serratus-bio/serratus-summary-api

Duplication of matches between result pages

Closed this issue · 0 comments

victorlin commented

Theory

The current filtering by score results in many "ties" and order may vary for results with the same score:

serratus-summary-api/query/base.py

Line 18 in 159bf66

.order_by(table.score.desc())

This can probably be fixed with additional ordering on run_id to ensure a unique order.

Example

https://dev.serratus.io/explorer-rdrp?family=Bornaviridae&identity=50-87&score=72-100#result

1

API call: https://api.serratus.io/matches/rdrp/paged?page=1&perPage=10&scoreMin=72&scoreMax=100&identityMin=50&identityMax=87&family=Bornaviridae

2

API call: https://api.serratus.io/matches/rdrp/paged?page=2&perPage=10&scoreMin=72&scoreMax=100&identityMin=50&identityMax=87&family=Bornaviridae