Unit testing a PyQt standalone "blocking" dialog

I'm creating a series of "standalone PyQt widgets" that can be used in procedural programs; the original idea is from the easygui project. To give a concrete idea of their use, instead of writing:

>>> name = input("What is your name? )

I could write

>>> name = get_string("What is your name? )

and a dialog would pop up, inviting the user to enter the response. Like Python's input, get_string is blocking the execution of the program, waiting for input.

I would like to set up some automatic testing of these widgets. I tried with a third-party module (pyautogui) which interacts with GUI programs, but the result is not totally reliable and probably not suitable for using a service like https://travis-ci.org/. I would prefer to use PyQt's QTest but do not know how to connect with the dialog: I suspect I may have to use threading (as I had to do with the pyautogui solution) as the dialog is effectively blocking the execution of the program.

Here is a simple implementation of get_string() mentioned above:

from PyQt4 import QtGui

def get_string(prompt="What is your name? ", title="Title",
               default_response="PyQt4", app=None):
    """GUI equivalent of input()."""

    if app is None:
        app = QtGui.QApplication([])
    app.dialog = QtGui.QInputDialog()
    text, ok = app.dialog.getText(None, title, prompt,
                              QtGui.QLineEdit.Normal,
                              default_response)
    app.quit()
    if ok:
        return text

if __name__ == '__main__':
    print(get_string())  # normal blocking mode
    app2 = QtGui.QApplication([])
    # perhaps start a delayed thread here, using QTest
    print(get_string(app=app2))

I have found one example of unit-testing a PyQt application (http://johnnado.com/pyqt-qtest-example/) but it did not help me find a solution in terms of connecting to a dialog.

Any help would be appreciated.