FZJ-INM1-BDA / siibra-python

Software interfaces for interacting with brain atlases - Python client
Apache License 2.0
46 stars 8 forks source link

Use precalculated bounding boxes of image-features to speed up image feature query #583

Open AhmetNSimsek opened 3 months ago

AhmetNSimsek commented 3 months ago

The profiling revealed that most time is spend fetching the boundingboxes of images. When this information is cached, the query itself takes less than a second (on 11th Gen Intel(R) Core(TM) i5-1145G7, 16 GB memory). Therefore, this PR allows the bounding box information to be precalculated and stored in the configs to be used in run time. I suggest that we utilize this method for all image features until the new spatial query system is settled.

EDT: Tests are failing due to issues with data-proxy. (lint is failing due to unused import which is caused by using a specific config branch)

EDT 2: Requires: