Yiorgis Gozadinos
d1fdbdeb21
Refactor MCP to use proper models, test
2026-02-20 17:39:02 +02:00
Yiorgis Gozadinos
8c2a101a8a
Fix out-of-memory in list_documents by adding column projection
2026-02-06 14:11:36 +01:00
Yiorgis Gozadinos
4241a4b09e
Security fixes for RLM. Fix type() builtin, AST validation for dict key access, sql injection
2026-02-06 12:07:19 +01:00
Yiorgis Gozadinos
626f5cdd0e
Make list_documents return structured output and give feedback about total count and pages
2026-01-27 17:00:39 +02:00
Yiorgis Gozadinos
b4c158ffcf
Fix potential 0.25.0 migration crash
2026-01-22 17:29:26 +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
5e55c0df54
replace pyright with ty type checker
2026-01-19 15:50:14 +02:00
Yiorgis Gozadinos
5464d24483
require explicit migrate command for database migrations
2026-01-19 13:22:09 +02:00
Yiorgis Gozadinos
fae274a04d
Remove defensive checks in info()
2026-01-15 10:30:45 +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
061095e855
Remove obsolete migrations
2026-01-12 12:07:49 +02:00
Yiorgis Gozadinos
29c682d8eb
Add migration
2026-01-11 11:40:57 +02:00
Yiorgis Gozadinos
afc5b02285
Update schema to store docling_document as compressed bytes
...
- Change DocumentRecord.docling_document_json (str) to docling_document (bytes)
- Use large_binary Arrow type (64-bit offsets) to avoid 2GB column limit
- Decompress in Document.get_docling_document() using gzip
- Update DocumentRepository field mappings
2026-01-10 15:17:27 +02:00
Yiorgis Gozadinos
face6f95bc
compression utils
2026-01-10 15:11:58 +02:00
Yiorgis Gozadinos
c9cb005d07
Use FTS on contextualized text
2025-12-29 14:50:07 +02:00
Yiorgis Gozadinos
132b8a36bc
Differentiate between embedding a query and a document
2025-12-24 12:43:11 +02:00
Yiorgis Gozadinos
1e5eebfbf0
Checkout version logic and datetime utils
2025-12-19 12:03:16 +02:00
Yiorgis Gozadinos
1a8f72f99a
Add is_read_only property to Store. Guard all write operations in repositories
2025-12-18 15:55:24 +02:00
Yiorgis Gozadinos
5c4799164c
Escape when getting by URI
2025-12-12 10:56:37 +02:00
Yiorgis Gozadinos
448d00f9b2
Add pagination to chunk repo's get_by_document_id
2025-12-11 15:38:19 +02:00
Yiorgis Gozadinos
1b14a4643d
Remove unused bounding box calculations
2025-12-08 15:56:21 +02:00
Yiorgis Gozadinos
9a5f8e4368
Add type-aware context expansion for search results
2025-12-08 15:56:21 +02:00
Yiorgis Gozadinos
4f0f214f26
Remove unecessary tests, add no coverage pragmas where appropriate
2025-12-08 15:56:02 +02:00
Yiorgis Gozadinos
2572125804
Move embedding logic from ChunkRepository to client._ensure_chunks_embedded()
2025-12-08 15:56:02 +02:00
Yiorgis Gozadinos
16ba3a9962
Remove create_chunks_for_document from repositories
2025-12-08 15:56:01 +02:00
Yiorgis Gozadinos
17c7147a49
Remove pre-processor, no longer needed
2025-12-08 15:56:01 +02:00
Yiorgis Gozadinos
69b1afa534
Remove _create_and_chunk & _update_and_rechunk, keep storage independent of chunking
2025-12-08 15:55:31 +02:00
Yiorgis Gozadinos
ab57c55eaa
Drop ChunkWithMetadata, use good old Chunk
2025-12-08 15:55:30 +02:00
Yiorgis Gozadinos
ac439b6c69
Store raw text (do not contextualize) in chunk.text. Embed with contextualized text.
2025-12-08 15:55:26 +02:00
Yiorgis Gozadinos
c69b934eac
Incorporate additional docling meta in search(), expand_context() and relevant qa/research agents
2025-12-08 15:54:56 +02:00
Yiorgis Gozadinos
fe4ac530be
DB migration
2025-12-08 15:54:56 +02:00
Yiorgis Gozadinos
d267b2c433
LRU cache for DoclingDocument
2025-12-08 15:54:31 +02:00
Yiorgis Gozadinos
a5d8af07e9
Resolve DoclingDocument
2025-12-08 15:54:31 +02:00
Yiorgis Gozadinos
343272d6fb
Chunkers return chunks with metadata, i.e. (refs, labels, headings, page_numbers)
2025-12-08 15:54:31 +02:00
Yiorgis Gozadinos
44e8d7b340
Update Document & Chunk schemas to include DoclingDocument references.
2025-12-08 15:54:30 +02:00
Yiorgis Gozadinos
78f1d0cef5
No warnings, just info on db upgrade
2025-12-03 08:51:25 +02:00
Yiorgis Gozadinos
33b254ae3d
Always upgrade database on access. Explicitly create db by running init or HaikuRAG(path, create=True)
2025-12-02 16:49:55 +02:00
Yiorgis Gozadinos
3525fae625
Use EmbeddingModelConfig similar to ModelConfig for embeddings
2025-12-02 11:55:31 +02:00
Yiorgis Gozadinos
14bdb7cb27
async versions for convert_file() convert_text()
2025-11-26 17:46:25 +02:00
Yiorgis Gozadinos
2f54ee7d59
Replace allow_create with cleaner read_only. Elaborate on the test_info tests
2025-11-21 14:37:33 +02:00
Yiorgis Gozadinos
fdcf9a1a12
Show index stats in cli info command
2025-11-21 13:54:04 +02:00
Yiorgis Gozadinos
6a2f33b464
Create vector index from CLI, use configurable refine_factor
2025-11-21 13:54:01 +02:00
Yiorgis Gozadinos
30a0bbcff2
Accept both single chunks and lists for batch insertion in ChunkRepository.create()
2025-11-19 12:35:02 +02:00
Yiorgis Gozadinos
de5bb117fe
Chunker abstraction
2025-11-17 13:05:06 +02:00
Yiorgis Gozadinos
7ba1c2376e
Introduce converters for supporting more than local docling document conversion. Transform existing FileReader and utils to "docling-local" converter
2025-11-17 13:05:05 +02:00
Yiorgis Gozadinos
4599dd1481
Drop disable_auto_create flag, simply create only on write operations
2025-11-13 13:53:11 +02:00
Yiorgis Gozadinos
908296e155
Do not create unecessary DoclingDocument when chunks are provided. Rename private _create_with_docling & _update_with_docling to _create_and_chunk & _update_and_rechunk
2025-11-13 13:16:15 +02:00
Yiorgis Gozadinos
b8bc6d9b68
Let ruff know about our package structure
2025-11-05 17:47:27 +02:00
Yiorgis Gozadinos
5cd951a5f3
When checking for updates, use haiku.rag-slim as the package, since it is always installed
2025-11-05 10:30:59 +02:00
Yiorgis Gozadinos
2f9c907031
Restructure into uv workspace to support minimal and full installations
2025-11-04 17:59:12 +02:00
