CTests

Tests compilation can be enabled at configure time using the option -DBUILD_TESTING=ON (OFF by default). After build, tests can be run with ctest:

cd _build

ctest -C Release --output-on-failure

All tests are associated to a label and multiple labels can be defined. You can choose which tests to execute when launching ctest. The list of available labels is the following:

Name	Label	Description
`unit_logger`	`unit`	Unit test for logger use.
`unit_launcher`	`unit`	Unit test Antares Xpansion python launcher.
`unit_solver`	`unit`	Unit test of multisolver interface(COIN only).
`unit_lpnamer`	`unit`	Unit test of lpnamer.
`unit_sensitivity`	`unit`	Unit test for sensitivity analysis.
`output_writer`	`unit`	Unit test for the output writer.
`helpers_test`	`unit`	Unit test for helpers (json reader, AntaresVersionProvider).
`lpnamer_end_to_end`	`lpnamer` `end_to_end`	End-to-end tests for lpnamer.
`examples_short_sequential`	`short` `short_sequential` `end_to_end`	Non-parallel end-to-end tests using benders algorithm on antares studies (short duration).
`examples_short_memory`	`short` `short_memory` `end_to_end`	Non-parallel end-to-end tests using benders algorithm on antares studies (short duration), use antares lib.
`examples_short_mpi`	`short` `short_mpi` `end_to_end`	Parallel end-to-end tests using benders algorithm on antares studies (short duration).
`examples_short_benders_by_batch_mpi`	`short` `short_mpi` `end_to_end` `benders_by_batch`	Parallel end-to-end tests using benders by batch algorithm on antares studies (short duration).
`examples_medium_sequential`	`medium` `medium_sequential` `end_to_end`	Non-parallel end-to-end tests using benders algorithm on antares studies (medium duration).
`examples_medium_mpi`	`medium` `medium_mpi` `end_to_end`	Parallel end-to-end tests using benders algorithm on antares studies (medium duration).
`examples_medium_benders_by_batch_mpi`	`medium` `medium_mpi` `end_to_end` `benders_by_batch`	Parallel end-to-end tests using benders by batch algorithm on antares studies (medium duration).
`examples_long_sequential`	`long` `long_sequential` `end_to_end`	Non-parallel end-to-end tests using benders algorithm on antares studies (long duration).
`examples_long_mpi`	`long` `long_mpi` `end_to_end`	Parallel end-to-end tests using benders algorithm on antares studies (long duration).
`mpibenders`	`benders`, `benders-mpi` `end_to_end`	End-to-end tests for benders mpi optimization.
`sequential`	`benders` `benders-sequential` `end_to_end`	End-to-end tests for benders sequential optimization.
`merge_mps`	`benders` `merge-mps` `end_to_end`	End-to-end tests for merge mps optimization.

!!! Note Use ctest -N to see all available tests.

Here is an example for running only examples_medium tests (use of Name with -R option):

ctest -C Release --output-on-failure -R examples_medium

Here is an example for running only unit tests (use of Label with -L option):

ctest -C Release --output-on-failure -L unit

To run all test, don't indicate any label or name:

ctest -C Release --output-on-failure

Cucumber Tests (BDD)

Overview

Cucumber tests use the behave framework (Python BDD testing) to run end-to-end feature tests. These tests validate high-level functionality through human-readable scenarios.

Location

tests/end_to_end/cucumber/
├── features/
│   ├── *.feature          # Feature files with Gherkin syntax
│   ├── environment.py     # Setup/teardown hooks
│   └── steps/             # Step implementations
│       ├── given.py
│       ├── when.py
│       ├── then.py
│       └── steps.py

Running Tests

# Install dependencies
pip install -r requirements-tests.txt
 
# Run all cucumber tests
cd tests/end_to_end
behave cucumber/features/ --no-capture --no-capture-stderr
 
# Run specific feature file
behave cucumber/features/Benders.feature --no-capture
 
# Run with tags (skip flaky or CI-excluded tests)
behave cucumber/features/ --tags "not @flaky and not @noci" --no-capture
 
# Run specific scenario by name
behave cucumber/features/ -n "Benders_handle_mixed_order_of_var"

Writing Tests

Feature file (*.feature):

Feature: Benders tests
 
  @fast @short @Benders
  Scenario: Test benders with specific study
    Given the study path is "data_test/Benders_handle_mixed_order_of_var"
    When I run benders with 1 proc(s)
    Then the simulation succeeds
    And the simulation takes less than 5 seconds

Step implementations are in the steps/ directory. Each step definition maps Gherkin steps to Python code.

Common Tags

Tag	Description
`@fast`	Quick tests
`@short`	Short duration
`@Benders`	Benders-related tests
`@flaky`	Unstable tests (skipped by default in CI)
`@noci`	Tests not run in CI
`@xpress`	Xpress solver-specific tests

GitHub Action

The workflow uses the action at .github/workflows/cucumber-tests/action.yml which:

Accepts feature (specific feature file/folder, default: features)
Accepts tags (tag expression, default: "not @flaky and not @noci")
Requires mpi_path for MPI support
Runs behave with pretty formatting and no output capture