Closed jonrkarr closed 3 years ago
The error appears to be with the taxonomic distance calculation because the query succeeds without the species. Example: https://testapi.datanator.info/proteins/proximity_abundance/proximity_abundance_kegg/?kegg_id=997616at2&distance=40
Cases where two different species have the same canonical common ancestors were not properly accounted for. It has now been fixed. Performance should also be greatly improved now the endpoint leverage native canonical ancestor information.
Note: having the keyword fixed
in commit message in front of the issue link seems to automatically close the referenced issue.
This looks to be fixed in the API, but the test API still has the error.
I learned about these shortcuts from Bilal. He's been using them. This works in comments on pull requests too. GitHub recognizes several keywords. See https://docs.github.com/en/free-pro-team@latest/github/managing-your-work-on-github/linking-a-pull-request-to-an-issue#linking-a-pull-request-to-an-issue-using-a-keyword.
This looks to be fixed in the API, but the test API still has the error.
I didn't deploy the test API. Should be fixed once the deployment is done. I also found out a big part of the performance issue for this endpoint is due to MongoDB's handling of sparse compound indexes. I've trying different combinations to see if performance can be improved.
Improvement to the performance would be great
Original issue is working now. Moved performance improvement issue to other repo.
Examples: