don't recreate index files when symlinking imported datasets

[crutching]

I'm using bioblend to import datasets to Galaxy histories for processing batches of clinical sequencing data. We are currently looking at processing some of our data through Illumina DRAGEN hardware, and would like to start workflows with BAMs. What I generally do is symlink the BAMs (alongside indexes), keeping them out of the galaxy-managed portion of the filesystem. The problem is that 10-15min per sample of setting metadata on BAMs that have already been indexed is a bit prohibitive for our use case, so I'm looking at workarounds. Can there be an option to not set metadata in these situations? One idea I had is to create a fake datatype (eg. bam_nometa) and then modify all of the downstream tools to accept this new type. Seems pretty hacky, so am trying to avoid it. Thoughts? This is somewhat similar to #6181.