Simplify retrieve

[ssssarah]

• 2 (main+metadata through resource endpoint) or 3 (main fail through resource endpoint, main+metadata through resolver endpoint) requests were necessary to retrieve resources:

  • When using cross_bucket = False, the resource endpoint is used. There was an issue with the retrieval of the metadata with annotate = True. It has since then been fixed Annotated source does not behave as expected nexus#4737
  • When using cross_bucket = True, the resolvers endpoint is used. It was not possible to get metadata along with the data if retrieve_source = True, now it is, with the annotate=True query parameter. Add the option to get the annotated original payload when fetching a resource via resolvers nexus#4717

• Remove async force complete for retrieval of one resource. Synchronous requests are now used, since the body of the function is simpler, a little bit of duplication doesn't hurt as much.

• To avoid the client session staying open too long, leading to a ServerDisconnectedError. + all tasks/resources no longer share the same client session (a batch of tasks/resources per client session)

• When using the resource endpoint, the annotate query parameter only works for regular resources. It fails if we attempt to retrieve a schema with it. The best thing would be to use the schema endpoint with the annotate query parameter but that isn't supported by delta yet. So the mechanism to retrieve data (not source) + metadata, then data (source) without metadata, and then to merge them has been kept in a retrieve_schema method.

• Allow to return annotated payload for schemas nexus#5079 RDFModel's validate uses the resource retrieve implementation to get schemas and ontologies. It was not possible to use the annotate query parameter in order to retrieve schemas through the resource endpoint until this change in the API.

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 9.90991% with 100 lines in your changes missing coverage. Please review.

Project coverage is 74.01%. Comparing base (4a0ee65) to head (da98e7d).

Files with missing lines	Patch %	Lines
kgforge/specializations/stores/bluebrain_nexus.py	6.15%	61 Missing ⚠️
...cializations/stores/nexus/batch_request_handler.py	13.33%	39 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #420      +/-   ##
==========================================
- Coverage   74.19%   74.01%   -0.18%     
==========================================
  Files         106      106              
  Lines        6998     7020      +22     
==========================================
+ Hits         5192     5196       +4     
- Misses       1806     1824      +18     

Flag	Coverage Δ
unittests	74.01% <9.90%> (-0.18%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[crisely09]

When using the resource endpoint, the annotate query parameter only works for regular resources. It fails if we attempt to retrieve a schema with it. The best thing would be to use the schema endpoint with the annotate query parameter but that isn't supported by delta yet. So the mechanism to retrieve data (not source) + metadata, then data (source) without metadata, and then to merge them has been kept in a retrieve_schema method.

BlueBrain/nexus#5079 RDFModel's validate uses the resource retrieve implementation to get schemas and ontologies. It was not possible to use the annotate query parameter in order to retrieve schemas through the resource endpoint until this change in the API.

Aren't these two points contradicting each other? if now the annotate query parameter is available for schemas, then the retrieve_schema could use the simplified retrieval?