Linked ProtoParticle creation to selections

After #160 (closed), we need to be able to use the ProtoParticle containers created in the HLT2 job in the HLT2 line selections. Right now the lines are hard-coded to take the reconstruction from the input file (as created by Brunel).

This might need a bit of configuration juggling so that we can switch between "use the reconstruction from the file" and "run the reconstruction on the fly".