Add chip ID filter to allow decoding data-merged streams

It is already possible to just instantiate 4 DecCores (1 per chip ID) and forward the data frame by frame to the correct one. For full flexibility, it is now also possible to point the instances to the same data and let them filter by chip ID themselves, removing the need to copy the data (which is internally done when giving data frame by frame). Closes #12 (closed) because the rest of the ticket would depend on the use case specifics, which I cannot tell right now.