What's Changed
- feat: optimize repeated entity extraction by @lxobr in #1682
- fix: Resolve issue with text classification by @dexters1 in #1704
- Update dev from main by @dexters1 in #1707
- fix: update unsupported vector db log by @hande-k in #1708
- Fix ollama tests by @pazone in #1714
- CI: (dev)Extract Windows and MacOS tests to separate job by @pazone in #1715
- docs: Fix Readme Instructions by @davidmyriel in #1706
- CI: Fix ollama tests by @pazone in #1713
- fix: Add logs to docker by @Vasilije1990 in #1656
- CI: Limit deletion integration tests to 60 minutes by @pazone in #1730
- feat: Add initial custom pipeline by @dexters1 in #1716
- feat: introduces memify pipeline to save cache sessions into cognee by @hajdul88 in #1731
- refactor: Enable multi user mode by default if graph and vector db pr… by @dexters1 in #1695
- test: add load tests by @siillee in #1573
- Fix: fixes session history test for multiuser mode by @hajdul88 in #1746
- chore: update videos by @hande-k in #1748
- feature: text chunker with overlap by @lxobr in #1732
- feat: add structured output support to retrievers by @siillee in #1734
- CI: Initial Release test workflow by @pazone in #1749
- Fix cypher search by @dexters1 in #1739
- Cherry-pick: Fix cypher search (#1739) by @pazone in #1763
- fix: added release update version by @Vasilije1990 in #1764
- feat:add emptiness check to neptune adapter by @siillee in #1757
- test: fix weighted edges example by @siillee in #1745
- feat: Add multi-tenancy by @dexters1 in #1560
- feat: enable multi user for falkor by @siillee in #1689
- refactor: Disable telemetry for all non telemetry tests by @dexters1 in #1777
- fix: Resolve issue with empty node set by @dexters1 in #1744
- feat: fs-cache by @daukadolt in #1645
- Chore: Acceptance Criteria for PRs by @pazone in #1781
- Relational DB migration test search by @dexters1 in #1752
- fix: remove duplicate mistral adapter creation by @siillee in #1787
- Fix: MCP remove cognee.add() preprequisite from the doc by @pazone in #1788
- Fix: Remove cognee script from pyproject.toml by @pazone in #1785
- Feature/cog 3407 fixing integration test in ci by @hajdul88 in #1810
- docs(mcp): update available tools documentation by @armelhbobdad in #1740
- Handle multiple response formats in OllamaEmbeddingEngine by @weikao in #1735
- chore: changes url for crawler tests by @hajdul88 in #1816
- CI: Release drafter configuration by @pazone in #1817
- CI: mergify config by @pazone in #1818
- ci(Mergify): configuration update by @pazone in #1820
- feat: add instructor mode env variable and config parameter by @siillee in #1789
- feat: csv ingestion & chunking by @EricXiao95 in #1574
- feat: add ontology endpoint in REST API by @fahad-shoaib in #1724
- feat: make notebook cognee usage identical across uis by @borisarzentar in #1747
- version: 0.5.0.dev0 by @borisarzentar in #1827
- fix: fixes ontology api endpoint tests by @hajdul88 in #1824
- feature: Introduces wide subgraph search in graph completion and improves QA speed by @hajdul88 in #1736
- backport: Adds lance-namespace version fix to toml (fixes lancedb issue with 0.2.0 lance-namespace version) + crawler ingetration test url fix by @hajdul88 in #1842
- COG-3437: Chore: CodeRabbit config by @pazone in #1833
- CI: 32 GB machine for Ollama tests by @pazone in #1857
- Fix distributed issues with latest pydantic version by @dexters1 in #1859
- feature: adds triplet embedding via memify by @hajdul88 in #1832
- chore: uv lock check in pre-test workflow by @pazone in #1773
- Main merge vol4 by @dexters1 in #1856
- refactor: move codify pipeline out of main repo by @siillee in #1738
- fix: fixes lancedb batch handling by @hajdul88 in #1872
- fix: Resolve issue with BAML rate limit handling by @dexters1 in #1813
- fix: install nvm and node for -ui cli command by @borisarzentar in #1836
- fix: remove double quotes from llmconfig str params by @daukadolt in #1758
- Removed check_permissions_on_dataset.py and related references by @martin0731 in #1786
- feat: avoid double edge vector search in triplet search by @lxobr in #1877
- feat: Adds edge centered payload and embedding structure during ingestion by @hajdul88 in #1853
- Chore: Remove Ontology file size limit. Code duplications by @pazone in #1880
- feat: add dataset database handler logic and neo4j/lancedb/kuzu handlers by @dexters1 in #1776
- Merge main vol7 by @dexters1 in #1891
- feat: make pipeline processing cache optional by @dexters1 in #1876
- feat: Add dataset database handler info by @dexters1 in #1887
- fix: make ontology key an optional param in cognify by @dexters1 in #1894
- test: remove anything codify related from mcp test by @siillee in #1890
- feat: add bedrock as supported llm provider by @siillee in #1830
- chore: adds id generation to memify triplet embedding pipeline by @hajdul88 in #1895
- feat: add support to pass custom parameters in llm adapters during cognify by @siillee in #1802
- COG-3546: Initial release pipeline by @pazone in #1883
- feat: Add database deletion on dataset delete by @dexters1 in #1893
- chore: introduces 1 file upload in ontology endpoint by @hajdul88 in #1899
- Release 0.5.0 Merge dev to main by @pazone in #1900
- Bump cognee and MCP versions to 0.5.0 by @pazone in #1901
- Release 0.5.0: uv lock by @pazone in #1902
New Contributors
- @davidmyriel made their first contribution in #1706
- @armelhbobdad made their first contribution in #1740
- @fahad-shoaib made their first contribution in #1724
- @martin0731 made their first contribution in #1786
Full Changelog: v0.3.8...v0.5.0