Sampling in parallel when a model reads outputs from the file

Our group is also looking for answer to getting the chain id to reference an input file. Any answers been discovered?