在 Flask 单元测试中，如何模拟请求全局“g"对象上的对象?

2022-01-08 00:00:00 python mocking flask unit-testing werkzeug

问题描述

我有一个烧瓶应用程序，它在 before_filter 中设置数据库连接，非常类似于 这个:

I have a flask application that is setting up a database connection in a before_filter, very similar to this:

@app.before_request
def before_request():
    g.db = connect_db()

现在:我正在编写一些单元测试，但我确实不希望它们进入数据库.我想用可以设置期望的模拟对象替换 g.db.

Now: I am writing some unit-tests and I do not want them to hit the database. I want to replace g.db with a mock object that I can set expectations on.

我的测试使用 app.test_client()，正如烧瓶文档 这里.示例测试类似于

My tests are using app.test_client(), as is demonstrated in the flask documentation here. An example test looks something like

def test(self):
    response = app.test_client().post('/endpoint', data={..})
    self.assertEqual(response.status_code, 200)
    ...

测试工作并通过，但它们正在访问数据库，正如我所说，我想用模拟对象替换 db 访问.我在 test_client 中看不到任何访问 g 对象或更改 before_filters 的方法.

The tests work and pass, but they are hitting the database and as I said I want to replace db access with mock objects. I do not see any way in test_client to access the g object or alter the before_filters.

解决方案

这可行

test_app.py

from flask import Flask, g

app = Flask(__name__)

def connect_db():
    print 'I ended up inside the actual function'
    return object()

@app.before_request
def before_request():
    g.db = connect_db()


@app.route('/')
def root():
    return 'Hello, World'

test.py

from mock import patch
import unittest

from test_app import app


def not_a_db_hit():
    print 'I did not hit the db'

class FlaskTest(unittest.TestCase):

    @patch('test_app.connect_db')
    def test_root(self, mock_connect_db):
        mock_connect_db.side_effect = not_a_db_hit
        response = app.test_client().get('/')
        self.assertEqual(response.status_code, 200)

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

所以这将打印出我没有命中数据库"，而不是我最终进入了实际函数".显然，您需要根据实际用例调整模拟.

So this will print out 'I did not hit the db', rather than 'I ended up inside the actual function'. Obviously you'll need to adapt the mocks to your actual use case.

相关文章