A make-like build system written on python
Create oak_build.py file in your project directory.
Every method marked with @task decorator can be called from CLI.
from pathlib import Path
from oak_build import task
@task
def create_file():
with open(Path("result.txt"), "w") as txt:
txt.write("test content\n")To execute create_file task call oak create_file from console.
You can link dependent tasks with depends_on parameter.
from oak_build import task, run
@task
def unit_tests():
run("poetry run pytest tests")
@task
def integration_tests():
run("poetry run pytest integration_tests")
@task(
depends_on=[
unit_tests,
integration_tests,
]
)
def tests():
passWhen oak tests is called oak build will execute unit_tests and integration_tests tasks as well.
You can pass parameters to tasks with --param key=value parameter of command line.
Oak will try to parse all task function params regarding its annotation type.
Example:
from enum import Enum, auto
from oak_build import task
class TaskEnum(Enum):
VALUE = auto()
@task
def enum_param(param: TaskEnum):
passAll int return value from task will be treated as return code. If return is integer and not zero task will be marked as failed and further execution will be stopped.
from oak_build import task
@task
def exit_code_task():
return 0All dict return value will be treated as result. Those results can be used in another tasks with tas name prefix (see example below).
from oak_build import task
@task
def source():
return {
"result": 123,
}
@task(depends_on=[source])
def result_consumer(source_result):
passPlease note that CLI params have greater priority than arguments deducted from task results.
You can return Tuple[int, Dict[str, Any]] from task, and it will be treated as return code and result dict.
If code is not zero task is considered failed.
The result dict processing is the same as in dict result
If task throws exception it is considered faled and further execution will be stopped.
For examples see integration tests files and self build oak_file.py.