ENH: Data management improvements in wave 2 collection

[oesteban]

Just FYI @celprov, so that you can check how to automatically generate PRs out of data. The new script add_session.sh does the following:

1. Takes the path of the DICOM folder downloaded from PACs, and extracts the session code (this is why it is important that the correct 'participant id' is encoded when registering a new patient).
2. Updates the git/git-annex with datalad update and branches into a clean branch called `add/
3. Runs HeuDiConv, based off of the script/parameters we had for wave 1
4. Clean up spurious files added by HeuDiConv and adds the new files to the annex.
5. Pushes to github and to ria-storage. The push to ria-storage is critical: this is when data are uploaded into curnagl
6. Creates a PR with the GitHub CLI tool <-- DON'T MISS THIS, this will be useful in the future to automate stuff and look like a real geek where you work.
7. Compact the DICOMS into a tarball and clean up the uncompressed data (critical to save space on curnagl).

I'm opening it as a draft for now, as I want to improve the documentation.