This process will allow third parties to update in an automated way the identities found in Korma.

So far there's documentation in the Community Wiki regarding to the process to manually update affiliation data. This was done in T88277: Instructions to update user data in korma.

With the use of Sorting Hat this process is simpler than manually editing that file, but this forces you to use Sorting Hat.

The automation is as follows:

• Sorting Hat 'git pull' the last version of the JSON file (bitbuket stored)
• Sorting Hat overwrite information in the database with changes in the JSON file (so Sorting Hat assumes this is always the right information).
  • If there were an issue, we can simply checkout a previous version.
• Sorting Hat add new identities information coming from data sources (for instance a new developer committed a change to the source code)
• Sorting Hat exports that identities/affiliations information and 'git push' that new JSON file to the bitbucket repository.
• Sorting Hat updates the private JSON file (so far in bitbucket) each day this is run.