Indirect File Uploads

[browjm4]

Reason

Certain customers are interested in using existing file storage as their object storage layer for DeepLynx

Design

1. Upload file metadata to DeepLynx (eg file path, file name, file size, etc) without actually attaching a file to the POST request. The file should live somewhere other than the default file storage directory. If upload "works", see if DL has the permissions to access the file (by attempting to download). This should be tested on both azure and filesystem.
2. If step 1 didn't work, modify the upload file pipeline (and perhaps download) to not be heavily constrained to the /containers/x/datasources/x filepath (but only if an alternative file path is provided)
3. You could probably make a PR at that point

Impact

changes to the behavior supported in the upload file API endpoint (POST /containers/:containerID/import/datasources/:sourceID/files). Add documentation on this process from the user perspective (eg what the json body should look like for the request to get it to work).