The first pass uploader will be direct S3 uploads from filesystem: #188
After that, it would be great to have pull options rather than only push:
- Provide a URL.
- Link to Google Drive.
- Link to Dropbox file.
Implementation
- While there may be some virtue in integrating with Google / Dropbox login APIs, it significantly complicates things.
- My personal opinion is that we mandate the raw data be made public. The pull based imagery upload should work in the same way for all datasets (pretty much), by simply downloading a publicly accessible dataset via URL (no matter where it's hosted).
- Perhaps there are complications to simply downloading from Google / Dropbox as a URL that I have missed - to be assessed.
The first pass uploader will be direct S3 uploads from filesystem: #188
After that, it would be great to have pull options rather than only push:
Implementation