Move extraction code for SBND PTB HLT and LLT into sbndcode; remove need for CAFMaker to access it via artDAQ

[kjplows]

The coupling between experiment specific code and high-level code is too high in this module. See comments below from @PetrilloAtWork

I am going to approve the PR as I suppose that's "urgently" needed.
However, I invite the author to rethink the pattern and to update it: here we have a direct extraction from raw detector-specific data to the highest analysis level. That is a very strong coupling.
My recommendation is to move the extraction code into sbndcode producer module which puts the extracted information into a data product (std::vector<PTBInfo_t>), allowing CAF maker and anything else to access it without the need for artDAQ and the knowledge of low-level details on the data format.

Originally posted by @PetrilloAtWork in #614 (review)