Add /links #95
Add /links #95
Conversation
Since map_back is the same for several endpoints, it has been moved to ResourceMapper and will only be modified (at this point) for the /links endpoint mapper.
TODO: Turn providers.json into proper MongoDB entries
Codecov Report
@@ Coverage Diff @@
## master #95 +/- ##
==========================================
- Coverage 85.74% 84.99% -0.76%
==========================================
Files 38 39 +1
Lines 2315 2352 +37
==========================================
+ Hits 1985 1999 +14
- Misses 330 353 +23
Continue to review full report at Codecov.
|
| def get_links(request: Request, params: EntryListingQueryParams = Depends()): | ||
| for str_param in ["filter", "sort"]: | ||
| if getattr(params, str_param): | ||
| setattr(params, str_param, "") | ||
| for int_param in [ | ||
| "page_offset", | ||
| "page_page", | ||
| "page_cursor", | ||
| "page_above", | ||
| "page_below", | ||
| ]: | ||
| if getattr(params, int_param): | ||
| setattr(params, int_param, 0) | ||
| params.page_limit = CONFIG.page_limit | ||
| return u.get_entries(links, LinksResponse, request, params) |
There was a problem hiding this comment.
Looks like you ran into the same problems as me, there must be a better way of doing this!
There was a problem hiding this comment.
I think there is in pydantic v1 maybe? Here's hoping.
|
I've just added two validator changes to this PR: one that checks the new links endpoint, and one minor change that stops the validator from failing if there are no references in the database. |
Better documentation of `map_back` and _only_ use `task_id` for structures ids.
CasperWA
force-pushed
the
close_89_add_links
branch
from
21752a7
to
8ce3e80
Compare
This should not have been in this PR and has now been removed. It's purpose is for the upcoming "extra" server, serving the index meta-database. A PR will be created for this after this and #99 have been merged. |
Closes #89
Several things are happening here:
map_backin the parent class. This should provide an idea of going forward and perhaps removing the sub-classed mappers or similar./linksendpoint with a pseudoindexparent implementation and all the providers fromproviders.json.Comments:
I am unsure on how to treat the
/linksendpoint concerning whether/links/{entry_id}is valid. And why is it demanded that only theSingleEntryQueryParamsare valid query parameters for the endpoint, (see here)?I think some of these questions should be put to the consortium and changes should be suggested to the spec.
Missing:/linksendpoint. (Added by @ml-evs)