lianbieandzhoufazhao 686b51f0d7 fix(rag): keep embedding vectors aligned when one in a batch is dropped (#45569) ...

Co-authored-by: zhoufazhao <fazhao.zhou@wetrial.com>

2026-11-05 13:03:13 +00:00

..

cleaner

fix(api): keep email-like text in Markdown link URLs when removing urls and emails (#43712)

2026-10-09 02:22:45 +00:00

data_post_processor

feat: knowledge add more trace (#38959)

2026-07-24 02:27:16 +00:00

datasource

refactor(api): type-check vector store provider unit tests (#44458)

2026-10-21 09:01:19 +00:00

docstore

refactor(api): establish dataset application boundaries (#41729)

2026-09-29 08:42:31 +00:00

embedding

fix(rag): keep embedding vectors aligned when one in a batch is dropped (#45569)

2026-11-05 13:03:13 +00:00

entities

fix(rag): reject negative segmentation overlap before indexing (#45455)

2026-11-03 11:41:04 +00:00

extractor

docs: fix the WaterCrawlWebExtractor Args block (#42304)

2026-11-04 16:50:39 +00:00

index_processor

fix: strip reasoning/thinking blocks from knowledge base summary output (#38623)

2026-11-04 12:34:26 +00:00

models

chore: reorg imports (#35308)

2026-04-16 08:50:02 +00:00

pipeline

chore: use from __future__ import annotations (#30254)

2026-01-06 23:57:20 +09:00

rerank

perf(rag): reduce knowledge retrieval post-processing overhead (#43665)

2026-10-15 02:47:15 +00:00

retrieval

feat: retrieve external_documents add doc_metadata (#22496)

2026-11-04 08:14:02 +00:00

splitter

fix: non-ASCII separators broken in FixedRecursiveCharacterTextSplitter (#33119)

2026-11-04 06:44:30 +00:00

summary_index

ci: enable Ruff E712 (#44114)

2026-10-15 07:37:57 +00:00

__init__.py

Feat/dify rag (#2528)

2024-02-22 23:31:57 +08:00