Re-introduce already-demultiplexed uploads and clarify upload types

[adamcantor22]

Is your feature request related to a problem? Please describe. As of PR #397, data uploads happen like this:

1. User selects 'Upload New Sequencing Run' and chooses a name.
2. User presented with a screen with four file upload sections (for_read, rev_read, for_barcode, rev_barcode) not all of which must be filled out.
3. User submits upload Having four file upload fields, three of which are considered 'optional' is confusing to the user. Also, it prevents users from uploading already-demultiplexed data.

Describe the solution you'd like Before the upload page, the user should be prompted with reads and barcodes types. Then on the upload page, there should be the correct number of file inputs based on that data, with none of the fields being optional. Also, the user should have the option to upload already-demultiplexed files as a zip instead of the raw sequencing run data.

[cleme]

Maybe we can simplify the solution by providing a list of possible upload strategies, which can then request the required files. Something like:

• non-demultiplexed, single barcodes
• non-demultiplexed, dual barcodes
• demultiplexed, single barcodes
• demultiplexed, dual barcodess Based on the user selection, it is then trivial to request the appropriate number of files.

[adamcantor22]

There are sections of code built to handle already-demuxed zips in server.py, but they were using the old framework of 1:1 studies to runs. So there may be helpful lines of code in there, but it'll need to be completely restructured.