Visualizing study logic

When we are faced with complex data we always feel lost, aren't we? We don't know anything about the essence of it but still we need to find what is wrong with it. Why does it not work as expected? It could be that the signals are wrongly mapped or the logic of a block is wrong. It could be a bug. It could be many things but since we don't have any information about the logic inside we don't know where to start. It would be very nice if we had a way to visualize the logic graph behind the study. How signals are routed between different resources? Which one goes where? Which ones are inputs and which ones are outputs? Which ones are provided by a PLC? etc. Eventually this may help to find our way easily or maybe discourage us. Imagine looking at an enormously complex graph. Will you be able to tell anything based on it? So many resources interconnected with so many signals. You can hardly tell apart one line from the rest ...

In an attempt to provide such a solution I came across the DOT file format: https://en.wikipedia.org/wiki/DOT_(graph_description_language) It seems designed exactly for the purpose. It uses a simple textual language to describe the different nodes and links (edges as they call them) between them. In addition a set of attributes can provide how the nodes should be visualized. Needless to mention I gave it a try. The result is a new command (Export Study Logic Graph). In version 2512 it sits behind the STUDY_LOGIC_GRAPH feature toggle. The command simply creates a .dot file that you can load in a DOT file reader like Graphviz (https://dreampuf.github.io/GraphvizOnline/)

The implementation is not perfect I admit. It has a lot of way to go to be fully functional. For now only signals, sensors, robots, SCL blocks and LB blocks are covered. Still with this initial version I could finally see the topology of the study logic.

Remember the demo cell with three robots?

Here is how it partially looks in terms of logic:

You can clearly see what goes in/out of the PLC and one of the robots. The same is true for the logic blocks and the sensors. I chose the color coding randomly:

1. grey - signal

2. light blue - input of a logic block

3. light green - output of a logic block

4. blue - logic block boundary

5. violet - sensor

6. orange - robot/plc

7. black - edges

What immediately became obvious is that this small study produces a quite big logic graph. Imagine really big studies. It takes time to produce the file and it takes even more time to visualize it. Unfortunately this is a downside that may limit the usage of this tool at the end. The good thing is that you can download the graph as an image and review it later.

Images also become large ...

Anyway one day this can become an invaluable tool when reviewing customer data. It is a directional graph so you can see the exact flow of data.

We have not seen anything like it yet. I hope we could bring it into production sooner or later.

Try it!