Support resumable/chunked uploads

We considered supporting uploading via the CLI intermine_boot client, but after some discussion found that it would be more user-friendly and probably just as reliable to use chunked uploads in the browser.

One good library that provides this is [flow.js](https://github.com/flowjs/flow.js).
We will need to have the upload endpoint support the [API](https://github.com/flowjs/flow.js#how-do-i-set-it-up-with-my-server).

The next question would be where to save the files.
- Upload to intermine_compose and save file to local filesystem **(how it's done currently)**
- Upload to intermine_compose and use [GCP’s app engine cloud client library](https://cloud.google.com/appengine/docs/standard/python3/using-cloud-storage) to save to cloud bucket **(we should probably move over to this, as the files will be large and numerous; the previous approach should still be used when this is not available due to running locally with intermine_boot)**
- Upload directly to Google Cloud Storage bucket **(a different alternative; might be more difficult and not sure if it's worth the extra effort)**
- This can be done for large files using either flow.js ([issue](https://github.com/flowjs/flow.js/issues/170), [PR](https://github.com/flowjs/flow.js/pull/276)) or GCS directly via resumable uploads API ([info](https://cloud.google.com/storage/docs/resumable-uploads), [guide](https://cloud.google.com/storage/docs/performing-resumable-uploads))

Note that the files need to be available to both the wizard and InterMine instance without being duplicated. (With a GCP bucket: after [acquiring an OAuth token](https://cloud.google.com/storage/docs/authentication), files can be [downloaded from buckets with a GET request](https://cloud.google.com/storage/docs/downloading-objects#rest-download-object).) This could be a problem with our current Auth approach, so we may need to explore using Daniela’s OAuth server