Closed ssssarah closed 7 months ago
Attention: 136 lines
in your changes are missing coverage. Please review.
Comparison is base (
6ddd5f3
) 74.69% compared to head (cb80eb9
) 74.41%.:exclamation: Current head cb80eb9 differs from pull request most recent head 20179b2. Consider uploading reports for the commit 20179b2 to get more accurate results
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
Generic catching of http errors on requests now also catches the errors thrown by
aiohttp
In the case of retrieval, there is more logic that goes into making requests (there may be a single request, just like there may be 3 of them)
Therefore, a different
task_creator
is used (can be found inBlueBrainNexus
)The entire logic of retrieval is located inside a co-routine (async-await block), see
_retrieve
. In the case of single resource retrieval, we force a blocking call. In the case of multiple resources retrieval, a differenttask_creator
is provided toBatchRequest.batch_request
, one that schedules tasks that are based on the asynchronous retrieval logic.TODO: figure out how to return errors of failed retrievals. Right now, it's through an empty resource and its _last_action says it was not successful. If None is returned instead, there should be a way to propagate the error to the user still (other than just printing)