Optimize Variant Classification Functions to Avoid ORM Relationship Loading #657
Description
The variant classification functions in classification.py currently use ORM relationship membership checks (variant in functional_range.variants) to determine which functional classification range a variant belongs to. This approach has significant performance implications:
- Eager Loading Required: The score set API must eagerly load the variants relationship on all ScoreCalibrationFunctionalClassification objects to avoid N+1 queries
- Memory Overhead: Loading all variants for all ranges means loading potentially thousands of variant records into memory even when only checking a single variant
- O(n) Lookup: The membership check requires iterating through all variants in a range, resulting in O(n) time complexity
- API Performance Degradation: Score set API endpoints are slowed down due to the required relationship loading
It seems prudent to pre-resolve these variant classifications to reduce DB fetches and minimize the amount of variant data to hold in memory.