v1.3.4
HKUDS/LightRAGv1.3.4Apr 26, 2025by LarFii
AI Summary
WebUI enhancements including Markdown/Mermaid support and PostgreSQL query performance improvements.
Key Highlights
- Added Markdown & Mermaid support for WebUI
- Added custom tokenizer support
- Added query mode prefix support in retrieval test
- Improved PostgreSQL AGE query speed with customized BFS implementation
- Improved parallel handling for documents processing
- Fixed dead lock problem for FAISS vector storage
New Features
- Markdown & Mermaid support
- Custom tokenizer
- Query mode prefix support
- PostgreSQL query optimization
- Parallel document processing
- FAISS deadlock fix
Full Release Notes
## What's Changed * Feat: Add Markdown & Mermaid support for WebUI by @choizhang in https://github.com/HKUDS/LightRAG/pull/1426 * Feat: add custom tokenizer by @drahnreb in https://github.com/HKUDS/LightRAG/pull/1396 * Feat: Add query mode prefix support in retrieval test of WebUI by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/1433 * Improve PostgreSQL AGE query speed by write a customized BFS implementation by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/1446 * Improve parallel handling for documents processing by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/1447 * Postpond mermaid rendering untill streaming streaming respond stopped. by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/1453 * Hot Fix: solved dead lock problem for FAISS vector storage by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/1465 * Prioritize high-degree neighbors in BFS traversal for NetoworkX storage by @danielaskdd in https://github.com/HKUDS/LightRAG/pull/1468 * fix: correct parentheses. `system_prompt` was never formatted. by @drahnreb in https://github.com/HKUDS/LightRAG/pull/1402 * fix: add miss key for ainsert_custom_chunks by @zhudongwork in https://github.com/HKUDS/LightRAG/pull/1389 * feat: PGKVStorage add get_all() by @zhudongwork in https://github.com/HKUDS/LightRAG/pull/1392 * fix streaming "list index out of range" by @Enoughappens in https://github.com/HKUDS/LightRAG/pull/1409 * refactor(lightrag): Optimize error handling and logging by @stevezhangishero in https://github.com/HKUDS/LightRAG/pull/1393 * Update prompt.py by @cywzl in https://github.com/HKUDS/LightRAG/pull/1422 * feat: Add `aget_docs_by_ids` to check document processing status by @earayu in https://github.com/HKUDS/LightRAG/pull/1419 * use a single parameter placeholder for the array; by @OxidBurn in https://github.com/HKUDS/LightRAG/pull/1427 * Revise the context format of chunks from CSV to JSON to enhance compatibility with LLM by @Inighty in https://github.com/HKUDS/LightRAG/pull/1412 * Update postgres_impl.py by @widgit in https://github.com/HKUDS/LightRAG/pull/1431 * Friendly implementation of entity extraction and relationship weight extract for Low-Capability LLMs by @tackhwa in https://github.com/HKUDS/LightRAG/pull/1423 * Update README-zh.md by @feeingHe in https://github.com/HKUDS/LightRAG/pull/1466 ## New Contributors * @drahnreb made their first contribution in https://github.com/HKUDS/LightRAG/pull/1402 * @Enoughappens made their first contribution in https://github.com/HKUDS/LightRAG/pull/1409 * @stevezhangishero made their first contribution in https://github.com/HKUDS/LightRAG/pull/1393 * @cywzl made their first contribution in https://github.com/HKUDS/LightRAG/pull/1422 * @Inighty made their first contribution in https://github.com/HKUDS/LightRAG/pull/1412 * @widgit made their first contribution in https://github.com/HKUDS/LightRAG/pull/1431 * @feeingHe made their first contribution in https://github.com/HKUDS/LightRAG/pull/1466 **Full Changelog**: https://github.com/HKUDS/LightRAG/compare/v1.3.3...v1.3.4