I have a computer vision project, built with python and OpenCV, that I would like to use robotframework to perform tests. For example, I’d like one case to be creating a “Capture” object and setting some properties on it and have the test flag for any properties that failed to be set.
But I haven’t really grasped how I could use RF for this? None of the standard libraries/keywords etc would allow for this kind of testing? So would I need to create my own custom keyword library?
Hi Ben,
To help you better we’ll need to better understand what you’re testing
- Is it purely a software application solution that takes inputs from cameras?
- Or is it an application embedded into a device and you need to test the device as well as the software?
1 is easier than 2, but the approach to both is quite different.
Go back to basics and think about the steps you’d take to test your test case manually, if you can list the steps for an example test case here, then you go step by step and identify how to automate each step into a keyword, then it’s simply a matter of joining them together in a test case.
For example if it’s test type 2 a physical device and your first step is capture an image, I’d probably start with placing a monitor from the robot framework machine in front of the camera and getting robot framework to display a test image on that screen (this is easily accomplished with existing robot framework keywords), if it’s more complex you might also then need to trigger a button press you might need to create a custom keyword to activate a solenoid that presses the button.
Even if you can’t list the steps for an example test case here, hopefully this has helped.
Dave.