def test_pull_config_missing(self, fake_kedro_cli, dummy_project, env, alias):
"""
Test for pulling a valid wheel file locally, but `config` directory is missing
from the wheel file.
"""
# pylint: disable=too-many-locals
self.call_pipeline_create(fake_kedro_cli)
config_path = dummy_project / "conf" / "base" / "pipelines" / PIPELINE_NAME
shutil.rmtree(config_path)
assert not config_path.exists()
self.call_pipeline_package(fake_kedro_cli)
self.call_pipeline_delete(fake_kedro_cli)
source_path = dummy_project / "src" / PACKAGE_NAME / "pipelines" / PIPELINE_NAME
test_path = dummy_project / "src" / "tests" / "pipelines" / PIPELINE_NAME
# Make sure the files actually deleted before pulling from the wheel file.
assert not source_path.exists()
assert not test_path.exists()
wheel_file = (
dummy_project
/ "src"
/ "dist"
/ _get_wheel_name(name=PIPELINE_NAME, version="0.1")
)
assert wheel_file.is_file()
options = ["-e", env] if env else []
options += ["--alias", alias] if alias else []
result = CliRunner().invoke(
fake_kedro_cli.cli, ["pipeline", "pull", str(wheel_file), *options]
)
assert result.exit_code == 0
pipeline_name = alias or PIPELINE_NAME
source_dest = dummy_project / "src" / PACKAGE_NAME / "pipelines" / pipeline_name
config_env = env or "base"
config_dest = dummy_project / "conf" / config_env / "pipelines" / pipeline_name
test_dest = dummy_project / "src" / "tests" / "pipelines" / pipeline_name
self.assert_package_files_exist(source_dest)
assert not config_dest.exists()
assert {f.name for f in test_dest.iterdir()} == {
"__init__.py",
"test_pipeline.py",
}
def test_package_pipeline_with_deep_nested_parameters(
self, fake_repo_path, fake_project_cli, fake_metadata):
CliRunner().invoke(fake_project_cli, ["pipeline", "create", "retail"],
obj=fake_metadata)
deep_nested_param_path = Path(fake_repo_path / "conf" / "base" /
"parameters" / "deep" / "retail")
deep_nested_param_path.mkdir(parents=True, exist_ok=True)
(deep_nested_param_path / "params1.yml").touch()
deep_nested_param_path2 = Path(fake_repo_path / "conf" / "base" /
"parameters" / "retail" / "deep")
deep_nested_param_path2.mkdir(parents=True, exist_ok=True)
(deep_nested_param_path2 / "params1.yml").touch()
deep_nested_param_path3 = Path(fake_repo_path / "conf" / "base" /
"parameters" / "deep")
deep_nested_param_path3.mkdir(parents=True, exist_ok=True)
(deep_nested_param_path3 / "retail.yml").touch()
super_deep_nested_param_path = Path(fake_repo_path / "conf" / "base" /
"parameters" / "a" / "b" / "c" /
"d" / "retail")
super_deep_nested_param_path.mkdir(parents=True, exist_ok=True)
(super_deep_nested_param_path / "params3.yml").touch()
result = CliRunner().invoke(fake_project_cli,
["pipeline", "package", "retail"],
obj=fake_metadata)
assert result.exit_code == 0
assert "Pipeline `retail` packaged!" in result.output
wheel_location = fake_repo_path / "src" / "dist"
assert f"Location: {wheel_location}" in result.output
wheel_name = _get_wheel_name(name="retail", version="0.1")
wheel_file = wheel_location / wheel_name
assert wheel_file.is_file()
assert len(list(wheel_location.iterdir())) == 1
# pylint: disable=consider-using-with
wheel_contents = set(ZipFile(str(wheel_file)).namelist())
assert "retail/config/parameters/deep/retail/params1.yml" in wheel_contents
assert "retail/config/parameters/retail/deep/params1.yml" in wheel_contents
assert "retail/config/parameters/retail.yml" in wheel_contents
assert "retail/config/parameters/deep/retail.yml" in wheel_contents
assert "retail/config/parameters/a/b/c/d/retail/params3.yml" in wheel_contents
def test_pipeline_pull_all( # pylint: disable=too-many-locals
self, fake_repo_path, fake_project_cli, fake_metadata, mocker):
# pylint: disable=import-outside-toplevel
from kedro.framework.cli import pipeline
spy = mocker.spy(pipeline, "_pull_package")
pyproject_toml = fake_repo_path / "pyproject.toml"
wheel_file = (lambda name: fake_repo_path / "src" / "dist" /
_get_wheel_name(name=name, version="0.1"))
project_toml_str = textwrap.dedent(f"""
[tool.kedro.pipeline.pull]
"{wheel_file("first")}" = {{alias = "dp"}}
"{wheel_file("second")}" = {{alias = "ds", env = "local"}}
"{wheel_file("third")}" = {{}}
""")
with pyproject_toml.open(mode="a") as file:
file.write(project_toml_str)
for name in ("first", "second", "third"):
call_pipeline_create(fake_project_cli,
fake_metadata,
pipeline_name=name)
call_pipeline_package(fake_project_cli,
fake_metadata,
pipeline_name=name)
call_pipeline_delete(fake_project_cli,
fake_metadata,
pipeline_name=name)
result = CliRunner().invoke(fake_project_cli,
["pipeline", "pull", "--all"],
obj=fake_metadata)
assert result.exit_code == 0
assert "Pipelines pulled and unpacked!" in result.output
assert spy.call_count == 3
build_config = toml.loads(project_toml_str)
pull_manifest = build_config["tool"]["kedro"]["pipeline"]["pull"]
for wheel_file, pull_specs in pull_manifest.items():
expected_call = mocker.call(wheel_file, fake_metadata,
**pull_specs)
assert expected_call in spy.call_args_list
def assert_wheel_contents_correct(
self, wheel_location, package_name=PIPELINE_NAME, version="0.1"
):
wheel_name = _get_wheel_name(name=package_name, version=version)
wheel_file = wheel_location / wheel_name
assert wheel_file.is_file()
assert len(list((wheel_location).iterdir())) == 1
wheel_contents = set(ZipFile(str(wheel_file)).namelist())
expected_files = {
f"{package_name}/__init__.py",
f"{package_name}/README.md",
f"{package_name}/nodes.py",
f"{package_name}/pipeline.py",
f"{package_name}/config/parameters/{package_name}.yml",
"tests/__init__.py",
"tests/test_pipeline.py",
}
assert expected_files <= wheel_contents
def test_pipeline_alias_refactors_imports(
self, fake_project_cli, fake_package_path, fake_repo_path, fake_metadata
):
call_pipeline_create(fake_project_cli, fake_metadata)
pipeline_file = fake_package_path / "pipelines" / PIPELINE_NAME / "pipeline.py"
import_stmt = (
f"import {fake_metadata.package_name}.pipelines.{PIPELINE_NAME}.nodes"
)
with pipeline_file.open("a") as f:
f.write(import_stmt)
package_alias = "alpha"
pull_alias = "beta"
call_pipeline_package(
cli=fake_project_cli, metadata=fake_metadata, alias=package_alias
)
wheel_file = (
fake_repo_path
/ "src"
/ "dist"
/ _get_wheel_name(name=package_alias, version="0.1")
)
CliRunner().invoke(
fake_project_cli, ["pipeline", "pull", str(wheel_file)], obj=fake_metadata
)
CliRunner().invoke(
fake_project_cli,
["pipeline", "pull", str(wheel_file), "--alias", pull_alias],
obj=fake_metadata,
)
for alias in (package_alias, pull_alias):
path = fake_package_path / "pipelines" / alias / "pipeline.py"
file_content = path.read_text()
expected_stmt = (
f"import {fake_metadata.package_name}.pipelines.{alias}.nodes"
)
assert expected_stmt in file_content
def test_pull_whl_fs_args(self, fake_project_cli, fake_repo_path, mocker,
tmp_path, fake_metadata):
"""Test for pulling a wheel file with custom fs_args specified."""
call_pipeline_create(fake_project_cli, fake_metadata)
call_pipeline_package(fake_project_cli, fake_metadata)
call_pipeline_delete(fake_project_cli, fake_metadata)
fs_args_config = tmp_path / "fs_args_config.yml"
with fs_args_config.open(mode="w") as f:
yaml.dump({"fs_arg_1": 1, "fs_arg_2": {"fs_arg_2_nested_1": 2}}, f)
mocked_filesystem = mocker.patch("fsspec.filesystem")
wheel_file = (fake_repo_path / "src" / "dist" /
_get_wheel_name(name=PIPELINE_NAME, version="0.1"))
options = ["--fs-args", str(fs_args_config)]
CliRunner().invoke(
fake_project_cli,
["pipeline", "pull", str(wheel_file), *options])
mocked_filesystem.assert_called_once_with(
"file", fs_arg_1=1, fs_arg_2=dict(fs_arg_2_nested_1=2))
def test_pull_two_dist_info(self, fake_kedro_cli, dummy_project, mocker,
tmp_path):
"""
Test for pulling a wheel file with more than one dist-info directory.
"""
self.call_pipeline_create(fake_kedro_cli)
self.call_pipeline_package(fake_kedro_cli)
wheel_file = (dummy_project / "src" / "dist" /
_get_wheel_name(name=PIPELINE_NAME, version="0.1"))
assert wheel_file.is_file()
(tmp_path / "dummy.dist-info").mkdir()
mocker.patch(
"kedro.framework.cli.pipeline.tempfile.TemporaryDirectory",
return_value=tmp_path,
)
result = CliRunner().invoke(
fake_kedro_cli.cli,
["pipeline", "pull", str(wheel_file)])
assert result.exit_code
assert "Error: More than 1 or no dist-info files found" in result.output
def test_package_pipeline_no_config(
self, fake_repo_path, fake_project_cli, fake_metadata
):
result = CliRunner().invoke(
fake_project_cli.cli,
["pipeline", "create", PIPELINE_NAME, "--skip-config"],
obj=fake_metadata,
)
assert result.exit_code == 0
result = CliRunner().invoke(
fake_project_cli.cli,
["pipeline", "package", PIPELINE_NAME],
obj=fake_metadata,
)
assert result.exit_code == 0
assert f"Pipeline `{PIPELINE_NAME}` packaged!" in result.output
wheel_location = fake_repo_path / "src" / "dist"
assert f"Location: {wheel_location}" in result.output
# the wheel contents are slightly different (config shouldn't be included),
# which is why we can't call self.assert_wheel_contents_correct here
wheel_file = wheel_location / _get_wheel_name(name=PIPELINE_NAME, version="0.1")
assert wheel_file.is_file()
assert len(list((fake_repo_path / "src" / "dist").iterdir())) == 1
wheel_contents = set(ZipFile(str(wheel_file)).namelist())
expected_files = {
f"{PIPELINE_NAME}/__init__.py",
f"{PIPELINE_NAME}/README.md",
f"{PIPELINE_NAME}/nodes.py",
f"{PIPELINE_NAME}/pipeline.py",
"tests/__init__.py",
"tests/test_pipeline.py",
}
assert expected_files <= wheel_contents
assert f"{PIPELINE_NAME}/config/parameters.yml" not in wheel_contents
def test_pull_local_whl_compare(self, fake_kedro_cli, dummy_project, env, alias):
"""
Test for pulling a valid wheel file locally, unpack it into another location and
check that unpacked files are identical to the ones in the original modular pipeline.
"""
# pylint: disable=too-many-locals
pipeline_name = "another_pipeline"
self.call_pipeline_create(fake_kedro_cli)
self.call_pipeline_package(fake_kedro_cli, pipeline_name)
source_path = dummy_project / "src" / PACKAGE_NAME / "pipelines" / PIPELINE_NAME
config_path = dummy_project / "conf" / "base" / "pipelines" / PIPELINE_NAME
test_path = dummy_project / "src" / "tests" / "pipelines" / PIPELINE_NAME
wheel_file = (
dummy_project
/ "src"
/ "dist"
/ _get_wheel_name(name=pipeline_name, version="0.1")
)
assert wheel_file.is_file()
options = ["-e", env] if env else []
options += ["--alias", alias] if alias else []
result = CliRunner().invoke(
fake_kedro_cli.cli, ["pipeline", "pull", str(wheel_file), *options]
)
assert result.exit_code == 0
pipeline_name = alias or pipeline_name
source_dest = dummy_project / "src" / PACKAGE_NAME / "pipelines" / pipeline_name
config_env = env or "base"
config_dest = dummy_project / "conf" / config_env / "pipelines" / pipeline_name
test_dest = dummy_project / "src" / "tests" / "pipelines" / pipeline_name
assert not filecmp.dircmp(source_path, source_dest).diff_files
assert not filecmp.dircmp(config_path, config_dest).diff_files
assert not filecmp.dircmp(test_path, test_dest).diff_files
def test_pipeline_package_version(self, fake_repo_path, fake_package_path,
fake_project_cli, fake_metadata):
_pipeline_name = "data_engineering"
# the test version value is set separately in
# features/steps/test_starter/<repo>/src/<package>/pipelines/data_engineering/__init__.py
_test_version = "4.20.69"
pipelines_dir = fake_package_path / "pipelines" / _pipeline_name
assert pipelines_dir.is_dir()
result = CliRunner().invoke(fake_project_cli,
["pipeline", "package", _pipeline_name],
obj=fake_metadata)
assert result.exit_code == 0
# test for actual version
wheel_location = fake_repo_path / "src" / "dist"
wheel_name = _get_wheel_name(name=_pipeline_name,
version=_test_version)
wheel_file = wheel_location / wheel_name
assert wheel_file.is_file()
assert len(list(wheel_location.iterdir())) == 1
def test_pipeline_package_overwrites_wheel(self, fake_kedro_cli, tmp_path):
destination = (tmp_path / "in" / "here").resolve()
destination.mkdir(parents=True)
wheel_file = destination / _get_wheel_name(name=PIPELINE_NAME, version="0.1")
wheel_file.touch()
result = CliRunner().invoke(
fake_kedro_cli.cli, ["pipeline", "create", PIPELINE_NAME]
)
assert result.exit_code == 0
result = CliRunner().invoke(
fake_kedro_cli.cli,
["pipeline", "package", PIPELINE_NAME, "--destination", str(destination)],
)
assert result.exit_code == 0
warning_message = f"Package file {wheel_file} will be overwritten!"
success_message = (
f"Pipeline `{PIPELINE_NAME}` packaged! Location: {destination}"
)
assert warning_message in result.output
assert success_message in result.output
self.assert_wheel_contents_correct(wheel_location=destination)
def test_pull_from_pypi(
self,
fake_project_cli,
fake_repo_path,
mocker,
tmp_path,
fake_package_path,
env,
alias,
fake_metadata,
):
"""
Test for pulling a valid wheel file from pypi.
"""
# pylint: disable=too-many-locals
self.call_pipeline_create(fake_project_cli.cli, fake_metadata)
# We mock the `pip download` call, and manually create a package wheel file
# to simulate the pypi scenario instead
self.call_pipeline_package(
fake_project_cli.cli, fake_metadata, destination=tmp_path
)
wheel_file = tmp_path / _get_wheel_name(name=PIPELINE_NAME, version="0.1")
assert wheel_file.is_file()
self.call_pipeline_delete(fake_project_cli.cli, fake_metadata)
source_path = fake_package_path / "pipelines" / PIPELINE_NAME
test_path = fake_repo_path / "src" / "tests" / "pipelines" / PIPELINE_NAME
source_params_config = (
fake_repo_path / CONF_ROOT / "base" / "parameters" / f"{PIPELINE_NAME}.yml"
)
# Make sure the files actually deleted before pulling from pypi.
assert not source_path.exists()
assert not test_path.exists()
assert not source_params_config.exists()
python_call_mock = mocker.patch("kedro.framework.cli.pipeline.python_call")
mocker.patch(
"kedro.framework.cli.pipeline.tempfile.TemporaryDirectory",
return_value=tmp_path,
)
options = ["-e", env] if env else []
options += ["--alias", alias] if alias else []
result = CliRunner().invoke(
fake_project_cli.cli,
["pipeline", "pull", PIPELINE_NAME, *options],
obj=fake_metadata,
)
assert result.exit_code == 0
python_call_mock.assert_called_once_with(
"pip", ["download", "--no-deps", "--dest", str(tmp_path), PIPELINE_NAME],
)
pipeline_name = alias or PIPELINE_NAME
source_dest = fake_package_path / "pipelines" / pipeline_name
test_dest = fake_repo_path / "src" / "tests" / "pipelines" / pipeline_name
config_env = env or "base"
dest_params_config = (
fake_repo_path
/ CONF_ROOT
/ config_env
/ "parameters"
/ f"{pipeline_name}.yml"
)
self.assert_package_files_exist(source_dest)
assert dest_params_config.is_file()
actual_test_files = {f.name for f in test_dest.iterdir()}
expected_test_files = {"__init__.py", "test_pipeline.py"}
assert actual_test_files == expected_test_files
def test_pull_config_missing(
self,
fake_project_cli,
fake_repo_path,
fake_package_path,
env,
alias,
fake_metadata,
):
"""
Test for pulling a valid wheel file locally, but `config` directory is missing
from the wheel file.
"""
# pylint: disable=too-many-locals
self.call_pipeline_create(fake_project_cli.cli, fake_metadata)
source_params_config = (
fake_repo_path / CONF_ROOT / "base" / "parameters" / f"{PIPELINE_NAME}.yml"
)
source_params_config.unlink()
self.call_pipeline_package(fake_project_cli.cli, fake_metadata)
self.call_pipeline_delete(fake_project_cli.cli, fake_metadata)
source_path = fake_package_path / "pipelines" / PIPELINE_NAME
test_path = fake_repo_path / "src" / "tests" / "pipelines" / PIPELINE_NAME
# Make sure the files actually deleted before pulling from the wheel file.
assert not source_path.exists()
assert not test_path.exists()
wheel_file = (
fake_repo_path
/ "src"
/ "dist"
/ _get_wheel_name(name=PIPELINE_NAME, version="0.1")
)
assert wheel_file.is_file()
options = ["-e", env] if env else []
options += ["--alias", alias] if alias else []
result = CliRunner().invoke(
fake_project_cli.cli,
["pipeline", "pull", str(wheel_file), *options],
obj=fake_metadata,
)
assert result.exit_code == 0
pipeline_name = alias or PIPELINE_NAME
source_dest = fake_package_path / "pipelines" / pipeline_name
test_dest = fake_repo_path / "src" / "tests" / "pipelines" / pipeline_name
config_env = env or "base"
dest_params_config = (
fake_repo_path
/ CONF_ROOT
/ config_env
/ "parameters"
/ f"{pipeline_name}.yml"
)
self.assert_package_files_exist(source_dest)
assert not dest_params_config.exists()
actual_test_files = {f.name for f in test_dest.iterdir()}
expected_test_files = {"__init__.py", "test_pipeline.py"}
assert actual_test_files == expected_test_files
def test_pull_local_whl_compare(
self,
fake_project_cli,
fake_repo_path,
fake_package_path,
env,
alias,
fake_metadata,
):
"""Test for pulling a valid wheel file locally, unpack it
into another location and check that unpacked files
are identical to the ones in the original modular pipeline.
"""
# pylint: disable=too-many-locals
pipeline_name = "another_pipeline"
call_pipeline_create(fake_project_cli, fake_metadata)
call_pipeline_package(fake_project_cli, fake_metadata, alias=pipeline_name)
source_path = fake_package_path / "pipelines" / PIPELINE_NAME
test_path = fake_repo_path / "src" / "tests" / "pipelines" / PIPELINE_NAME
source_params_config = (
fake_repo_path
/ settings.CONF_ROOT
/ "base"
/ "parameters"
/ f"{PIPELINE_NAME}.yml"
)
wheel_file = (
fake_repo_path
/ "src"
/ "dist"
/ _get_wheel_name(name=pipeline_name, version="0.1")
)
assert wheel_file.is_file()
options = ["-e", env] if env else []
options += ["--alias", alias] if alias else []
result = CliRunner().invoke(
fake_project_cli,
["pipeline", "pull", str(wheel_file), *options],
obj=fake_metadata,
)
assert result.exit_code == 0, result.output
assert "pulled and unpacked" in result.output
pipeline_name = alias or pipeline_name
source_dest = fake_package_path / "pipelines" / pipeline_name
test_dest = fake_repo_path / "src" / "tests" / "pipelines" / pipeline_name
config_env = env or "base"
dest_params_config = (
fake_repo_path
/ settings.CONF_ROOT
/ config_env
/ "parameters"
/ f"{pipeline_name}.yml"
)
assert not filecmp.dircmp(source_path, source_dest).diff_files
assert not filecmp.dircmp(test_path, test_dest).diff_files
assert source_params_config.read_bytes() == dest_params_config.read_bytes()
