jina-ai/examples

Fix batching logic for ZarrIndexer

JoanFM opened this issue · 1 comments

ZarrIndexer is broken since latest batching refactoring. batching is quite hard to mantain and to fix. I would just duplicate some code for the NumpyIndexer in the ZarrIndexer and rewrite the batching specifically for ZarrIndexer. If we start adding exceptions to batching it will explode

This does not belong here