tests(user-interaction-e2e): add module scoped e2e tests with user interaction; closes #508

tests/end-2-end/conftest.py

@@ -4,8 +4,7 @@ import random
 
 import pytest
 
-from bec_widgets.cli.client_utils import BECGuiClient, _start_plot_process
-from bec_widgets.utils import BECDispatcher
+from bec_widgets.cli.client_utils import BECGuiClient
 
 # pylint: disable=unused-argument
 # pylint: disable=redefined-outer-name
@@ -28,7 +27,7 @@ def gui_id():
     return f"figure_{random.randint(0,100)}"  # make a new gui id each time, to ensure no 'gui is alive' zombie key can perturbate
 
 
-@pytest.fixture
+@pytest.fixture(scope="function")
 def connected_client_gui_obj(qtbot, gui_id, bec_client_lib):
     """
     Fixture to create a new BECGuiClient object and start a server in the background.
@@ -42,22 +41,3 @@ def connected_client_gui_obj(qtbot, gui_id, bec_client_lib):
         yield gui
     finally:
         gui.kill_server()
-
-
-@pytest.fixture(scope="session")
-def connected_gui_with_scope_session(qtbot, gui_id, bec_client_lib):
-    """
-    Fixture to create a new BECGuiClient object and start a server in the background.
-
-    This fixture is scoped to the session, meaning it remains alive for all tests in the session.
-    We can use this fixture to create a gui object that is used across multiple tests, and
-    simulate a real-world scenario where the gui is not restarted for each test.
-    """
-    gui = BECGuiClient(gui_id=gui_id)
-    try:
-        gui.start(wait=True)
-        # After the server started, we need to wait until the bec exists in the namespace
-        qtbot.waitUntil(lambda: hasattr(gui, "bec"), timeout=5000)
-        yield gui
-    finally:
-        gui.kill_server()