Skip to content

Avoid repeated name normalization during matching #460

Description

@GibranAlfa

Matching about 1,550 RMS wellbore names against SMDA wellbore names took about 159 seconds because MatchService normalizes both names and reapplies replacement rules for every source-target pair. Normalize each target once per request and each source once before calculating similarity scores, then reuse the normalized values while preserving the existing top-three results and ordering. This reduced the same comparison to about 515 ms locally.

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions