相关疑难解决方法(0)

简单的例子.两种方法,一种叫另一种方法:

def method_a(arg):
    some_data = method_b(arg)

def method_b(arg):
    return some_data

在Python中,我们可以def在另一个内部声明def.因此,如果method_b需要并且仅从中调用method_a,我应该method_b在内部声明method_a吗？像这样 :

def method_a(arg):

    def method_b(arg):
        return some_data

    some_data = method_b

或者我应该避免这样做？

如果我有这个功能,我应该怎么做用我自己的自定义版本替换内部函数？

def foo():
    def bar():
        # I want to change this
        pass

    # here starts a long list of functions I want to keep unchanged
    def baz():
        pass

使用类可以轻松完成覆盖该方法.虽然,我无法弄清楚如何使用嵌套函数.更改foo为类(或其他任何内容)不是一个选项,因为它来自我无法修改的给定导入模块.

目前，我的 django 项目中有一个 google Drive API 客户端，可以按预期工作。

import unittest
from unittest import mock

DRIVE_API_VERSION = "v3"
DRIVE_API_SERVICE_NAME = "drive"
DRIVE_AUTHORIZED_USER_FILE = "path/to/secrets/json/file"
DRIVE_SCOPES = ['https://www.googleapis.com/auth/drive', 'https://www.googleapis.com/auth/drive.file ', 'https://www.googleapis.com/auth/drive.appdata']

def construct_drive_service():
    try:
        drive_credentials = google.oauth2.credentials.Credentials.from_authorized_user_file(
            DRIVE_AUTHORIZED_USER_FILE, scopes=DRIVE_SCOPES)
    except FileNotFoundError:
        print('Drive credentials not created')
        pass
    if drive_credentials:
        return build(DRIVE_API_SERVICE_NAME, DRIVE_API_VERSION, credentials=drive_credentials, cache_discovery=False)
    else:
        return None

现在我面临的挑战是为这个函数编写测试。但我不知道该使用什么策略。我试过这个

class TestAPICalls(unittest.TestCase):

    @mock.patch('api_calls.google.oauth2.credentials', autospec=True)
    def setUp(self, mocked_drive_cred):
        self.mocked_drive_cred = mocked_drive_cred

    @mock.patch('api_calls.DRIVE_AUTHORIZED_USER_FILE')
    def test_drive_service_creation(self, mocked_file):
        mocked_file.return_value = "some/file.json"
        self.mocked_drive_cred.Credentials.return_value = mock.sentinel.Credentials
        construct_drive_service()
        self.mocked_drive_cred.Credentials.from_authorized_user_file.assert_called_with(mocked_file)

但我的测试失败并出现以下错误

    with io.open(filename, 'r', …

我有一个模块 utils.py，它有这个 run_cmd() 方法

def run_cmd(cmd):
    pipe = subprocess.Popen(cmd,
                            shell=True,
                            stdout=subprocess.PIPE,
                            stderr=subprocess.PIPE)
    print(pipe.communicate())
    print(pipe.returncode)
    stdout, stderr = [stream.strip() for stream in pipe.communicate()]
    output = ' - STDOUT: "%s"' % stdout if len(stdout) > 0 else ''
    error = ' - STDERR: "%s"' % stdout if len(stderr) > 0 else ''
    logger.debug("Running [{command}] returns: [{rc}]{output}{error}".format(
                 command=cmd,
                 rc=pipe.returncode,
                 output=output,
                 error=error))

    return pipe.returncode, stdout, stderr

我使用模拟和此链接stackoverflow作为参考编写了一个单元测试

  @patch('subprocess.Popen')
  @patch('utils.logger.debug')
  def test_run_cmd(self, mock_popen, mock_log):
    cmd = 'mock_command'
    mocked_pipe = Mock()
    attrs = {'communicate.return_value': …

我正在尝试使用Mock来测试从检索设置的函数返回的结果：

def apply_delta_to_date(original_date):
    delta = MySettings.get_delta()
    result_date = original_date + timedelta(delta)
    return result_date

在测试apply_delta_to_date函数时，我试图模拟调用MySettings.get_delta以模拟它返回特定结果：

class TestAppyDelta():
    def setUp(self):
        self.MySettings = Mock()
        self.MySettings.get_delta = Mock(return_value=10)

    def test_apply_delta(self):
        result = apply_delta_to_date(today)

问题是我无法“模拟”MySettings.get_delta()我真正想要测试的函数内部的函数调用。

我如何模拟我正在测试的函数中的内部函数返回的结果？