The opsim output contents are described here:
(there is only one important table of observations in current databases).
As to what the best approach is after that … I guess it depends on what the existing tool needs or does.
The opsim schema gives you the overall pointing; you can then determine whether a particular RA/Dec falls onto science silicon (onto a chip) using the rubin_sim camera footprint tool. A demonstration of that tool is in the Camera Footprint Demo notebook.
You might want to use your tool to see if an object is within 2.1 degrees of the center of these pointings, and then apply the camera footprint (this is approximately what the SSSC simulator and the code in rubin_sim.movingObjects do).
But if you really want to add the 189 CCDs (fixed: 189 CCDs! many more amps…) for the ~2M pointings, you could likely modify how we use the Camera Footprint or extend it to use the underlying map (which defines the corners of each CCD) to pass back the approximate CCD footprint on the sky (if you want an exact camera footprint on the sky, it might make sense to go a step further and use the full DM camera model, instead of what is essentially our tangent plane approximation based on the geometry of the focal plane).