Implement personal data archive

[gwennlbh]

To do:

• Rename endpoint to GET /personal_data_archive
• Implementation:
  • Get all data from COLLECTIONS where owner_key is the current user
  • Zip it
  • Send the bytes with the appropriate MIME type
• Rate-limit it aggressively (within legal bounds, see GDPR. EDIT: since no personally-identifiable info is stored outside of /users/, the data doesn't fall under GDPR)
• Define a model for the response (so that we have an idea of what the response is in the docs)