Check the channel id usage for pixel gain and pixel threshold

[YifanC]

Check the channel id usage for pixel gain and pixel threshold (with input files). Make sure we are not only reading the values for the first module.

[jaafar-chakrani]

I confirmed during the validation of the pixel thresholds that they are properly set on the expected channels for all modules. The only subtlety that I would like to mention is related to pixel IDs for Module 2:

• For Module 0, the pixel IDs go from 0 to 2 * N_PIXELS[0] * N_PIXELS[1] - 1 with N_PIXELS = (140, 280), i.e. from 0 to 78,399.
• For Module 1, the pixel IDs go from 2 * N_PIXELS[0] * N_PIXELS[1] to 4 * N_PIXELS[0] * N_PIXELS[1] - 1 with N_PIXELS unchanged, i.e. from 78,400 to 156,799.
• For Module 2, the pixel IDs go from 4 * N_PIXELS[0] * N_PIXELS[1] to 6 * N_PIXELS[0] * N_PIXELS[1] - 1, except that N_PIXELS gets reloaded and now would correspond to N_PIXELS = (160, 320), so they now range from 204,800 to 409,599.
• For Module 3, the pixel IDs go from 6 * N_PIXELS[0] * N_PIXELS[1] to 8 * N_PIXELS[0] * N_PIXELS[1] - 1, but N_PIXELS gets reloaded back to N_PIXELS = (140, 280), and the range is from 235,800 to 313,599.

So it may look like there is an overlap in pixel IDs between Module 2 and Module 3, but I think it's fine since larnd-sim processes one module at a time independently and this doesn't seem to have any impact. It's only important to keep in mind when we want to make pixel gains/thresholds inputs.

[krwood]

IMO, we should try to address this... I understand @jaafar-chakrani's point about having some protection due to module-by-module batching, but that's a configurable parameter in the simulation and we should try to avoid putting in these types of "unspoken assumptions". I.e. in principal, the simulation should work under different batching schemes. If a user changes the batching scheme on their end, they will be susceptible to bugs. Probably worth sorting this out now to avoid any potential confusion down the road.

We really should try and standardize the unique pixel ID scheme and use it here. We have functions to convert from unique pixel IDs to (io_group, io_channel, chip_id, channel_id) and vice versa in many different places, which is potential for bugs (and we've been bitten by them in the past).