Skip merging projects with multiplexed libraries

[allyhawkins]

Closes #716 Closes #717

Here I added a line to filter out any projects that have multiplexed libraries from the channel that is used as input to create merged objects.

Because we have decided against including multiplexed libraries being included in merged objects, I removed any support we had for merging multiplexed libraries in the merge_sces.R script. I was a little torn on if we should keep this, but if the plan is to eventually create merged objects for all non-multiplexed libraries in a project, then we don't need this anymore. I can add it back if we want the flexibility to use it in the future, but I think a better change to that script would be to spit out an error if any libraries being merged contained multiplexed data when we address #718.

As part of that, I also removed the columns to keep referring to cell hash stats and demuxing results, resolving the issue in #717.

[allyhawkins]

I updated this to use the branching you suggested and then also removed the changes to the script, so that should be the same now, other than addressing the capitalization of ID for the demuxing columns.

I am testing this with running one multiplexed and one non-multiplexed libraries, and I think I like the log.warn? It comes up as a message pretty immediately when running so you can see why not as many jobs started as you might expect if a multiplexed library is there: