Simulate hdf5 file and hdf5 group with MagicMock in python3

Question

I am trying to simulate the reading of an hdf5 file in python3 for testing. The function controls, if the file contains all the required keywords. The structure of the file is:

• Group: parent
  • Dataset: A
  • Dataset: B

My plan is to use MagicMock object for the file object and the group object.

Testfile:

def test_checkInput(self):
    file = MagicMock()
    my_values = MagicMock()

    my_values.keys.return_value = ['A', 'B']
    file.keys.return_value = {'parent': my_values}

    self.assertTrue(reader_class.checkInput(file))

reader_class is the module where the function checkInput(file) is defined

reader_class - file

def checkInput(file):
    if not file.keys:
        return False

    if 'parent' in file.keys():
        group = file['parent']
        if 'A' in group.keys():
            if 'B' in group.keys():
                return True
            else:
                print(f"[ {__file__} ] : No Group: 'A'")
                return False
        else:
            print(f"[ {__file__} ] : No Group: 'B'")
            return False
    else:
        print(f"[ {__file__} ] : No Group: 'parent'")
        return False

The problem is, that group.keys() does not return ['A', 'B'] as expected in the checkInput function. It returns a MagicMock-Object instead. How can a get the set values?

Answer 1

The problem was, that the MagicMock objects in the the reader_class file are different. file.keys() returns another mock as file['parent'] I found a simple solution: in the Testfile just use one MagicMock object and specify the return values

def test_checkInput(self):
    file = MagicMock()    
    file.keys.return_value = 'parent'
    file['parent'].keys.return_value = ['A', 'B']

    self.assertTrue(reader_class.checkInput(file))