Address a couple of spatial upload hiccups

[adkinsrs]

Observations from loading a larger spatial dataset

• For visiumhd, simplify the tar.gz structure. Instead of requiring a specific structure, make sure the files are present, then extract to the area they need to be.
• The tarball was extracted pretty quickly but the processing step was lengthy, and needs to be written out to the jobs status.json file
• After file processing is done and the zarr is created, a preliminary analysis is run. This is also lengthy but the "process" step in the UI showed as "complete". Need to communicate this step is running
• @JPReceveur ran into a potential issues with add_primary_analysis_to_dataset_upload.cgi where the gene input file had duplications, but this was not reported in the uploader. Need to propogate these errors back up to the UI, or for this particular step, perform this check before uploading.