What's Changed

• feat: improve late chunking and optimize pgvector settings by @lsorber in #51
  • Add a workaround for #24 to increase the embedder's context size from 512 to a user-definable size.
  • Increase the default embedder context size to 1024 tokens (more degrades bge-m3's performance).
  • Upgrade llama-cpp-python to the latest version.
  • More robust testing of rerankers with Kendall's rank correlation coefficient.
  • Optimise pgvector's settings.
  • Offer better control of oversampling in hybrid and vector search.
  • Upgrade to the PostgreSQL 17.

Full Changelog: v0.3.0...v0.4.0

What's Changed

• feat: support prompt caching and apply Anthropic's long-context prompt format by @undo76 in #52

Full Changelog: v0.2.1...v0.3.0

What's Changed

• fix: add fallbacks for model info by @undo76 in #44
• fix: improve unpacking of keyword search results by @lsorber in #46
• fix: upgrade rerankers and remove flashrank patch by @lsorber in #47
• fix: improve structured output extraction and query adapter updates by @emilradix in #34

New Contributors

• @undo76 made their first contribution in #44
• @emilradix made their first contribution in #34

Full Changelog: v0.2.0...v0.2.1

What's Changed

• feat: add Chainlit frontend by @lsorber in #33

Full Changelog: v0.1.4...v0.2.0

What's Changed

• fix: fix optimal chunking edge cases by @lsorber in #32

Full Changelog: v0.1.3...v0.1.4

What's Changed

• fix: improve chunk and segment ordering by @lsorber in #29
• fix: upgrade pdftext by @lsorber in #30
• ci: expand test matrix with Python 3.11 by @lsorber in #31

Full Changelog: v0.1.2...v0.1.3

What's Changed

• fix: avoid pdftext v0.3.11 by @lsorber in #27

Full Changelog: v0.1.1...v0.1.2

What's Changed

• docs: improve installation instructions by @lsorber in #21
• fix: patch rerankers flashrank issue by @lsorber in #22

Full Changelog: v0.1.0...v0.1.1

What's Changed

• feat: implement basic features by @lsorber in #2
• feat: simplify document insertion by @lsorber in #6
• feat: optimize config for CPU and GPU by @lsorber in #7
• fix: improve indexing of multiple documents by @lsorber in #8
• feat: improve exception feedback for extraction by @lsorber in #9
• feat: automatically adjust number of RAG contexts by @lsorber in #10
• fix: lazily import optional dependencies by @lsorber in #11
• feat: infer missing font sizes by @lsorber in #12
• feat: add evaluation by @lsorber in #14
• feat: upgrade default CPU model to Phi-3.5-mini by @lsorber in #15
• feat: make query adapter minimally invasive by @lsorber in #16
• feat: add PostgreSQL support by @lsorber in #18
• feat: add LiteLLM and late chunking by @lsorber in #19
• feat: add reranking by @lsorber in #20

Full Changelog: https://github.com/superlinear-ai/raglite/commits/v0.1.0