gui layer#

The gui layer provides GUI frontend for meas layer (or inst layer sometimes). This is implementated as mahos.gui package. We use Qt as the GUI toolkit, and frequently use PyQtGraph for data visualization.

Qt has its own (inter-or-intra thread) communication mechanism called signal and slot. The GUI event (such as this button is clicked) emits the signal. If signal is connected to slots (or methods), the methods are called.

Basic structure#

Because GUI is basically a client of the nodes, we have to implement two functions: Req (Requester) and Sub (Subscriber). The Req is relatively simple, because Req-Rep can be implemented as single method call (it’s a slot in Qt terminology). For Sub, it becomes quite tidy if we emit the signal on subscribed message’s arrival. Let’s see this point by observing the structure of an example IVCurveWidget from Tutorial 2: Basic Measurement.

The QBasicMeasClient can be used as Qt-version client of BasicMeasNode. As noted, Req-Rep looks quite simple. When startButton is clicked, the click signal invokes request_start() method (slot), which subsequently calls QBasicMeasClient.change_state() to send the request. The data topic published by IVCurve is subscribed by QStatusDataSubWorker, which is working in a dedicated thread (different from main thread running GUI main loop). When data arrives, QStatusDataSubWorker emits dataUpdated signal and received by the QBasicMeasClient.check_data() (here, inter-thread communication is done). check_data() again emits dataUpdated signal, which eventually updates the data visualized by plot_item.

It is important that QBasicMeasClient does all the mahos communication stuffs (Req-Rep and Pub-Sub); it converts Qt-communication to mahos-communication in other words. As a result, IVCurveWidget don’t have to care about mahos communications and can focus on using signals and slots provided by QBasicMeasClient.