feat: search improvements #3583
Conversation
…euse and dataservice
Working towards having better tests for `udata-search-service` integration (useful for #3583)
maudetes
left a comment
maudetes
left a comment
There was a problem hiding this comment.
Most of my comments are on consistency, but the overall code is quite clean! 🎉
| "topic": ModelTermsFilter(model=Topic), | ||
| "access_type": Filter(), | ||
| "format_family": Filter(choices=list(FormatFamily)), | ||
| "producer_type": Filter(choices=list(PRODUCER_TYPES)), |
There was a problem hiding this comment.
Don't we have confusing similar filters between organization_badge and producer_type? Should we deprecate the first one if we use producer_type instead?
| return dataservice.is_visible | ||
|
|
||
| @classmethod | ||
| def fetch_documentation_content(cls, url: str) -> str | None: |
There was a problem hiding this comment.
I have some notes here:
- the documentation content is stored at index time only. We can ignore this for now as this is minor but it means hard-to-spot de-synchronization.
- Shouldn't we check for mime-type?
- Should we activate it for certified orgs only? I'm not very satisfied about indexing anything anywhere
Do you think we could have a better approach, maybe more job oriented and not at the index time?
There was a problem hiding this comment.
As a general note, I think the idea was to have the same filters on mongo list vs search service search.
We're adding some filters on search service only here, but we can try to converge when we refacto the search service?
| "description": organization.description, | ||
| "url": organization.url, | ||
| "badges": [badge.kind for badge in organization.badges], | ||
| "producer_type": producer_types, |
There was a problem hiding this comment.
I don't think we need producer_types on organization if we already have an existing filter on badges? Example: https://www.data.gouv.fr/organizations?badge=association
…s, add organization_name to dataservices
No description provided.