Releases: nuclia/nucliadb
Releases · nuclia/nucliadb
Release 2.46.0
v2.46.0 chore: autopublish 2024-04-01T15:33:13Z
Release 2.45.0
v2.45.0 chore: autopublish 2024-03-26T23:59:35Z
Release 2.44.2
v2.44.2 chore: autopublish 2024-03-18T15:11:56Z
Release 2.44.1
What's Changed
- Better handle resumable uri not found errors on tus patch by @lferran in #1887
- Bump dependencies to fix vulnerabilities by @lferran in #1881
- Collect learning id on chat audit by @lferran in #1890
- Remove Unimplemented facets feature from find endpoints by @lferran in #1882
- merge on commit by @hermeGarcia in #1896
- Add shards and node info on catalog response by @lferran in #1895
- Place new shards on nodes based on disk space available by @lferran in #1891
- remove fst by @hermeGarcia in #1869
- Add generative_model in chat request by @lferran in #1866
- remove async merge by @hermeGarcia in #1898
- Handle shard not found errors while indexing by @lferran in #1900
- fix stack rebuild by @hermeGarcia in #1903
- Parallelize KB migrations by @lferran in #1905
- Handle more shard not found node errors by @lferran in #1902
- Whitelist proxied headers when proxing to learning services by @lferran in #1909
Full Changelog: v2.44.0...v2.44.1
Release 2.44.0
What's Changed
- Properly validate RAG strategies model by @lferran in #1847
- Fix Filter model validation by @lferran in #1853
- More logging info for migrator when waiting for node by @lferran in #1836
- reduce max max shard size by @vangheem in #1857
- Revert "Is CPAN really needed?" by @javitonino in #1859
- Remove unused gRPC methods from ingest servicer by @jotare in #1851
- fix replicate relations by @vangheem in #1858
- Retry tikv more by @lferran in #1861
- Add Merge gRPC method to index node writer by @jotare in #1860
- Bump replication id after forced merge by @javitonino in #1865
- Do not scan_keys on getting labelsets by @lferran in #1863
- Models management missing endpoint by @lferran in #1867
- Scan keys less often on ingest by @lferran in #1868
- fix node scheduling for rollover shards by @vangheem in #1873
- Proxy learning collector feedback by @lferran in #1874
- updated processing integration by @vangheem in #1872
- Add BucketNotEmpty as S3 string error codes by @jotare in #1878
- Purge orphan shards by @jotare in #1838
- Add resources for jaeger agent container by @lferran in #1880
- disk space computed and propagated by @hermeGarcia in #1870
- Do not rollover twice by @vangheem in #1889
Full Changelog: v2.43.1...v2.44.0
Release 2.43.1
What's Changed
- Delete nucliadb-sdk v1 by @lferran in #1827
- Remove too much context from unrelated entities log by @jotare in #1840
- Fix GCS delete bucket status code by @jotare in #1835
- Fix error code parsing for AWS S3 responses by @jotare in #1837
- Use new public Predict API by @lferran in #1841
- Run merging relations in executor by @lferran in #1844
- Unpin some more dependencies by @javitonino in #1843
- Support for more complex filtering expressions by @lferran in #1801
- Fix pypi release for sdk by @lferran in #1845
- Do not check proxy headers anymore by @jotare in #1846
- New KB creation logic by @lferran in #1842
- Model metadata is still received via grpc for hosted by @lferran in #1849
- Ingest - use defined SA name in purgeCron RBAC by @zepellin in #1850
Full Changelog: v2.43.0...v2.43.1
Release 2.43.0
What's Changed
- Add nuclia zone as context for feature flags by @jotare in #1807
- processing v2 cleanup by @vangheem in #1803
- Remove all kb configuration logic by @lferran in #1808
- Summary improvements by @lferran in #1809
- sc 8711/push to artifact registry too by @funollet in #1791
- Buffered read for vectors journal by @javitonino in #1811
- Add debug logs for notifications by @lferran in #1812
- Misc improvements by @jotare in #1802
- Fix missing await on GCS debug info by @jotare in #1813
- versioned internal predict api by @lferran in #1814
- lock deps for profile tools by @javitonino in #1817
- missing retries for tus gcs driver by @vangheem in #1818
- Proxy Learning Config by @lferran in #1810
- Do not use default full_jitter by @javitonino in #1821
- stop using ndb proxy, update urls by @vangheem in #1820
- Update python dependency versions by @javitonino in #1825
- fix delete from processing by @vangheem in #1826
- Fix purge command errors by @jotare in #1824
- Update rust version by @javitonino in #1822
- Fix some global state in our tests by @jotare in #1831
- Remove unused nucliadb_node settings by @jotare in #1828
Full Changelog: v2.42.1...v2.43.0
Release 2.42.1
v2.42.1 chore: autopublish 2024-02-07T11:00:42Z
Release 2.42.0
What's Changed
- safer tantivy replication by @hermeGarcia in #1755
- Offer different strategies for building RAG prompt context by @lferran in #1767
- Use read only transaction more on search by @lferran in #1726
- Reduce heavy search integration tests by @lferran in #1777
- Add prometheus metrics to resource cache by @javitonino in #1774
- Let the install script create the systemd unit by @javitonino in #1772
Full Changelog: v2.41.1...v2.42.0
Release 2.41.1
What's Changed
- Add summary_kind to /summarize endpoint by @jotare in #1764
- support path origin value by @vangheem in #1754
- compose title for processing status results by @vangheem in #1769
- New LRU cache implementation by @javitonino in #1738
Full Changelog: v2.41.0...v2.41.1