根据 json 文件中的列表测试数据对测试进行参数化

Question

当测试有一个 different/multiple 测试数据列表时，有没有办法参数化测试？

example_test_data.json

{ "test_one" : [1,2,3], # this is the case, where the `test_one` test need to be parametrize.
  "test_two" : "split",
  "test_three" : {"three":3},
  "test_four" : {"four":4},
  "test_set_comparison" : "1234"
}

目录结构：

主要--
- conftest.py # 我的灯具的 conftest 文件
- 测试用例
  - project_1 （包含这些文件 -- test_suite_1.py、config.json）
  - project_2 （包含这些文件 -- test_suite_2.py、config.json）
- 工作流程
  - 库

在顶级目录级别的 conftest.py 中使用以下代码，能够 get/map 来自特定测试用例的 json 文件的测试数据。

@pytest.yield_fixture(scope="class", autouse=True)
def test_config(request):
  f = pathlib.Path(request.node.fspath.strpath)
  print "File : %s" % f
  config = f.with_name("config.json")
  print "Config json file : %s" % config
  with config.open() as fd:
    testdata = json.loads(fd.read())
  print "test data :", testdata
  yield testdata


@pytest.yield_fixture(scope="function", autouse=True)
def config_data(request, test_config):
  testdata = test_config
  test = request.function.__name__
  print "Class Name : %s" % request.cls.__name__
  print "Testcase Name : %s" % test
  if test in testdata:
    test_args = testdata[test]
    yield test_args
  else:
    yield {}

就我而言：

@pytest.yield_fixture(scope="function", autouse=True)
def config_data(request, test_config):
  testdata = test_config
  test = request.function.__name__
  print "Class Name : %s" % request.cls.__name__
  print "Testcase Name : %s" % test
  if test in testdata:
    test_args = testdata[test]
    if isinstance(test_args, list):
       # How to parametrize the test 
       # yield test_args
  else:
    yield {}

Answer 1

我会在 pytest_generate_tests 钩子中处理特殊的参数化情况：

# conftest.py
import json
import pathlib
import pytest


@pytest.fixture(scope="class")
def test_config(request):
    f = pathlib.Path(request.node.fspath.strpath)
    config = f.with_name("config.json")
    with config.open() as fd:
        testdata = json.loads(fd.read())
    yield testdata


@pytest.fixture(scope="function")
def config_data(request, test_config):
    testdata = test_config
    test = request.function.__name__
    if test in testdata:
        test_args = testdata[test]
        yield test_args
    else:
        yield {}


def pytest_generate_tests(metafunc):
    if 'config_data' not in metafunc.fixturenames:
        return
    config = pathlib.Path(metafunc.module.__file__).with_name('config.json')
    testdata = json.loads(config.read_text())
    param = testdata.get(metafunc.function.__name__, None)
    if isinstance(param, list):
        metafunc.parametrize('config_data', param)

一些注意事项：yield_fixture is deprecated 所以我用普通的 fixture 替换了它。此外，在 return 值的固定装置中不需要 autouse=True - 无论如何你都可以调用它们。

我使用的示例测试和配置：

# testcases/project_1/config.json
{
    "test_one": [1, 2, 3],
    "test_two": "split"
}

# testcases/project_1/test_suite_1.py
def test_one(config_data):
    assert config_data >= 0

def test_two(config_data):
    assert config_data == 'split'


# testcases/project_2/config.json
{
    "test_three": {"three": 3},
    "test_four": {"four": 4}
}

# testcases/project_2/test_suite_2.py
def test_three(config_data):
    assert config_data['three'] == 3

def test_four(config_data):
    assert config_data['four'] == 4

运行测试结果：

$ pytest -vs
============================== test session starts ================================
platform linux -- Python 3.6.5, pytest-3.4.1, py-1.5.3, pluggy-0.6.0 --
/data/gentoo64/usr/bin/python3.6
cachedir: .pytest_cache
rootdir: /data/gentoo64/home/u0_a82/projects/Whosebug/so-50815777, inifile:
plugins: mock-1.6.3, cov-2.5.1
collected 6 items

testcases/project_1/test_suite_1.py::test_one[1] PASSED
testcases/project_1/test_suite_1.py::test_one[2] PASSED
testcases/project_1/test_suite_1.py::test_one[3] PASSED
testcases/project_1/test_suite_1.py::test_two PASSED
testcases/project_2/test_suite_2.py::test_three PASSED
testcases/project_2/test_suite_2.py::test_four PASSED

============================ 6 passed in 0.12 seconds =============================

根据 json 文件中的列表测试数据对测试进行参数化

Parametrize the test based on the list test-data from a json file

python

testing

pytest

python-2.7

example_test_data.json

目录结构：