docs: scoped discovery

[v-rocheleau]

Discovery configuration hierarchy

Discovery configs can be provided at the following levels:

• Node:
  • Mounted JSON file
  • None => no public data
• Project:
  • Optional at project creation
  • None => node discovery fallback
• Dataset:
  • Optional at dataset creation
  • None => project discovery fallback

When querying a scoped Katsu endpoint, the user can scope the query for a specific project/dataset by using the project and/or dataset query parameters. This does not apply to endpoints that expect URL parameters for project/dataset IDs, like datasets/<str:dataset_id>/summary.

Katsu handles scoped queries this way:

1. {endpoint}?dataset={dataset_uuid}
   • uses the dataset discovery config
2. {endpoint}?project={project_uuid}&dataset={dataset_uuid}
   • uses the dataset discovery config
   • the dataset MUST belong to the project
3. {endpoint}?project={project_uuid}
   • uses the project's discovery config
4. {endpoint} OR {endpoint}?project=""&dataset=""
   • uses the node's discovery config
   • compatibility with older bento_web and bento_public versions

Katsu endpoints scoping

The following endpoints can now be queried with the project and/or dataset query parameters.

1. datasets/str:dataset_id/summary
2. public_search_fields
3. search_overview
4. public_overview
5. overview
6. public
7. beacon_search

New public endpoint on katsu:

• public_rules: returns the scoped public rules (max query params and count threshold)
  • related gateway PR

[davidlougheed]

minor, otherwise looks good

[davidlougheed]

lgtm