elasticsearch-py
elasticsearch-py copied to clipboard
elastic
Official Python client for Elasticsearch
I encountered an issue with pickling/unpickling BulkindexError. `unpickle_exception\n inst = func(*args)\n ^^^^^^^^^^^\nTypeError: BulkIndexError.**init**() missing 1 required positional argument: 'errors` I'm adding an implementation of `__reduce__` to this (and ScanError) so...
It seems like BulkIndexError cannot be unpickled: `unpickle_exception\n inst = func(*args)\n ^^^^^^^^^^^\nTypeError: BulkIndexError.**init**() missing 1 required positional argument: 'errors` This creates problems with some libraries that expect errors to be...
This is mostly adapted from the Elasticsearch Serverless tests that already do this. @miguelgrinberg This also bumps the tested version of Elasticsearch to 8.16. How do you want to handle...
This change updates the `SparseVectorStrategy` class to use the newer `sparse_vector` query instead of `text_expansion`, which generates a deprecation warning since 8.15.
It would be nice to have the ability to pass filter clauses to [max_marginal_relevance_search](https://github.com/elastic/elasticsearch-py/blob/6521b55761ccb6e7033ea643037fd1c94ea818c4/elasticsearch/helpers/vectorstore/_async/vectorstore.py#L344-L357) The search method that is being called inside this method does accept filter clauses.
ES Cloud id does not work when using [GCP Private Service Connect traffic filters](https://www.elastic.co/docs/deploy-manage/security/gcp-private-service-connect-traffic-filters) I think https://github.com/elastic/elastic-transport-python/blob/main/elastic_transport/client_utils.py#L127 is the culprit but I'm not certain. We were able to make things...
