buua436
2476e2fbfa
feat: configure knowledge compilation runtime settings ( #19254 )
2026-09-04 17:05:22 +08:00
chanx
14306f42bd
feat(preview): add EPUB document preview with encoding normalization ( #19242 )
2026-09-04 16:56:21 +08:00
balibabu
dbeaab1648
Fix: After clearing data from pages other than the first page, the data in the filters and search box should also be cleared. ( #19255 )
2026-09-04 16:54:41 +08:00
Lynn
2177c3ad5e
Fix: check required fields when use template ( #19228 )
2026-09-04 16:52:50 +08:00
chanx
2ed042b17c
fix(web): sanitize worksheet names that crash the excel previewer ( #19260 )
2026-09-04 16:51:03 +08:00
euvre
9d8df641dd
fix(ingestion): show data pipeline on failed parse log entries ( #19231 )
2026-09-04 16:46:10 +08:00
Sbaaoui Idriss
c280ff0bec
fix(ingestion-logs): normalize dataset status filters ( #19261 )
2026-09-04 16:43:29 +08:00
Lynn
2f8dab2199
Fix: upload image to model in chat ( #19257 )
2026-09-04 16:37:25 +08:00
MarMar Labs
2f13f103a0
fix(token_utils): build the tiktoken encoder on first use, not at import ( #18986 )
2026-09-04 15:46:30 +08:00
Sbaaoui Idriss
ebb2174236
fix(parser): decode non-UTF-8 document text ( #19253 )
2026-09-04 15:40:34 +08:00
Harsh Kashyap
eea0f1a296
test(web): cover decodeBlobText across all five decoder branches ( #19240 )
...
Add 8 unit tests for decodeBlobText covering UTF-16LE/UTF-16BE/UTF-8 BOM, strict UTF-8, GBK fallback, and buffer edge cases. No production code change. Fixes #19239 .
2026-09-04 15:36:47 +08:00
xugangqiang
4298700cb2
chore(deps): depend on org mirror infiniflow/onnxruntime_go ( #19169 )
2026-09-04 15:26:46 +08:00
xugangqiang
4f7c69c06c
fix(chunker): align canvas TokenChunker default delimiters with General parser ( #19226 )
2026-09-04 15:22:24 +08:00
Wang Qi
b48cfc131e
Fix missing content_ltks when rerank ( #19230 )
2026-09-04 14:48:20 +08:00
jay77721
5bcaa73cb7
fix(parser): support non-UTF-8 decoding across text, csv, and epub parsers ( #19243 )
2026-09-04 14:23:08 +08:00
maoyifeng
5d0e2253cf
fix ci unit test not found in installed package 'wordnet' ( #19251 )
2026-09-04 13:52:43 +08:00
lei_lei
62f09abe4e
fix(ingestion): route xlsx QA JSON table items to extractQATable ( #19198 )
2026-09-04 11:23:33 +08:00
buua436
447f87d79f
feat: adapt knowledge compilation concurrency ( #19156 )
2026-09-04 11:17:56 +08:00
euvre
60fcf20c89
fix(agent): wire document service so dataflow rerun works instead of 500ing ( #19094 )
2026-09-04 10:02:08 +08:00
dawn
70601910ef
fix(model_meta): discover TEI reranker models ( #19210 )
2026-09-03 20:18:27 +08:00
Wang Qi
3c2c3e1743
Docs: update indexer ( #19227 )
2026-09-03 19:39:06 +08:00
euvre
a41d0d3a94
fix(agent): accept user_id node param on the Retrieval agent tool ( #19090 )
2026-09-03 19:37:10 +08:00
jay77721 and CommandCodeBot
45c3a70c2f
refactor(go): merge messagefit into internal/tokenizer ( #19225 )
...
Co-authored-by: CommandCodeBot <noreply@commandcode.ai >
2026-09-03 19:33:29 +08:00
xugangqiang
20b6075d92
refactor deepdoc backend register ( #19224 )
...
Refactor deep doc registration
2026-09-03 19:32:02 +08:00
buua436
ef8b424888
fix: normalize reasoning stream markers ( #19216 )
2026-09-03 19:06:53 +08:00
chanx
020e079780
fix(preview): handle non-UTF-8 encodings in CSV/TXT preview ( #19222 )
2026-09-03 19:05:42 +08:00
balibabu
9e294429ad
Fix: The filter data did not update after deleting the agent's item. ( #19220 )
2026-09-03 18:14:03 +08:00
Eric Curtin and Jin Hai
c051f40593
feat: add llmman as a model provider ( #19030 )
...
Add llmman support (local OCI-distributed models) — Python + Go chat/embeddings/list. Uses OpenAI API shape.
Signed-off-by: Eric Curtin <eric.curtin@docker.com >
Co-authored-by: Jin Hai <haijin.chn@gmail.com >
2026-09-03 18:13:37 +08:00
Sbaaoui Idriss
ec0cf792a9
fix(storage): scope configured single-bucket S3 removals correctly ( #19217 )
2026-09-03 17:59:06 +08:00
chanx
f9c3260870
fix(dynamic-form): allow non-string values in Segmented fields ( #15106 ) ( #19212 )
2026-09-03 17:56:16 +08:00
Wang Qi
ae556538ac
[Go] Fix switch embeding report error ( #19201 )
2026-09-03 17:55:58 +08:00
Sbaaoui Idriss
cbaf0669b8
fix(dataset): fail closed on unknown permissions ( #19213 )
2026-09-03 17:30:10 +08:00
euvre
4bb522d11d
fix(rag): unify parser_config delimiter defaults on a single constant ( #19058 )
2026-09-03 17:16:37 +08:00
Sbaaoui Idriss
9f73625f53
fix(storage): handle versioned S3 bucket deletion ( #19199 )
2026-09-03 16:45:20 +08:00
xugangqiang
a5a83a5653
refactor(parser): collapse docElementText into docxElementText ( #19135 )
2026-09-03 16:42:49 +08:00
balibabu
68959037f3
Fix: Install only oxlint and oxfmt to ensure CI checks for web files. ( #19197 )
2026-09-03 16:33:42 +08:00
balibabu
c28337b78e
Fix: Hide Skills Entry ( #19205 )
2026-09-03 16:32:54 +08:00
Wang Qi
4ab91c703c
Dev: launch_backend_service.sh to support --debug ( #19215 )
2026-09-03 16:27:48 +08:00
Caobaining1
d0e22f4dfc
Fix/final answer streaming phase is not interrupted by the research budget ( #19209 )
...
- Remove the whole-graph timeout around graph.ainvoke() so the final
answer streaming phase is not interrupted by the research budget.
- Keep research-stage safeguards unchanged, including per-node timeouts,
the 180-second research deadline, routing guards, and recursion limits.
The final answer generation can now continue until the model finishes.
Retrieval remains bounded by the existing per-node timeouts, research
deadline, routing guards, and recursion limit.
2026-09-03 16:04:52 +08:00
Wang Qi
2390b20dc8
Refactor: go default template no more _0 ( #19204 )
2026-09-03 15:36:59 +08:00
helloxjade
070fa6b15c
docs: add macOS vm.max_map_count setup guidance ( #19177 )
...
Fix self-hosting docs: clarify vm.max_map_count is Linux-only; add macOS Docker Desktop and Colima configuration instructions. Closes #19118 .
2026-09-03 15:23:42 +08:00
68cdb795f5
Fix OAuth redirect URI examples ( #15683 )
...
Fix OAuth callback URL examples in docs — now use /api/v1/auth/oauth/<channel>/callback (was stale /v1/user/oauth/callback).
Co-authored-by: 谢金城 <xiejincheng@xgd.com >
Co-authored-by: Jin Hai <haijin.chn@gmail.com >
2026-09-03 15:20:42 +08:00
Huang-404-Q
028d7fa45a
fix(storage): make bucket deletion work on the AWS_S3 backend ( #19162 )
...
Fix #19100 — S3 deletion now works for knowledge bases/users. Renamed rm_bucket → remove_bucket, rewrote with boto3 batch delete, single-bucket mode uses prefix-scoped deletion (keeps physical bucket).
2026-09-03 15:12:38 +08:00
helloxjade
6ed4970d6d
fix(pdf): make _table_transformer_job honor TABLE_AUTO_ROTATE by default ( #16079 )
2026-09-03 14:51:22 +08:00
jay77721
363f77ef80
refactor(ingestion): unify memory task execution with heartbeat leases ( #19189 )
2026-09-03 13:28:28 +08:00
Sbaaoui Idriss
330b988b61
fix(security): generate random UUIDv4 identifiers ( #19159 )
2026-09-03 13:13:57 +08:00
jay77721
471dd91e43
feat(ingestion): surface queued document task status ( #19137 )
2026-09-03 12:05:41 +08:00
mkaaad and coderabbitai[bot]
d29a4fc26b
fix(py-go-connector): enforce kb authorization in rebuild ( #19190 )
...
Based on community PR #15272 by @rene0422
---------
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
2026-09-03 12:01:28 +08:00
Yash Raj Pandey
b502ddf6ed
Require exactly two cells for a CSV QA pair ( #19008 )
2026-09-03 11:56:58 +08:00
dependabot[bot]
383a441dab
build(deps): bump mistune from 3.3.0 to 3.3.3 ( #19193 )
2026-09-03 11:25:52 +08:00
buua436
chanx
balibabu
Lynn
euvre
Sbaaoui Idriss
MarMar Labs
Harsh Kashyap
xugangqiang
Wang Qi
jay77721
maoyifeng
lei_lei
dawn
Eric CurtinandJin Hai
Caobaining1
helloxjade
Huang-404-Q
mkaaadandcoderabbitai[bot]
Yash Raj Pandey
dependabot[bot]