testing framework

[larsbuntemeyer]

Just some ideas, could we maybe use a python test framework for data validation? That would be nice, e.g., to generate reports. If we don't want to stop on the first assertion, we could use, e.g.

• unittest, subtests
• pytest-subtests
• pandera (mostly for pandas)
• allure
• xarray-schema
• tsdat
• pydantic
• compliance-checker plugin (https://github.com/AtMoDat/cc-plugin-cmip6-cv)

[larsbuntemeyer]

Example to use subtests:

import unittest

class Car(object):
  def __init__(self, make, model):
    self.make = make
    self.model = make  # Copy and paste error: should be model.
    self.has_seats = True
    self.wheel_count = 3  # Typo: should be 4.

class CarTest(unittest.TestCase):
  def test_init(self):
    make = "Ford"
    model = "Model T"
    car = Car(make=make, model=model)
    with self.subTest(msg='Car.make check'):
        self.assertEqual(car.make, make)
    with self.subTest(msg='Car.model check'):
        self.assertEqual(car.model, model)
    with self.subTest(msg='Car.has_seats check'):
        self.assertTrue(car.has_seats)
    with self.subTest(msg='Car.wheel_count check'):
        self.assertEqual(car.wheel_count, 4)

if __name__ == "__main__":
    unittest.main()

[larsbuntemeyer]

examples using pydantic:

from pydantic import BaseModel, field_validator
class User(BaseModel):
    username: str
    password: str
    age: int
    @field_validator('password')
    def password_must_be_strong(cls, v):
        if len(v) < 16:
            raise ValueError('Password must be at least 16 characters long.')
            return v
    @field_validator('username')
    def username_must_be_strong(cls, v):
        if len(v) < 4:
            raise ValueError('Username must be at least 4 characters long.')
            return v

# Validate incoming user_data
user_data = {'username': 'App', 'password': 'password', 'age': 25}
user = User(**user_data)

[larsbuntemeyer]

it seems that data validation for n-d array/xarray is not really implemented yet. I guess for now, i will simply work with a logger and logging levels...

[larsbuntemeyer]

using conftest for input parameters with pytest:

import glob


def pytest_addoption(parser):
    parser.addoption(
        "--filename",
        action="store",
        type=str,
        help="list of files to pass to test functions",
    )
    parser.addoption("--cv", action="store", default="default name")


def pytest_generate_tests(metafunc):
    if "filename" in metafunc.fixturenames:
        metafunc.parametrize(
            "filename", glob.glob(str(metafunc.config.getoption("filename")))
        )

run with pytest, e.g.

pytest -s -v cmor_check/tests/test_cmor.py --filename "/work/bb1203/g300046_CMOR/_CMOR/NUKLEUS/output/EUR-11/GERICS/ECMWF-ERA5/evaluation/r1i1p1f1/GERICS-REMO2020/v1/day/tas/v20240402/tas_E    UR-11_ECMWF-ERA5_evaluation_r1i1p1f1_GERICS-REMO2020_v1_day_*" --cv /work/bb1203/g300046_CMOR/cmor-tables/Tables/CORDEX-CMIP6_CV.json --html=report.html --self-contained-html