feat: Add Pinecone online store integration - #6585
Conversation
804ff38 to
020a0e6
Compare
a9607d7 to
0fa28bc
Compare
|
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #6585 +/- ##
==========================================
+ Coverage 47.09% 47.41% +0.32%
==========================================
Files 419 422 +3
Lines 51878 52272 +394
Branches 7525 7613 +88
==========================================
+ Hits 24430 24786 +356
- Misses 25700 25716 +16
- Partials 1748 1770 +22
... and 1 file with indirect coverage changes Continue to review full report in Codecov by Harness.
🚀 New features to boost your workflow:
|
3d25634 to
5db79c2
Compare
Signed-off-by: Siddhesh Khairnar <khairnarsiddhesh4057@gmail.com>
5db79c2 to
134f76d
Compare
Signed-off-by: Siddhesh Khairnar <khairnarsiddhesh4057@gmail.com>
Signed-off-by: Siddhesh Khairnar <khairnarsiddhesh4057@gmail.com>
Signed-off-by: Siddhesh Khairnar <khairnarsiddhesh4057@gmail.com>
jyejare
left a comment
There was a problem hiding this comment.
This PR adds a comprehensive Pinecone online store integration for Feast, including vector similarity search capabilities. The implementation is well-structured with proper error handling, comprehensive tests, and good documentation. However, there are several areas that need improvement including error handling gaps, potential security concerns, and some code quality issues that should be addressed.
Signed-off-by: Siddhesh Khairnar <khairnarsiddhesh4057@gmail.com>
Signed-off-by: Siddhesh Khairnar <khairnarsiddhesh4057@gmail.com>
Signed-off-by: Siddhesh Khairnar <khairnarsiddhesh4057@gmail.com>
Signed-off-by: Siddhesh Khairnar <khairnarsiddhesh4057@gmail.com>
|
Thanks for the thorough review, @jyejare. I’ve addressed all of the comments in the latest commits (softened the API-key error, upsert/query error handling, metadata size checks, returning match vector values, |
What this PR does / why we need it:
This PR adds Pinecone as a new online store integration for Feast. Pinecone is a managed vector database that enables fast similarity search and retrieval, making it ideal for ML feature serving workloads that involve embeddings and vector-based lookups.
Key changes:
PineconeOnlineStoreclass with full CRUD operations (online_write_batch,online_read,teardown)PineconeOnlineStoreConfigWhich issue(s) this PR fixes:
Fixes #6582
Checks
git commit -s)Testing Strategy
Misc
pinecone-clientas an optional dependency (pip install feast[pinecone])docs/reference/online-stores/pinecone.md