lint clean, typecheck clean
121 lines
4.9 KiB
Gherkin
121 lines
4.9 KiB
Gherkin
Feature: Context service new coverage
|
|
As a developer
|
|
I want ContextService LangGraph analysis methods thoroughly tested
|
|
So that context analysis, streaming, and vector search stay stable
|
|
|
|
Scenario: _get_current_plan returns None for project without id
|
|
Given a context service with mocked dependencies
|
|
And a project without an id
|
|
When I get the current plan
|
|
Then the context result should be None
|
|
|
|
Scenario: _build_langsmith_config returns empty dict when disabled
|
|
Given a context service with mocked dependencies
|
|
And a project with id 1
|
|
When I build langsmith config with langsmith disabled
|
|
Then the config should be an empty dict
|
|
|
|
Scenario: _build_langsmith_config returns config when enabled
|
|
Given a context service with langsmith enabled
|
|
And a project with id 1
|
|
When I build langsmith config with langsmith enabled
|
|
Then the config should contain metadata and tags
|
|
|
|
Scenario: _prepare_analysis_config adds thread_id
|
|
Given a context service with mocked dependencies
|
|
And a project with id 1
|
|
When I prepare analysis config
|
|
Then the config should have a configurable thread_id
|
|
|
|
Scenario: _vector_store_available returns False when no store
|
|
Given a context service with no vector store
|
|
When I check vector store availability
|
|
Then the vector store check should return False
|
|
|
|
Scenario: _refresh_vector_index is a no-op when store unavailable
|
|
Given a context service with no vector store
|
|
When I refresh vector index for plan 1
|
|
Then no exception should be raised
|
|
|
|
Scenario: _refresh_vector_index handles ConfigurationError gracefully
|
|
Given a context service with a failing vector store
|
|
When I refresh vector index for plan 1
|
|
Then no exception should be raised
|
|
|
|
Scenario: _get_context_agent returns a ContextAnalysisAgent
|
|
Given a context service with mocked dependencies
|
|
When I get the context agent
|
|
Then the context agent should be a ContextAnalysisAgent instance
|
|
|
|
Scenario: analyze_context returns empty analysis for empty context
|
|
Given a context service with mocked dependencies
|
|
And a project that has no context files
|
|
When I analyze context
|
|
Then the result summary should indicate no files
|
|
|
|
Scenario: analyze_context runs full analysis for project with files
|
|
Given a context service with mocked dependencies
|
|
And a project with context files
|
|
When I analyze context
|
|
Then the result should have documents and summary
|
|
|
|
Scenario: get_context_summary returns summary string
|
|
Given a context service with mocked dependencies
|
|
And a project with context files
|
|
When I retrieve the context summary
|
|
Then the context summary should be a nonempty string
|
|
|
|
Scenario: get_context_dependencies returns dependencies dict
|
|
Given a context service with mocked dependencies
|
|
And a project with context files
|
|
When I get context dependencies
|
|
Then the context deps result should be a dict
|
|
|
|
Scenario: get_relevant_files filters by threshold
|
|
Given a context service with mocked dependencies
|
|
And a project with context files
|
|
When I fetch relevant files above threshold 0.5
|
|
Then the relevant files should be a list of tuples
|
|
|
|
Scenario: search_context returns empty for blank query
|
|
Given a context service with mocked dependencies
|
|
And a project with id 1
|
|
When I search context with empty query
|
|
Then the context search result should be an empty list
|
|
|
|
Scenario: search_context returns empty when no plan
|
|
Given a context service with mocked dependencies
|
|
And a project with id 1 but no current plan
|
|
When I search context with query "test"
|
|
Then the context search result should be an empty list
|
|
|
|
Scenario: search_context returns empty when vector store unavailable
|
|
Given a context service with no vector store
|
|
And a project with id 1 and a current plan
|
|
When I search context with query "test"
|
|
Then the context search result should be an empty list
|
|
|
|
Scenario: search_context delegates to vector store
|
|
Given a context service with a working vector store
|
|
And a project with id 1 and a current plan
|
|
When I search context with query "test"
|
|
Then the vector store search should be called
|
|
|
|
Scenario: search_context handles ConfigurationError from vector store
|
|
Given a context service with a vector store that raises ConfigurationError
|
|
And a project with id 1 and a current plan
|
|
When I search context with query "test"
|
|
Then the context search result should be an empty list
|
|
|
|
Scenario: analyze_context_streaming yields complete for empty context
|
|
Given a context service with mocked dependencies
|
|
And a project that has no context files
|
|
When I stream analyze context
|
|
Then the first event should indicate no files
|
|
|
|
Scenario: analyze_context_streaming yields events for project with files
|
|
Given a context service with mocked dependencies
|
|
And a project with context files
|
|
When I stream analyze context
|
|
Then at least one streaming event should be yielded
|