pytest-qt

Repository:GitHub
Version:2.0.0
License:MIT
Author:Bruno Oliveira

Introduction

pytest-qt is a pytest plugin that provides fixtures to help programmers write tests for PySide and PyQt.

The main usage is to use the qtbot fixture, which provides methods to simulate user interaction, like key presses and mouse clicks:

def test_hello(qtbot):
    widget = HelloWidget()
    qtbot.addWidget(widget)

    # click in the Greet button and make sure it updates the appropriate label
    qtbot.mouseClick(window.button_greet, QtCore.Qt.LeftButton)

    assert window.greet_label.text() == 'Hello!'

Requirements

Python 2.7 or later, including Python 3.4+.

Requires pytest version 2.7 or later.

Works with either PyQt5, PySide or PyQt4, picking whichever is available on the system giving preference to the first one installed in this order:

  • PyQt5
  • PySide
  • PyQt4

To force a particular API, set the configuration variable qt_api in your pytest.ini file to pyqt5, pyside, pyqt4 or pyqt4v2. pyqt4v2 sets the PyQt4 API to version 2.

[pytest]
qt_api=pyqt5

Alternatively, you can set the PYTEST_QT_API environment variable to the same values described above (the environment variable wins over the configuration if both are set).

Installation

The package may be installed by running:

pip install pytest-qt

Or alternatively, download the package from pypi, extract and execute:

python setup.py install

Both methods will automatically register it for usage in pytest.

Development

If you intend to develop pytest-qt itself, use virtualenv to activate a new fresh environment and execute:

git clone https://github.com/pytest-dev/pytest-qt.git
cd pytest-qt
python setup.py develop
pip install pyside # or pyqt4/pyqt5

Versioning

This projects follows semantic versioning.

Tutorial

pytest-qt registers a new fixture named qtbot, which acts as bot in the sense that it can send keyboard and mouse events to any widgets being tested. This way, the programmer can simulate user interaction while checking if GUI controls are behaving in the expected manner.

To illustrate that, consider a widget constructed to allow the user to find files in a given directory inside an application.

_images/find_files_dialog.png

It is a very simple dialog, where the user enters a standard file mask, optionally enters file text to search for and a button to browse for the desired directory. Its source code is available here,

To test this widget’s basic functionality, create a test function:

def test_basic_search(qtbot, tmpdir):
    '''
    test to ensure basic find files functionality is working.
    '''
    tmpdir.join('video1.avi').ensure()
    tmpdir.join('video1.srt').ensure()

    tmpdir.join('video2.avi').ensure()
    tmpdir.join('video2.srt').ensure()

Here the first parameter indicates that we will be using a qtbot fixture to control our widget. The other parameter is pytest’s standard tmpdir that we use to create some files that will be used during our test.

Now we create the widget to test and register it:

window = Window()
window.show()
qtbot.addWidget(window)

Tip

Registering widgets is not required, but recommended because it will ensure those widgets get properly closed after each test is done.

Now we use qtbot methods to simulate user interaction with the dialog:

window.fileComboBox.clear()
qtbot.keyClicks(window.fileComboBox, '*.avi')

window.directoryComboBox.clear()
qtbot.keyClicks(window.directoryComboBox, str(tmpdir))

The method keyClicks is used to enter text in the editable combo box, selecting the desired mask and directory.

We then simulate a user clicking the button with the mouseClick method:

qtbot.mouseClick(window.findButton, QtCore.Qt.LeftButton)

Once this is done, we inspect the results widget to ensure that it contains the expected files we created earlier:

assert window.filesTable.rowCount() == 2
assert window.filesTable.item(0, 0).text() == 'video1.avi'
assert window.filesTable.item(1, 0).text() == 'video2.avi'

Qt Logging Capture

New in version 1.4.

Qt features its own logging mechanism through qInstallMessageHandler (qInstallMsgHandler on Qt4) and qDebug, qWarning, qCritical functions. These are used by Qt to print warning messages when internal errors occur.

pytest-qt automatically captures these messages and displays them when a test fails, similar to what pytest does for stderr and stdout and the pytest-catchlog plugin. For example:

from pytestqt.qt_compat import qWarning

def do_something():
    qWarning('this is a WARNING message')

def test_foo():
    do_something()
    assert 0

$ py.test test.py -q
F
================================== FAILURES ===================================
_________________________________ test_types __________________________________

    def test_foo():
        do_something()
>       assert 0
E       assert 0

test.py:8: AssertionError
---------------------------- Captured Qt messages -----------------------------
QtWarningMsg: this is a WARNING message
1 failed in 0.01 seconds

Disabling Logging Capture

Qt logging capture can be disabled altogether by passing the --no-qt-log to the command line, which will fallback to the default Qt bahavior of printing emitted messages directly to stderr:

py.test test.py -q --no-qt-log
F
================================== FAILURES ===================================
_________________________________ test_types __________________________________

    def test_foo():
        do_something()
>       assert 0
E       assert 0

test.py:8: AssertionError
---------------------------- Captured stderr call -----------------------------
this is a WARNING message

qtlog fixture

pytest-qt also provides a qtlog fixture that can used to check if certain messages were emitted during a test:

def do_something():
    qWarning('this is a WARNING message')

def test_foo(qtlog):
    do_something()
    emitted = [(m.type, m.message.strip()) for m in qtlog.records]
    assert emitted == [(QtWarningMsg, 'this is a WARNING message')]

qtlog.records is a list of Record instances.

Logging can also be disabled on a block of code using the qtlog.disabled() context manager, or with the pytest.mark.no_qt_log mark:

def test_foo(qtlog):
    with qtlog.disabled():
        # logging is disabled within the context manager
        do_something()

@pytest.mark.no_qt_log
def test_bar():
    # logging is disabled for the entire test
    do_something()

Keep in mind that when logging is disabled, qtlog.records will always be an empty list.

Log Formatting

The output format of the messages can also be controlled by using the --qt-log-format command line option, which accepts a string with standard {} formatting which can make use of attribute interpolation of the record objects:

$ py.test test.py --qt-log-format="{rec.when} {rec.type_name}: {rec.message}"

Keep in mind that you can make any of the options above the default for your project by using pytest’s standard addopts option in you pytest.ini file:

[pytest]
qt_log_format = {rec.when} {rec.type_name}: {rec.message}

Automatically failing tests when logging messages are emitted

Printing messages to stderr is not the best solution to notice that something might not be working as expected, specially when running in a continuous integration server where errors in logs are rarely noticed.

You can configure pytest-qt to automatically fail a test if it emits a message of a certain level or above using the qt_log_level_fail ini option:

[pytest]
qt_log_level_fail = CRITICAL

With this configuration, any test which emits a CRITICAL message or above will fail, even if no actual asserts fail within the test:

from pytestqt.qt_compat import qCritical

def do_something():
    qCritical('WM_PAINT failed')

def test_foo(qtlog):
    do_something()

>py.test test.py --color=no -q
F
================================== FAILURES ===================================
__________________________________ test_foo ___________________________________
test.py:5: Failure: Qt messages with level CRITICAL or above emitted
---------------------------- Captured Qt messages -----------------------------
QtCriticalMsg: WM_PAINT failed

The possible values for qt_log_level_fail are:

  • NO: disables test failure by log messages.
  • DEBUG: messages emitted by qDebug function or above.
  • WARNING: messages emitted by qWarning function or above.
  • CRITICAL: messages emitted by qCritical function only.

If some failures are known to happen and considered harmless, they can be ignored by using the qt_log_ignore ini option, which is a list of regular expressions matched using re.search:

[pytest]
qt_log_level_fail = CRITICAL
qt_log_ignore =
    WM_DESTROY.*sent
    WM_PAINT failed

py.test test.py --color=no -q
.
1 passed in 0.01 seconds

Messages which do not match any of the regular expressions defined by qt_log_ignore make tests fail as usual:

def do_something():
    qCritical('WM_PAINT not handled')
    qCritical('QObject: widget destroyed in another thread')

def test_foo(qtlog):
    do_something()

py.test test.py --color=no -q
F
================================== FAILURES ===================================
__________________________________ test_foo ___________________________________
test.py:6: Failure: Qt messages with level CRITICAL or above emitted
---------------------------- Captured Qt messages -----------------------------
QtCriticalMsg: WM_PAINT not handled  (IGNORED)
QtCriticalMsg: QObject: widget destroyed in another thread

You can also override the qt_log_level_fail setting and extend qt_log_ignore patterns from pytest.ini in some tests by using a mark with the same name:

def do_something():
    qCritical('WM_PAINT not handled')
    qCritical('QObject: widget destroyed in another thread')

@pytest.mark.qt_log_level_fail('CRITICAL')
@pytest.mark.qt_log_ignore('WM_DESTROY.*sent', 'WM_PAINT failed')
def test_foo(qtlog):
    do_something()

If you would like to override the list of ignored patterns instead, pass extend=False to the qt_log_ignore mark:

@pytest.mark.qt_log_ignore('WM_DESTROY.*sent', extend=False)
def test_foo(qtlog):
    do_something()

waitSignal: Waiting for threads, processes, etc.

New in version 1.2.

If your program has long running computations running in other threads or processes, you can use qtbot.waitSignal to block a test until a signal is emitted (such as QThread.finished) or a timeout is reached. This makes it easy to write tests that wait until a computation running in another thread or process is completed before ensuring the results are correct:

def test_long_computation(qtbot):
    app = Application()

    # Watch for the app.worker.finished signal, then start the worker.
    with qtbot.waitSignal(app.worker.finished, timeout=10000) as blocker:
        blocker.connect(app.worker.failed)  # Can add other signals to blocker
        app.worker.start()
        # Test will block at this point until either the "finished" or the
        # "failed" signal is emitted. If 10 seconds passed without a signal,
        # SignalTimeoutError will be raised.

    assert_application_results(app)

raising parameter

New in version 1.4.

Changed in version 2.0.

You can pass raising=False to avoid raising a qtbot.SignalTimeoutError if the timeout is reached before the signal is triggered:

def test_long_computation(qtbot):
    ...
    with qtbot.waitSignal(app.worker.finished, raising=False) as blocker:
        app.worker.start()

    assert_application_results(app)

    # qtbot.SignalTimeoutError is not raised, but you can still manually
    # check whether the signal was triggered:
    assert blocker.signal_triggered, "process timed-out"

qt_wait_signal_raising ini option

New in version 1.11.

Changed in version 2.0.

The qt_wait_signal_raising ini option can be used to override the default value of the raising parameter of the qtbot.waitSignal and qtbot.waitSignals functions when omitted:

[pytest]
qt_wait_signal_raising = false

Calls which explicitly pass the raising parameter are not affected.

Getting arguments of the emitted signal

New in version 1.10.

The arguments emitted with the signal are available as the args attribute of the blocker:

def test_signal(qtbot):
    ...
    with qtbot.waitSignal(app.got_cmd) as blocker:
        app.listen()
    assert blocker.args == ['test']

Signals without arguments will set args to an empty list. If the time out is reached instead, args will be None.

waitSignals

New in version 1.4.

If you have to wait until all signals in a list are triggered, use qtbot.waitSignals, which receives a list of signals instead of a single signal. As with qtbot.waitSignal, it also supports the raising parameter:

def test_workers(qtbot):
    workers = spawn_workers()
    with qtbot.waitSignals([w.finished for w in workers]):
        for w in workers:
            w.start()

    # this will be reached after all workers emit their "finished"
    # signal or a qtbot.SignalTimeoutError will be raised
    assert_application_results(app)

Making sure a given signal is not emitted

New in version 1.11.

If you want to ensure a signal is not emitted in a given block of code, use the qtbot.assertNotEmitted context manager:

def test_no_error(qtbot):
    ...
    with qtbot.assertNotEmitted(app.worker.error):
        app.worker.start()

waitUntil: Waiting for arbitrary conditions

New in version 2.0.

Sometimes your tests need to wait a certain condition which does not trigger a signal, for example that a certain control gained focus or a QListView has been populated with all items.

For those situations you can use qtbot.waitUntil to wait until a certain condition has been met or a timeout is reached. This is specially important in X window systems due to their asynchronous nature, where you can’t rely on the fact that the result of an action will be immediately available.

For example:

def test_validate(qtbot):
    window = MyWindow()
    window.edit.setText('not a number')
    # after focusing, should update status label
    window.edit.setFocus()
    assert window.status.text() == 'Please input a number'

The window.edit.setFocus() may not be processed immediately, only in a future event loop, which might lead to this test to work sometimes and fail in others (a flaky test).

A better approach in situations like this is to use qtbot.waitUntil with a callback with your assertion:

def test_validate(qtbot):
    window = MyWindow()
    window.edit.setText('not a number')
    # after focusing, should update status label
    window.edit.setFocus()
    def check_label():
        assert window.status.text() == 'Please input a number'
    qtbot.waitUntil(check_label)

qtbot.waitUntil will periodically call check_label until it no longer raises AssertionError or a timeout is reached. If a timeout is reached, the last assertion error re-raised and the test will fail:

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
    def check_label():
>       assert window.status.text() == 'Please input a number'
E       assert 'OK' == 'Please input a number'
E         - OK
E         + Please input a number

A second way to use qtbot.waitUntil is to pass a callback which returns True when the condition is met or False otherwise. It is usually terser than using a separate callback with assert statement, but it produces a generic message when it fails because it can’t make use of pytest‘s assertion rewriting:

def test_validate(qtbot):
    window = MyWindow()
    window.edit.setText('not a number')
    # after focusing, should update status label
    window.edit.setFocus()
    qtbot.waitUntil(lambda: window.edit.hasFocus())
    assert window.status.text() == 'Please input a number'

Exceptions in virtual methods

New in version 1.1.

It is common in Qt programming to override virtual C++ methods to customize behavior, like listening for mouse events, implement drawing routines, etc.

Fortunately, both PyQt and PySide support overriding this virtual methods naturally in your python code:

class MyWidget(QWidget):

    # mouseReleaseEvent
    def mouseReleaseEvent(self, ev):
        print('mouse released at: %s' % ev.pos())

This works fine, but if python code in Qt virtual methods raise an exception PyQt4 and PySide will just print the exception traceback to standard error, since this method is called deep within Qt’s event loop handling and exceptions are not allowed at that point. In PyQt5.5+, exceptions in virtual methods will by default call abort(), which will crash the interpreter.

This might be surprising for python users which are used to exceptions being raised at the calling point: for example, the following code will just print a stack trace without raising any exception:

class MyWidget(QWidget):

    def mouseReleaseEvent(self, ev):
        raise RuntimeError('unexpected error')

w = MyWidget()
QTest.mouseClick(w, QtCore.Qt.LeftButton)

To make testing Qt code less surprising, pytest-qt automatically installs an exception hook which captures errors and fails tests when exceptions are raised inside virtual methods, like this:

E           Failed: Qt exceptions in virtual methods:
E           ________________________________________________________________________________
E             File "x:\pytest-qt\pytestqt\_tests\test_exceptions.py", line 14, in event
E               raise RuntimeError('unexpected error')
E
E           RuntimeError: unexpected error

Disabling the automatic exception hook

You can disable the automatic exception hook on individual tests by using a qt_no_exception_capture marker:

@pytest.mark.qt_no_exception_capture
def test_buttons(qtbot):
    ...

Or even disable it for your entire project in your pytest.ini file:

[pytest]
qt_no_exception_capture = 1

This might be desirable if you plan to install a custom exception hook.

Note

Starting with PyQt5.5, exceptions raised during virtual methods will actually trigger an abort(), crashing the Python interpreter. For this reason, disabling exception capture in PyQt5.5+ is not recommended unless you install your own exception hook.

Model Tester

New in version 2.0.

pytest-qt includes a fixture that helps testing QAbstractItemModel implementations. The implementation is copied from the C++ code as described on the Qt Wiki, and it continuously checks a model as it changes, helping to verify the state and catching many common errors the moment they show up.

Some of the conditions caught include:

  • Verifying X number of rows have been inserted in the correct place after the signal rowsAboutToBeInserted() says X rows will be inserted.
  • The parent of the first index of the first row is a QModelIndex()
  • Calling index() twice in a row with the same values will return the same QModelIndex
  • If rowCount() says there are X number of rows, model test will verify that is true.
  • Many possible off by one bugs
  • hasChildren() returns true if rowCount() is greater then zero.
  • and many more...

To use it, create an instance of your model implementation, fill it with some items and call qtmodeltester.check:

def test_standard_item_model(qtmodeltester):
    model = QStandardItemModel()
    items = [QStandardItem(str(i)) for i in range(4)]
    model.setItem(0, 0, items[0])
    model.setItem(0, 1, items[1])
    model.setItem(1, 0, items[2])
    model.setItem(1, 1, items[3])
    qtmodeltester.check(model)

If the tester finds a problem the test will fail with an assert pinpointing the issue.

The following attribute may influence the outcome of the check depending on your model implementation:

  • data_display_may_return_none (default: False): While you can technically return None (or an invalid QVariant) from data() for QtCore.Qt.DisplayRole, this usually is a sign of a bug in your implementation. Set this variable to True if this really is OK in your model.

The source code was ported from modeltest.cpp by Florian Bruhin, many thanks!

A note about QApplication.exit()

Some pytest-qt features, most notably waitSignal and waitSignals, depend on the Qt event loop being active. Calling QApplication.exit() from a test will cause the main event loop and auxiliary event loops to exit and all subsequent event loops to fail to start. This is a problem if some of your tests call an application functionality that calls QApplication.exit().

One solution is to monkeypatch QApplication.exit() in such tests to ensure it was called by the application code but without effectively calling it.

For example:

def test_exit_button(qtbot, monkeypatch):
    exit_calls = []
    monkeypatch.setattr(QApplication, 'exit', lambda: exit_calls.append(1))
    button = get_app_exit_button()
    qtbot.click(button)
    assert exit_calls == [1]

Or using the mock package:

def test_exit_button(qtbot):
    with mock.patch.object(QApplication, 'exit'):
        button = get_app_exit_button()
        qtbot.click(button)
        assert QApplication.exit.call_count == 1

Reference

QtBot

class pytestqt.qtbot.QtBot(request)[source]

Instances of this class are responsible for sending events to Qt objects (usually widgets), simulating user input.

Important

Instances of this class should be accessed only by using a qtbot fixture, never instantiated directly.

Widgets

addWidget(widget)[source]

Adds a widget to be tracked by this bot. This is not required, but will ensure that the widget gets closed by the end of the test, so it is highly recommended.

Parameters:widget (QWidget) – Widget to keep track of.

Note

This method is also available as add_widget (pep-8 alias)

waitForWindowShown(widget)[source]

Waits until the window is shown in the screen. This is mainly useful for asynchronous systems like X11, where a window will be mapped to screen some time after being asked to show itself on the screen.

Parameters:widget (QWidget) – Widget to wait on.

Note

In Qt5, the actual method called is qWaitForWindowExposed, but this name is kept for backward compatibility

Note

This method is also available as wait_for_window_shown (pep-8 alias)

stopForInteraction()[source]

Stops the current test flow, letting the user interact with any visible widget.

This is mainly useful so that you can verify the current state of the program while writing tests.

Closing the windows should resume the test run, with qtbot attempting to restore visibility of the widgets as they were before this call.

Note

As a convenience, it is also aliased as stop.

wait(ms)[source]

New in version 1.9.

Waits for ms milliseconds.

While waiting, events will be processed and your test will stay responsive to user interface events or network communication.

Signals and Events

waitSignal(signal=None, timeout=1000, raising=None, check_params_cb=None)[source]

New in version 1.2.

Stops current test until a signal is triggered.

Used to stop the control flow of a test until a signal is emitted, or a number of milliseconds, specified by timeout, has elapsed.

Best used as a context manager:

with qtbot.waitSignal(signal, timeout=1000):
    long_function_that_calls_signal()

Also, you can use the SignalBlocker directly if the context manager form is not convenient:

blocker = qtbot.waitSignal(signal, timeout=1000)
blocker.connect(another_signal)
long_function_that_calls_signal()
blocker.wait()

Any additional signal, when triggered, will make wait() return.

New in version 1.4: The raising parameter.

Parameters:
  • signal (Signal) – A signal to wait for. Set to None to just use timeout.
  • timeout (int) – How many milliseconds to wait before resuming control flow.
  • raising (bool) – If QtBot.SignalTimeoutError should be raised if a timeout occurred. This defaults to True unless qt_wait_signal_raising = false is set in the config.
  • check_params_cb (Callable) – Optional callable(*parameters) that compares the provided signal parameters to some expected parameters. It has to match the signature of signal (just like a slot function would) and return True if parameters match, False otherwise.
Returns:

SignalBlocker object. Call SignalBlocker.wait() to wait.

Note

Cannot have both signals and timeout equal None, or else you will block indefinitely. We throw an error if this occurs.

Note

This method is also available as wait_signal (pep-8 alias)

waitSignals(signals=None, timeout=1000, raising=None, check_params_cbs=None, order='none')[source]

New in version 1.4.

Stops current test until all given signals are triggered.

Used to stop the control flow of a test until all (and only all) signals are emitted or the number of milliseconds specified by timeout has elapsed.

Best used as a context manager:

with qtbot.waitSignals([signal1, signal2], timeout=1000):
    long_function_that_calls_signals()

Also, you can use the MultiSignalBlocker directly if the context manager form is not convenient:

blocker = qtbot.waitSignals(signals, timeout=1000)
long_function_that_calls_signal()
blocker.wait()
Parameters:
  • signals (list) – A list of Signal objects to wait for. Set to None to just use timeout.
  • timeout (int) – How many milliseconds to wait before resuming control flow.
  • raising (bool) – If QtBot.SignalTimeoutError should be raised if a timeout occurred. This defaults to True unless qt_wait_signal_raising = false is set in the config.
  • check_params_cbs (list) – optional list of callables that compare the provided signal parameters to some expected parameters. Each callable has to match the signature of the corresponding signal in signals (just like a slot function would) and return True if parameters match, False otherwise. Instead of a specific callable, None can be provided, to disable parameter checking for the corresponding signal. If the number of callbacks doesn’t match the number of signals ValueError will be raised.
  • order (str) –

    Determines the order in which to expect signals:

    • "none": no order is enforced
    • "strict": signals have to be emitted strictly in the provided order (e.g. fails when expecting signals [a, b] and [a, a, b] is emitted)
    • "simple": like “strict”, but signals may be emitted in-between the provided ones, e.g. expected signals == [a, b, c] and actually emitted signals = [a, a, b, a, c] works (would fail with "strict").
Returns:

MultiSignalBlocker object. Call MultiSignalBlocker.wait() to wait.

Note

Cannot have both signals and timeout equal None, or else you will block indefinitely. We throw an error if this occurs.

Note

This method is also available as wait_signals (pep-8 alias)

assertNotEmitted(*args, **kwds)[source]

New in version 1.11.

Make sure the given signal doesn’t get emitted.

This is intended to be used as a context manager.

Note

This method is also available as assert_not_emitted (pep-8 alias)

waitUntil(callback, timeout=1000)[source]

New in version 2.0.

Wait in a busy loop, calling the given callback periodically until timeout is reached.

callback() should raise AssertionError to indicate that the desired condition has not yet been reached, or just return None when it does. Useful to assert until some condition is satisfied:

def view_updated():
    assert view_model.count() > 10
qtbot.waitUntil(view_updated)

Another possibility is for callback() to return True when the desired condition is met, False otherwise. Useful specially with lambda for terser code, but keep in mind that the error message in those cases is usually not very useful because it is not using an assert expression.

qtbot.waitUntil(lambda: view_model.count() > 10)

Note that this usage only accepts returning actual True and False values, so returning an empty list to express “falseness” raises an ValueError.

Parameters:
  • callback – callable that will be called periodically.
  • timeout – timeout value in ms.
Raises:

ValueError – if the return value from the callback is anything other than None, True or False.

Note

This method is also available as wait_until (pep-8 alias)

Raw QTest API

Methods below provide very low level functions, as sending a single mouse click or a key event. Those methods are just forwarded directly to the QTest API. Consult the documentation for more information.

—

Below are methods used to simulate sending key events to widgets:

static keyClick(widget, key[, modifier=Qt.NoModifier[, delay=-1]])
static keyClicks(widget, key sequence[, modifier=Qt.NoModifier[, delay=-1]])
static keyEvent(action, widget, key[, modifier=Qt.NoModifier[, delay=-1]])
static keyPress(widget, key[, modifier=Qt.NoModifier[, delay=-1]])
static keyRelease(widget, key[, modifier=Qt.NoModifier[, delay=-1]])

Sends one or more keyword events to a widget.

Parameters:
  • widget (QWidget) – the widget that will receive the event
  • key (str|int) – key to send, it can be either a Qt.Key_* constant or a single character string.
Parameters:
  • modifier (Qt.KeyboardModifier) –

    flags OR’ed together representing other modifier keys also pressed. Possible flags are:

    • Qt.NoModifier: No modifier key is pressed.
    • Qt.ShiftModifier: A Shift key on the keyboard is pressed.
    • Qt.ControlModifier: A Ctrl key on the keyboard is pressed.
    • Qt.AltModifier: An Alt key on the keyboard is pressed.
    • Qt.MetaModifier: A Meta key on the keyboard is pressed.
    • Qt.KeypadModifier: A keypad button is pressed.
    • Qt.GroupSwitchModifier: X11 only. A Mode_switch key on the keyboard is pressed.
  • delay (int) – after the event, delay the test for this miliseconds (if > 0).
static keyToAscii(key)

Auxilliary method that converts the given constant ot its equivalent ascii.

Parameters:key (Qt.Key_*) – one of the constants for keys in the Qt namespace.
Return type:str
Returns:the equivalent character string.

Note

This method is not available in PyQt.

—

Below are methods used to simulate sending mouse events to widgets.

static mouseClick(widget, button[, stateKey=0[, pos=QPoint()[, delay=-1]]])
static mouseDClick(widget, button[, stateKey=0[, pos=QPoint()[, delay=-1]]])
static mouseEvent(action, widget, button, stateKey, pos[, delay=-1])
static mouseMove(widget[, pos=QPoint()[, delay=-1]])
static mousePress(widget, button[, stateKey=0[, pos=QPoint()[, delay=-1]]])
static mouseRelease(widget, button[, stateKey=0[, pos=QPoint()[, delay=-1]]])

Sends a mouse moves and clicks to a widget.

Parameters:
  • widget (QWidget) – the widget that will receive the event
  • button (Qt.MouseButton) –

    flags OR’ed together representing the button pressed. Possible flags are:

    • Qt.NoButton: The button state does not refer to any button (see QMouseEvent.button()).
    • Qt.LeftButton: The left button is pressed, or an event refers to the left button. (The left button may be the right button on left-handed mice.)
    • Qt.RightButton: The right button.
    • Qt.MidButton: The middle button.
    • Qt.MiddleButton: The middle button.
    • Qt.XButton1: The first X button.
    • Qt.XButton2: The second X button.
  • modifier (Qt.KeyboardModifier) – flags OR’ed together representing other modifier keys also pressed. See keyboard modifiers.
  • position (QPoint) – position of the mouse pointer.
  • delay (int) – after the event, delay the test for this miliseconds (if > 0).

SignalBlocker

class pytestqt.wait_signal.SignalBlocker(timeout=1000, raising=True, check_params_cb=None)[source]

Returned by pytestqt.qtbot.QtBot.waitSignal() method.

Variables:
  • timeout (int) – maximum time to wait for a signal to be triggered. Can be changed before wait() is called.
  • signal_triggered (bool) – set to True if a signal (or all signals in case of MultipleSignalBlocker) was triggered, or False if timeout was reached instead. Until wait() is called, this is set to None.
  • raising (bool) –

    If SignalTimeoutError should be raised if a timeout occurred.

    Note

    contrary to the parameter of same name in pytestqt.qtbot.QtBot.waitSignal(), this parameter does not consider the qt_wait_signal_raising ini option.

  • args (list) – The arguments which were emitted by the signal, or None if the signal wasn’t emitted at all.

New in version 1.10: The args attribute.

wait()

Waits until either a connected signal is triggered or timeout is reached.

Raises:ValueError – if no signals are connected and timeout is None; in this case it would wait forever.
connect(signal)[source]

Connects to the given signal, making wait() return once this signal is emitted.

More than one signal can be connected, in which case any one of them will make wait() return.

Parameters:signal – QtCore.Signal

MultiSignalBlocker

class pytestqt.wait_signal.MultiSignalBlocker(timeout=1000, raising=True, check_params_cbs=None, order='none')[source]

Returned by pytestqt.qtbot.QtBot.waitSignals() method, blocks until all signals connected to it are triggered or the timeout is reached.

Variables identical to SignalBlocker:
  • timeout
  • signal_triggered
  • raising
wait()

Waits until either a connected signal is triggered or timeout is reached.

Raises:ValueError – if no signals are connected and timeout is None; in this case it would wait forever.

SignalTimeoutError

class pytestqt.wait_signal.SignalTimeoutError[source]

New in version 1.4.

The exception thrown by pytestqt.qtbot.QtBot.waitSignal() if the raising parameter has been given and there was a timeout.

Record

class pytestqt.logging.Record(msg_type, message, ignored, context)[source]

Hold information about a message sent by one of Qt log functions.

Variables:
  • message (str) – message contents.
  • type (Qt.QtMsgType) – enum that identifies message type
  • type_name (str) – type as string: "QtDebugMsg", "QtWarningMsg" or "QtCriticalMsg".
  • log_type_name (str) – type name similar to the logging package: DEBUG, WARNING and CRITICAL.
  • when (datetime.datetime) – when the message was captured
  • ignored (bool) – If this record matches a regex from the “qt_log_ignore” option.
  • context – a namedtuple containing the attributes file, function, line. Only available in Qt5, otherwise is None.

Changelog

2.0

Breaking Changes

With pytest-qt 2.0, we changed some defaults to values we think are much better.

However, this required some backwards-incompatible changes:

  • pytest-qt now defaults to using PyQt5 if PYTEST_QT_API is not set. Before, it preferred PySide which is using the discontinued Qt4.
  • Python 3 versions prior to 3.4 are no longer supported.
  • The @pytest.mark.qt_log_ignore mark now defaults to extend=True, i.e. extends the patterns defined in the config file rather than overriding them. You can pass extend=False to get the old behaviour of overriding the patterns.
  • qtbot.waitSignal now defaults to raising=True and raises an exception on timeouts. You can set qt_wait_signal_raising = false in your config to get back the old behaviour.
  • PYTEST_QT_FORCE_PYQT environment variable is no longer supported. Set PYTEST_QT_API to the appropriate value instead.

New Features

  • From this version onward, pytest-qt is licensed under the MIT license (#134).
  • New qtmodeltester fixture to test QAbstractItemModel subclasses. Thanks @The-Compiler for the initiative and port of the original C++ code for ModelTester (#63).
  • New qtbot.waitUntil method, which continuously calls a callback until a condition is met or a timeout is reached. Useful for testing asynchronous features (like in X window environments for example).
  • waitSignal and waitSignals can receive an optional callback that can evaluate if the arguments of emitted signals should resume execution or not. Thanks @MShekow for the PR (#141).
  • Now which Qt binding pytest-qt will use can be configured by the qt_api config option. Thanks @The-Compiler for the request (#129).
  • While pytestqt.qt_compat is an internal module and shouldn’t be imported directly, it is known that some test suites did import it. This module now uses a lazy-load mechanism to load Qt classes and objects, so the old symbols (QtCore, QApplication, etc.) are no longer available from it.

Other Changes

  • Exceptions caught by pytest-qt in sys.excepthook are now also printed to stderr, making debugging them easier from within an IDE. Thanks @fabioz for the PR (126)!

1.11.0

Note

The default value for raising is planned to change to True starting in pytest-qt version 1.12. Users wishing to preserve the current behavior (raising is False by default) should make use of the new qt_wait_signal_raising ini option below.

  • New qt_wait_signal_raising ini option can be used to override the default value of the raising parameter of the qtbot.waitSignal and qtbot.waitSignals functions when omitted:

    [pytest]
qt_wait_signal_raising = true

    Calls which explicitly pass the raising parameter are not affected. Thanks @The-Compiler for idea and initial work on a PR (120).

  • qtbot now has a new assertNotEmitted context manager which can be used to ensure the given signal is not emitted (92). Thanks @The-Compiler for the PR!

1.10.0

  • SignalBlocker now has a args attribute with the arguments of the signal that triggered it, or None on a time out (115). Thanks @billyshambrook for the request and @The-Compiler for the PR.
  • MultiSignalBlocker is now properly disconnects from signals upon exit.

1.9.0

  • Exception capturing now happens as early/late as possible in order to catch all possible exceptions (including fixtures)(105). Thanks @The-Compiler for the request.
  • Widgets registered by qtbot.addWidget are now closed before all other fixtures are tear down (106). Thanks @The-Compiler for request.
  • qtbot now has a new wait method which does a blocking wait while the event loop continues to run, similar to QTest::qWait. Thanks @The-Compiler for the PR (closes 107)!
  • raise RuntimeError instead of ImportError when failing to import any Qt binding: raising the latter causes pluggy in pytest-2.8 to generate a subtle warning instead of a full blown error. Thanks @Sheeo for bringing this problem to attention (closes 109).

1.8.0

  • pytest.mark.qt_log_ignore now supports an extend parameter that will extend the list of regexes used to ignore Qt messages (defaults to False). Thanks @The-Compiler for the PR (99).
  • Fixed internal error when interacting with other plugins that raise an error, hiding the original exception (98). Thanks @The-Compiler for the PR!
  • Now pytest-qt is properly tested with PyQt5 on Travis-CI. Many thanks to @The-Compiler for the PR!

1.7.0

  • PYTEST_QT_API can now be set to pyqt4v2 in order to use version 2 of the PyQt4 API. Thanks @montefra for the PR (93)!

1.6.0

  • Reduced verbosity when exceptions are captured in virtual methods (77, thanks @The-Compiler).
  • pytestqt.plugin has been split in several files (74) and tests have been moved out of the pytestqt package. This should not affect users, but it is worth mentioning nonetheless.
  • QApplication.processEvents() is now called before and after other fixtures and teardown hooks, to better try to avoid non-processed events from leaking from one test to the next. (67, thanks @The-Compiler).
  • Show Qt/PyQt/PySide versions in pytest header (68, thanks @The-Compiler!).
  • Disconnect SignalBlocker functions after its loop exits to ensure second emissions that call the internal functions on the now-garbage-collected SignalBlocker instance (#69, thanks @The-Compiler for the PR).

1.5.1

  • Exceptions are now captured also during test tear down, as delayed events will get processed then and might raise exceptions in virtual methods; this is specially problematic in PyQt5.5, which changed the behavior to call abort by default, which will crash the interpreter. (65, thanks @The-Compiler).

1.5.0

  • Fixed log line number in messages, and provide better contextual information in Qt5 (55, thanks @The-Compiler);
  • Fixed issue where exceptions inside a waitSignals or waitSignal with-statement block would be swallowed and a SignalTimeoutError would be raised instead. (59, thanks @The-Compiler for bringing up the issue and providing a test case);
  • Fixed issue where the first usage of qapp fixture would return None. Thanks to @gqmelo for noticing and providing a PR;
  • New qtlog now sports a context manager method, disabled (58). Thanks @The-Compiler for the idea and testing;

1.4.0

  • Messages sent by qDebug, qWarning, qCritical are captured and displayed when tests fail, similar to pytest-catchlog. Also, tests can be configured to automatically fail if an unexpected message is generated.
  • New method waitSignals: will block untill all signals given are triggered (thanks @The-Compiler for idea and complete PR).
  • New parameter raising to waitSignals and waitSignals: when True will raise a qtbot.SignalTimeoutError exception when timeout is reached (defaults to False). (thanks again to @The-Compiler for idea and complete PR).
  • pytest-qt now requires pytest version >= 2.7.

Internal changes to improve memory management

  • QApplication.exit() is no longer called at the end of the test session and the QApplication instance is not garbage collected anymore;
  • QtBot no longer receives a QApplication as a parameter in the constructor, always referencing QApplication.instance() now; this avoids keeping an extra reference in the qtbot instances.
  • deleteLater is called on widgets added in QtBot.addWidget at the end of each test;
  • QApplication.processEvents() is called at the end of each test to make sure widgets are cleaned up;

1.3.0

  • pytest-qt now supports PyQt5!

    Which Qt api will be used is still detected automatically, but you can choose one using the PYTEST_QT_API environment variable (the old PYTEST_QT_FORCE_PYQT is still supported for backward compatibility).

    Many thanks to @jdreaver for helping to test this release!

1.2.3

  • Now the module ``qt_compat`` no longer sets QString and QVariant APIs to 2 for PyQt, making it compatible for those still using version 1 of the API.

1.2.2

  • Now it is possible to disable automatic exception capture by using markers or a pytest.ini option. Consult the documentation for more information. (26, thanks @datalyze-solutions for bringing this up).
  • QApplication instance is created only if it wasn’t created yet (21, thanks @fabioz!)
  • addWidget now keeps a weak reference its widgets (20, thanks @fabioz)

1.2.1

  • Fixed 16: a signal emitted immediately inside a waitSignal block now works as expected (thanks @baudren).

1.2.0

This version include the new waitSignal function, which makes it easy to write tests for long running computations that happen in other threads or processes:

def test_long_computation(qtbot):
    app = Application()

    # Watch for the app.worker.finished signal, then start the worker.
    with qtbot.waitSignal(app.worker.finished, timeout=10000) as blocker:
        blocker.connect(app.worker.failed)  # Can add other signals to blocker
        app.worker.start()
        # Test will wait here until either signal is emitted, or 10 seconds has elapsed

    assert blocker.signal_triggered  # Assuming the work took less than 10 seconds
    assert_application_results(app)

Many thanks to @jdreaver for discussion and complete PR! (12, 13)

1.1.1

  • Added stop as an alias for stopForInteraction (10, thanks @itghisi)

  • Now exceptions raised in virtual methods make tests fail, instead of silently passing (11). If an exception is raised, the test will fail and it exceptions that happened inside virtual calls will be printed as such:

    E           Failed: Qt exceptions in virtual methods:
E           ________________________________________________________________________________
E             File "x:\pytest-qt\pytestqt\_tests\test_exceptions.py", line 14, in event
E               raise ValueError('mistakes were made')
E
E           ValueError: mistakes were made
E           ________________________________________________________________________________
E             File "x:\pytest-qt\pytestqt\_tests\test_exceptions.py", line 14, in event
E               raise ValueError('mistakes were made')
E
E           ValueError: mistakes were made
E           ________________________________________________________________________________

    Thanks to @jdreaver for request and sample code!

  • Fixed documentation for QtBot: it was not being rendered in the docs due to an import error.

1.1.0

Python 3 support.

1.0.2

Minor documentation fixes.

1.0.1

Small bug fix release.

1.0.0

First working version.