Tags: ExpediaGroup/feast
Tags
fix: Cassandra Entity Key Serialization + Warning (#370) * fix: use config-driven entity key serialization version in Cassandra and Valkey stores Cassandra was hardcoding serialization version 2, which silently diverged from the Python default (v3) and from the Valkey store. Both stores now resolve the version from feature_store.yaml (entity_key_serialization_version, defaulting to v3 when unset), matching Python's SerializeEntityKey behavior. Also adds a shared warnPotentialEntityKeyVersionMismatch helper (extracted into entitykeyserialization.go) that fires a single deduplicated warning when an online read returns zero data for all requested feature views — a possible indicator of a write/read version mismatch. Covered by new unit tests in both store test files. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: use explicit time range in integration test materialization The test data parquet files have timestamps from April 2025. When using materialize-incremental with TTL=0 feature views, Feast defaults to a 1-year lookback window. Once the current date passed April 2026, the materialization found 0 rows to ingest, causing all Go integration tests to fail. Switch from materialize-incremental to materialize with an explicit start time of 2025-01-01 to ensure test data is always included. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: correct materialization time range for test data The test parquet file go/internal/test/feature_repo/driver_stats.parquet contains timestamps from 2021-2022, not 2025. The previous commit used 2025-01-01 as the start date which excluded all test data. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com> * fix: update operator test config to match upstream Sync operator test configuration with upstream feast-dev/feast: - Go version: 1.22.9 → 1.25.0 - ENVTEST_K8S_VERSION: 1.30.0 → 1.31.0 - ENVTEST_VERSION: release-0.18 → release-0.21 - Add GOTOOLCHAIN env var to test command - Add workflow concurrency to cancel in-progress runs The older envtest version was failing to download kubebuilder assets (etcd not found in $PATH) on newer Ubuntu CI runners. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com> * Revert "fix: update operator test config to match upstream" This reverts commit 38fc7ff. * fix: update operator tooling for Go 1.25 compatibility Update operator CI and tooling to work with Go 1.25: - Go version: 1.22.9 → 1.25.0 - controller-tools: v0.15.0 → v0.18.0 (fixes x/tools compile error) - golangci-lint: v1.59.1 → v2.12.2 (v1.x incompatible with Go 1.25) - envtest: release-0.18 → release-0.21 (fixes etcd not found) - Migrate .golangci.yml to v2 config format - Regenerate CRDs, RBAC, and docs with new tool versions Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com> * fix: use reflect.Pointer instead of deprecated reflect.Ptr The new golangci-lint v2 govet checker requires using reflect.Pointer (the newer name) instead of the deprecated reflect.Ptr constant. This aligns with the upstream feast-dev/feast repository. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com> --------- Co-authored-by: Zach Barnett <zbarnett@expediagroup.com> Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
feat: Add per-feature-view metrics for online read path (#365) * feat: Add per-feature-view metrics for online read path (latency, requests, errors) Emit per-feature-view latency, request count, error count in the Go feature server (HTTP + gRPC). This enables filtering latency/error distributions by feature view. Key changes: - Add Distribution() to StatsdClient interface - New FeatureViewReadMetrics emitter (fv_read_latency_ms, fv_read_requests, fv_read_errors) - Extend LookupMetricsAggregator with totalByFV for feature_lookup_requests - Extract FV names from request (works with fullFeatureNames=false) - New unified flag ENABLE_FV_LEVEL_METRICS - Instrument GetOnlineFeatures and GetOnlineFeaturesRange in both HTTP and gRPC handlers
feat: Add per-feature-view metrics for online read path (#365) * feat: Add per-feature-view metrics for online read path (latency, requests, errors) Emit per-feature-view latency, request count, error count in the Go feature server (HTTP + gRPC). This enables filtering latency/error distributions by feature view. Key changes: - Add Distribution() to StatsdClient interface - New FeatureViewReadMetrics emitter (fv_read_latency_ms, fv_read_requests, fv_read_errors) - Extend LookupMetricsAggregator with totalByFV for feature_lookup_requests - Extract FV names from request (works with fullFeatureNames=false) - New unified flag ENABLE_FV_LEVEL_METRICS - Instrument GetOnlineFeatures and GetOnlineFeaturesRange in both HTTP and gRPC handlers
fix: skip Cassandra rows whose TTL exceeds the 20-year maximum (#367) A SortedFeatureView row with a far-future event_timestamp produces a per-row TTL above Cassandra's 20-year cap (630,720,000s). Because the TTL is inlined into the CQL statement, session.prepare() rejects it ("ttl is too large" / "Unable to make int from ..."). That exception is raised during batch construction, bypasses the async on_failure handler, and aborts the entire micro-batch — stalling a stream indefinitely on a single bad record (the checkpoint can never advance past the poison row). Add a CASSANDRA_MAX_TTL guard mirroring the existing 'ttl < 0' skip: oversized-TTL rows are now skipped and logged at WARNING instead of crashing the batch, so good rows in the same batch keep flowing. The non-sorted insert path uses a static key_ttl_seconds and cannot overflow from row data, so it needs no guard. Co-authored-by: Manisha4 <Manisha4@github.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
feat: Add missing-key metrics for online feature lookups (#359) * feat: Add missing-key metrics for online feature lookups (Go + Python) Emit DogStatsD counters (feature_lookup_not_found, feature_lookup_null_or_expired) when online store reads return NOT_FOUND, NULL_VALUE, or OUTSIDE_MAX_AGE statuses. Gated behind ENABLE_MISSING_KEY_METRICS env var
feat: Add Vector Store Support in Valkey & Elastic Search (#364) * fix: Updated the elastic search to use vector_index defined in the feature view to identify vector fields (#348) * updated the elastic search to use vector_index defined in the feature view to identify vector fields * fix: formatting * Added logging and switched to use open source elastic search --------- Co-authored-by: vanitabhagwat <vbhagwat@expediagroup.com> * fix: ES integration tests (#350) * fix: ES integration tests * fix: Added fromisoformat() for converting timestamps --------- Co-authored-by: vanitabhagwat <vbhagwat@expediagroup.com> * fix:Elasticsearch online store — correctness, performance, and robustness fixes (#353) Co-authored-by: vanitabhagwat <vbhagwat@expediagroup.com> * Feature/vector store (#357) * feat: Valkey Online Write Batch Vector Search Support (#351) * Adding support for Valkey Search, adding changes to the online_write_batch functionality * Addressing PR comments * addressing linting error * fix tests * addressing PR comments * addressing PR comments * fixing linting --------- Co-authored-by: Manisha4 <Manisha4@github.com> * feat: Support Vector Search in Valkey (#354) * Adding support for Valkey Search, adding changes to the online_write_batch functionality * Addressing PR comments * addressing linting error * Adding changes to support search in valkey * fix tests * adding unit tests * reformatting files and adding checks and more tests * reformatting files and adding checks and more tests * reformatting files and adding checks and more tests * Fix linter errors: type annotations and code formatting - Add explicit type annotation for schema_fields to support both TagField and VectorField - Encode project string to bytes for consistency with other hash values - Decode doc_key bytes to string for hmget compatibility - Fix code formatting: break long lines and remove extra blank lines - Remove tests for multiple vector fields (Feast enforces one vector per feature view) - Fix config type: use 'eg-valkey' (hyphen) not 'eg_valkey' (underscore) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * addressing PR comments * addressing PR comments * fixing linting * Fix missing feature_name argument in retrieve_online_documents_v2 Add the third argument (vector_field.name) to _get_vector_index_name call to match the updated function signature. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com> * addressing comments, PR changes for some fixes and merge conflicts * fixing tests * fixing tests * fixing linting * fixing linting --------- Co-authored-by: Manisha4 <Manisha4@github.com> Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * fix: Valkey vector search - remove unsupported SORTBY (#356) * fix: Valkey vector search - remove unsupported SORTBY and fix tag filter syntax Valkey Search KNN queries return results pre-sorted by distance, so explicit SORTBY is not supported and causes a ResponseError. This removes the .sort_by() call from the query builder. Additionally, fixes the project tag filter to use unquoted syntax with backslash escaping for special characters (e.g. hyphens, dots) instead of the quoted syntax which was returning empty results. Updates unit tests to reflect both changes: replaces three metric-specific sort order tests with a single test asserting no SORTBY is set, and updates escaping assertions to match the new backslash-escape approach. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * style: apply ruff format to eg_valkey.py and test_valkey.py Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Manisha4 <Manisha4@github.com> Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Manisha4 <Manisha4@github.com> Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com> * feat: Quantization support for elastic search (#355) * feat: Add retrieve_online_documents_v3 SDK with multi-vector and hybrid fusion (#358) * feat: implement retrieve_online_documents_v3 SDK method - Multi-vector search with configurable fusion (RRF, WEIGHTED_LINEAR, VECTOR_ONLY) via the ES retriever API. Valkey gracefully degrades to single-vector KNN with warnings. - "embedding" magic key for V2→V3 migration convenience - Reserved output fields: final_score, signal_scores - include_signal_scores and distance_metric accepted as reserved params - ODFV and reserved-name collision validation - Shared signal_scores encoding via _signal_scores helper * update tests * update tests * fixing linting * docs: clarify final_score semantics in Valkey V3 docstring Correct the Valkey final_score description — Valkey's __distance__ is lower-is-better across all metrics (COSINE, L2, IP), not higher-is-better for IP. Call out the ordering inversion vs Elasticsearch so callers don't assume cross-backend score portability. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix: plumb include_signal_scores through V3 and align defaults to False Valkey/ES/provider/passthrough/online-store defaults were True, mismatching the SDK's False default. Align all layers on False and thread the parameter from retrieve_online_documents_v3 through the internal dispatcher, provider, and online stores so callers can opt in today and transparently pick up the explain-based per-signal path when it lands — no API change required. Tighten docstrings to describe the current best-effort behavior instead of hinting at latency tradeoffs that aren't wired yet. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * updating doc string * fix: preserve ranked row order in V3 retrieve_online_documents _retrieve_from_online_store_v3 was passing the driver's ranked rows through _get_unique_entities_from_values, which sorts and dedupes by entity-key bytes. That helper is correct for batch entity lookups but wrong here — ES/Valkey have already ordered rows by relevance, and the sort was scrambling them in the final DataFrame (e.g. doc_10 jumping ahead of doc_3 because "10" < "3" lexicographically). Replace the helper call with an identity mapping so the driver's rank order flows through untouched. No change to V1, V2, batch reads, or the helper itself; V3 output now matches the order returned by the online store. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> --------- Co-authored-by: Manisha4 <Manisha4@github.com> Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com> * fix: apply rescore_oversample to V3 ES kNN retrievers (#363) V3's retriever construction was silently ignoring rescore_oversample. V2 honored it (lines 483-486, 617-620) but V3 never added rescore_vector to its kNN clauses. On quantized indices (int8_hnsw / int4_hnsw / bbq_hnsw), this meant V3 queries returned lower recall than the config promised, with no error or warning. Wire rescore_oversample into each kNN retriever the same way V2 does. Covers single-vector and multi-vector V3 queries; BM25 retrievers skip the branch since they lack a "knn" key. Existing config validation (lines 102-127) already prevents rescore on non-quantized indices, so no new validation needed. Added three unit tests in TestRetrieveOnlineDocumentsV3QueryBuilding: - rescore_vector appears in single-vector query body when configured - rescore_vector appears on every kNN retriever in multi-vector query - rescore_vector absent when rescore_oversample is None Co-authored-by: Manisha4 <Manisha4@github.com> Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com> --------- Co-authored-by: vanitabhagwat <92561664+vanitabhagwat@users.noreply.github.com> Co-authored-by: vanitabhagwat <vbhagwat@expediagroup.com> Co-authored-by: Manisha4 <Manisha4@github.com> Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
fix: Add retry with exponential backoff for online store writes (stre… …aming) (#347) * fix(streaming): add retry with exponential backoff for online store writes * fix linting error * fix(build): pin setuptools_scm<10 to avoid vcs_versioning dependency * addressing PR comments, making timeout check more explicit and adding retry unit tests * fixing linting * fixing linting --------- Co-authored-by: Manisha4 <Manisha4@github.com>
fix: Wrap exception in cassandra on_failure for pickle issue (#345) * fix: Wrap exception in cassandra on_failure for pickle issue * fix: test issue * fix: update test to actually call cassandra code * fix: change test location * noop * revert last commit * ci: trigger workflow * ci: allow manual workflow dispatch for unit tests * remove manual workflow dispatch --------- Co-authored-by: Zach Barnett <zbarnett@expediagroup.com>
PreviousNext