This directory adds support running experiments against Pinecone - a managed vector database.
It supports connecting to both Pod-based and Serverless indexes.
To run VSB against a Pinecone index:
- Invoke VSB with
--database=pinecone
and provide your API key to VSB. A serverless index will be created for you, using aws/us-east-1.
vsb --database=pinecone --workload=mnist-test \
--pinecone_api_key=<YOUR_API_KEY>
If you wish to configure the created index or use an already existing one,
specify the name and/or spec with --pinecone_index_name
and --pinecone_index_spec
.
The --pinecone_index_spec
option takes a JSON string described in the Pinecone docs.
vsb --database=pinecone --workload=mnist-test \
--pinecone_api_key=<YOUR_API_KEY> \
--pinecone_index_name=<YOUR_INDEX_NAME> \
--pinecone_index_spec=<YOUR_INDEX_SPEC>
Tip
The API key and/or index name can also be passed via environment variables
(VSB__PINECONE_API_KEY
and VSB__PINECONE_INDEX_NAME
respectively).