Commit graph

40 commits

Author SHA1 Message Date
Yiorgis Gozadinos
9a480cee90
Return program in chat agent analyze; 2026-02-06 12:07:18 +01:00
Yiorgis Gozadinos
ed570633cd
Basic integration of RLM with chat agent 2026-02-06 12:07:18 +01:00
Yiorgis Gozadinos
198a8d5a88
Add pre-loaded documents support and improve RLM test coverage 2026-02-06 12:07:04 +01:00
Yiorgis Gozadinos
5e5fc4a7d9
Integration tests for RLM 2026-02-06 12:07:03 +01:00
Yiorgis Gozadinos
57d5b1bde5
Simplify research planning: remove gather_context, always use planner 2026-01-30 22:32:08 +02:00
Yiorgis Gozadinos
109f770a2a
Refactor research graph to iterative planning approach 2026-01-30 20:46:29 +02:00
Yiorgis Gozadinos
57fd08d600
Revert "Synthesis steps now select only relevant citations instead of including all" 2026-01-28 16:00:27 +02:00
Yiorgis Gozadinos
1bc2b5f05a
Add vcr test 2026-01-27 17:28:50 +02:00
Yiorgis Gozadinos
cca03dbe61
Give the conversational agent a list documents and a summarize tools 2026-01-27 16:39:48 +02:00
Yiorgis Gozadinos
8a587ec554
Synthesis steps now select only relevant citations instead of including all 2026-01-27 15:56:29 +02:00
Yiorgis Gozadinos
6f9bc369dc
Improve coverage 2026-01-26 16:43:12 +02:00
Yiorgis Gozadinos
5976cebe8c
Concolidate tests 2026-01-26 16:43:12 +02:00
Yiorgis Gozadinos
a7d7a09c22
Integrate recall functionality into ask tool 2026-01-26 16:43:11 +02:00
Yiorgis Gozadinos
77647b08bc
Recall tool in chat agent 2026-01-26 16:42:41 +02:00
Yiorgis Gozadinos
30e9ff3038
Use session document filter in agent, combine with document name filter if necessary 2026-01-26 16:42:40 +02:00
Yiorgis Gozadinos
93737852a8
Use previous context in summarization 2026-01-26 16:42:38 +02:00
Yiorgis Gozadinos
74eac57ffa
Remove qa_history ranking in favor of SessionContext, planner can skip searching if context adequate 2026-01-26 16:42:38 +02:00
Yiorgis Gozadinos
8441516a73
Tests, widget to show context in tui 2026-01-26 16:41:13 +02:00
Yiorgis Gozadinos
aaddf4a5a9
Dynamic session context to be used in the conversational agent 2026-01-26 16:41:12 +02:00
Yiorgis Gozadinos
b96e5a3be1
Additional prompts needing adapting. Missing cassettes. 2026-01-21 15:17:56 +02:00
Yiorgis Gozadinos
7f51c63ad7
Search results now show rank position instead of raw scores. Consolidate cassettes 2026-01-21 14:54:15 +02:00
Yiorgis Gozadinos
7ff014684a
Support for jina reranker, both local and API 2026-01-21 10:39:21 +02:00
Yiorgis Gozadinos
62d35d2907
Add ocr_engine field to ConversionOptions 2026-01-20 13:41:43 +02:00
Yiorgis Gozadinos
8bdfbe526c
Fix embed-only rebuild with changed vector dimensions
When a database was created with one embedding model and rebuild
--embed-only was run with a different model, it failed with a
vector dimension validation error.

- Store reads stored vector_dim when opening existing databases
- _rebuild_embed_only recreates chunks table to handle dimension changes
- Add test for rebuild with changed vector dimensions
2026-01-14 17:42:28 +02:00
Yiorgis Gozadinos
bff613b7dc
Better test coverage 2026-01-14 17:12:33 +02:00
Yiorgis Gozadinos
929c29250e
Additional tests for page & picture images for local and docling-serve 2026-01-14 12:48:13 +02:00
Yiorgis Gozadinos
98b3d114b3
Update to docling-document 1.9.0 2026-01-13 19:20:27 +02:00
Yiorgis Gozadinos
ce326d80ff
Disable ocr in tests previously marked as integration 2026-01-12 16:19:25 +02:00
Yiorgis Gozadinos
51836a1c55
Add tests for chat agent & friends 2026-01-12 16:00:07 +02:00
Yiorgis Gozadinos
e4a7d86348
Rank q/a history with respect to cosine similarity to new queries. Adapt the chat agent to only pass relevant q/as 2026-01-12 14:15:36 +02:00
Yiorgis Gozadinos
a9ed178a90
Research graph test cassette 2026-01-12 12:43:51 +02:00
Yiorgis Gozadinos
3603b00513
Fix base64 images leaking into expanded search results
PictureItem.export_to_markdown() defaults to ImageRefMode.EMBEDDED,
which embeds base64 image data. When _extract_item_text() called this
method during context expansion, base64 data would leak into results.

Now explicitly handles PictureItem with ImageRefMode.PLACEHOLDER to
prevent base64 while still including VLM descriptions and captions.
2026-01-08 18:42:58 +02:00
Yiorgis Gozadinos
f861664656
Extract shared DoclingServeClient for async workflow & use it for VLM support, chunking, converting 2026-01-07 11:09:23 +02:00
Yiorgis Gozadinos
3b7fd440cc
End-to-end test for VLM annotation 2026-01-07 11:09:23 +02:00
Yiorgis Gozadinos
22f22d2181
Add vcr to tests 2025-12-29 14:59:15 +02:00
Yiorgis Gozadinos
1609582c1f
Move cassettes to default location, add development.md to docs 2025-12-29 14:13:25 +02:00
Yiorgis Gozadinos
1e9a10a235
Record cassettes for embedder, ignore huggingface while recording 2025-12-29 13:23:53 +02:00
Yiorgis Gozadinos
0da8ba97c4
Docling-serve cassettes 2025-12-26 18:59:46 +02:00
Yiorgis Gozadinos
c3a3b4a5b4
Extended VCR cassette recording to embedder, client, research graph, and search filter tests 2025-12-26 18:53:17 +02:00
Yiorgis Gozadinos
4a82e47d03
Use cassette to record and reply llms in qa tests 2025-12-26 15:41:11 +02:00