Commit graph

  • 2831030f9b
    expose toc.json + heading_level/tree_depth in analysis VFS Yiorgis Gozadinos 2026-05-15 14:32:39 +0300
  • 1360303edf
    persist docling heading hierarchy on document_items Yiorgis Gozadinos 2026-05-12 12:06:30 +0300
  • f208730a5b
    Merge pull request #374 from tseaver/feat-cli-search_type Yiorgis Gozadinos 2026-05-18 16:39:28 +0300
  • 8026693450
    chore: ruff format Tres Seaver 2026-05-18 09:31:38 -0400
  • ad0161e6da
    chore: isort Tres Seaver 2026-05-18 09:31:12 -0400
  • f9ce3bc401
    refactor: move 'SearchType' to 'store.models.chunk' Tres Seaver 2026-05-18 09:27:34 -0400
  • d7e96f2658
    fix: typo Tres Seaver 2026-05-18 09:17:53 -0400
  • 4feac9683a
    fix: default 'search_type' to None Tres Seaver 2026-05-18 09:17:09 -0400
  • b98af1bf11
    Merge pull request #377 from ggozad/chore/deps Yiorgis Gozadinos 2026-05-18 16:13:12 +0300
  • 04415a9d09
    bump pydantic-monty, refresh deps, cap transformers<5 Yiorgis Gozadinos 2026-05-18 16:02:51 +0300
  • 1df6530d03
    Bump pydantic-ai, prepare for 2.* Yiorgis Gozadinos 2026-05-15 15:24:57 +0300
  • b1e8bdc839
    Bump docling Yiorgis Gozadinos 2026-05-15 15:09:05 +0300
  • b331dd0b76
    fix: default 'search_type' to 'None' in 'client.search.search' Tres Seaver 2026-05-18 06:52:03 -0400
  • 5079205d3d
    fix: use 'SearchType' for 'ChunkRepository.search' Tres Seaver 2026-05-18 06:50:49 -0400
  • 70da24b691
    docs: add '--search-type' examples to CLI docs Tres Seaver 2026-05-18 06:40:41 -0400
  • 7953de46b9
    fix: declare 'SearchType' for 'search_type' arg Tres Seaver 2026-05-18 06:33:04 -0400
  • b2ffc50168
    WIP: add '--search-type' option to CLI 'search' Tres Seaver 2026-05-17 08:55:43 -0400
  • f588e1150b
    add vllm Gemma-4-26B skill QA row to wix table Yiorgis Gozadinos 2026-05-18 09:43:13 +0300
  • c1923b85cd
    Add cross-encoder to all extras Yiorgis Gozadinos 2026-05-15 14:26:56 +0300
  • abc5a210a6
    update benchmarks Yiorgis Gozadinos 2026-05-15 10:04:17 +0300
  • 97a7eb246c
    vb 0.47.0 Yiorgis Gozadinos 2026-05-14 16:03:33 +0300
  • 9c19de93c0
    Merge pull request #373 from ggozad/feat/rerank-cross-encoder Yiorgis Gozadinos 2026-05-14 16:02:07 +0300
  • 5de4d50622
    coverage Yiorgis Gozadinos 2026-05-14 15:54:43 +0300
  • 57f273e000
    Minor fixes, CI should build Yiorgis Gozadinos 2026-05-14 15:41:57 +0300
  • 519afe6709
    Add cross-encoder reranking provider Yiorgis Gozadinos 2026-05-14 15:12:11 +0300
  • 0e4c462885
    Merge pull request #372 from ggozad/fix/rebuild-embed-memory Yiorgis Gozadinos 2026-05-14 15:37:00 +0300
  • dbb87e091f
    Ensure schema parity for _StagingChunkRecord Yiorgis Gozadinos 2026-05-14 15:02:02 +0300
  • 849eca94d0
    Make rebuild --embed-only idempotent across crashes Yiorgis Gozadinos 2026-05-14 13:59:08 +0300
  • a035b0f9e4
    Stream rebuild --embed-only through a staging table to bound memory Yiorgis Gozadinos 2026-05-14 13:34:51 +0300
  • 704bea6f4a
    Stabilize test_client_search ranking assertion Yiorgis Gozadinos 2026-05-14 14:37:27 +0300
  • d76c645862
    Update benchmarks for orb for gemma4 Yiorgis Gozadinos 2026-05-14 14:03:19 +0300
  • 862a209137
    vb 0.46.0 Yiorgis Gozadinos 2026-05-13 16:53:35 +0300
  • 129563b319
    Merge pull request #370 from ggozad/fix/picture-generation-opt-out Yiorgis Gozadinos 2026-05-13 16:32:31 +0300
  • a98ddc14b8
    Replace picture_description.enabled with processing.pictures enum Yiorgis Gozadinos 2026-05-13 15:29:32 +0300
  • bdab201c42
    Merge pull request #371 from ggozad/fix/image-results Yiorgis Gozadinos 2026-05-13 16:19:09 +0300
  • a6e5664c78
    Verify picture bytes before attaching to multimodal tool returns Yiorgis Gozadinos 2026-05-13 16:00:38 +0300
  • 072edad62f
    Merge pull request #368 from ggozad/feat/relax-compat-check Yiorgis Gozadinos 2026-05-13 15:43:37 +0300
  • 701e9d3632
    Relax embedding compat check to vector_dim only Yiorgis Gozadinos 2026-05-13 13:58:52 +0300
  • 675c9030a4
    Merge pull request #369 from ggozad/fix/embedding-batch-size Yiorgis Gozadinos 2026-05-13 14:21:10 +0300
  • 8b135c4d4d
    Expose embedding batch size as config Yiorgis Gozadinos 2026-05-13 14:09:47 +0300
  • d11b8a4e6e
    Merge pull request #367 from ggozad/feat/extra-body Yiorgis Gozadinos 2026-05-13 13:54:08 +0300
  • c6cd847299
    Add ModelConfig.extra_body for raw provider pass-through Yiorgis Gozadinos 2026-05-13 13:43:52 +0300
  • 0f4c4af495
    Cache the chunking tokenizer to avoid HF Hub 429 Yiorgis Gozadinos 2026-05-13 13:32:24 +0300
  • d512f9fbd1
    Disable Typer locals in CLI tracebacks Yiorgis Gozadinos 2026-05-13 13:24:05 +0300
  • bf1cd68f64
    Preserve cause chain on docling wrapper exceptions Yiorgis Gozadinos 2026-05-13 12:18:53 +0300
  • 2189002974
    Benchmark update Yiorgis Gozadinos 2026-05-12 17:20:48 +0300
  • 45dfd4b382
    Merge pull request #363 from ggozad/feat/chat-tui-markdown-streaming Yiorgis Gozadinos 2026-05-12 17:16:22 +0300
  • 674b80e8c9
    Support streaming for markdown Yiorgis Gozadinos 2026-05-12 11:03:31 +0300
  • bb39eb3809
    Merge pull request #364 from ggozad/fix/remote-image-ingestion Yiorgis Gozadinos 2026-05-12 15:58:05 +0300
  • a206f8bfcc
    Document fetch_remote_images and the docling-serve HTML gap Yiorgis Gozadinos 2026-05-12 15:50:33 +0300
  • df21286cc9
    add tests for mixed <img> sources and IMAGE-format wiring Yiorgis Gozadinos 2026-05-12 15:37:40 +0300
  • afd4665517
    Thread source_uri through URL ingest Yiorgis Gozadinos 2026-05-12 15:13:02 +0300
  • a25d9699ee
    Wire per-format options across docling-local Yiorgis Gozadinos 2026-05-12 15:01:05 +0300
  • c34a9dd466
    Merge pull request #360 from ggozad/feat/s3-monitor Yiorgis Gozadinos 2026-05-12 14:55:11 +0300
  • 9f8c9fe3b1
    cover s3 uri override, unsupported extension, watcher error path Yiorgis Gozadinos 2026-05-05 10:42:57 +0300
  • 6631e339ee
    record VCR cassettes for s3 source tests Yiorgis Gozadinos 2026-04-29 17:24:03 +0300
  • 6f6ebf27c8
    Document s3 storage & monitoring Yiorgis Gozadinos 2026-04-29 14:47:59 +0300
  • ef82e6eb08
    swap aioboto3 for obstore in the S3 client path Yiorgis Gozadinos 2026-04-29 14:38:53 +0300
  • f87d894b01
    add SeaweedFS integration tests for S3Watcher Yiorgis Gozadinos 2026-04-29 14:19:55 +0300
  • 37099988a6
    poll S3 prefixes with S3Watcher and wire into serve Yiorgis Gozadinos 2026-04-29 13:21:26 +0300
  • d009da06d6
    support s3:// document sources Yiorgis Gozadinos 2026-04-29 13:03:14 +0300
  • 8b8f44e7b2
    vb 0.45.0 Yiorgis Gozadinos 2026-05-08 12:22:35 +0300
  • 5c0598c3be
    Merge pull request #362 from ggozad/feat/multimodal-search Yiorgis Gozadinos 2026-05-08 12:20:04 +0300
  • 387032511e
    Simplify Yiorgis Gozadinos 2026-05-08 12:11:25 +0300
  • 788fac422e
    Clean up migration Yiorgis Gozadinos 2026-05-08 11:52:57 +0300
  • ab5cfdd04a
    Docs & cl Yiorgis Gozadinos 2026-05-08 11:04:32 +0300
  • 6af463699d
    record VCR cassette for vllm embedder end-to-end test Yiorgis Gozadinos 2026-05-06 20:01:22 +0300
  • 5e6a7148ed
    Update README Yiorgis Gozadinos 2026-05-06 13:24:44 +0300
  • cc59dc7203
    Use upload_large_folder for hf upload Yiorgis Gozadinos 2026-05-06 13:06:28 +0300
  • e55e3d13c8
    fix _patch_picture_descriptions silently dropping docling_pages Yiorgis Gozadinos 2026-05-06 12:57:41 +0300
  • ce7201271f
    split open_rag_bench dataset into orb_text and orb_multimodal variants Yiorgis Gozadinos 2026-05-06 12:49:03 +0300
  • 75f75ec505
    render attached pictures in inspector context modal under qa.model.vision Yiorgis Gozadinos 2026-05-06 12:00:12 +0300
  • 7fac35d2af
    rename embed_image_query to embed_image; run description check on text path Yiorgis Gozadinos 2026-05-05 16:15:47 +0300
  • 1ccb5b5fad
    auto-append /v1 to vllm base_url, matching ollama behavior Yiorgis Gozadinos 2026-05-05 16:08:14 +0300
  • 4478b7ce2b
    test v0.45.0 migration Yiorgis Gozadinos 2026-05-05 16:06:14 +0300
  • c96894a0f0
    fix cross-document picture dedup in agent and skill search tools Yiorgis Gozadinos 2026-05-05 15:56:10 +0300
  • 61a88edfd8
    drop bogus deprecation warning for processing.pictures Yiorgis Gozadinos 2026-05-05 14:46:41 +0300
  • 08696773fe
    Fix vcr test Yiorgis Gozadinos 2026-05-05 14:38:03 +0300
  • 49274ab51b
    attach picture bytes through skill search and document the schema Yiorgis Gozadinos 2026-05-05 14:33:57 +0300
  • 46b81f3fa1
    cover picture-description provider and rebuild edge cases Yiorgis Gozadinos 2026-05-05 14:20:37 +0300
  • ffc7b95375
    yield per document during rebuild for live progress reporting Yiorgis Gozadinos 2026-05-05 14:04:37 +0300
  • a45d82f206
    Update benchmark Yiorgis Gozadinos 2026-05-05 13:10:37 +0300
  • 5c3a864a0b
    rename open_rag_bench eval DBs to text/multimodal variants Yiorgis Gozadinos 2026-05-05 12:55:06 +0300
  • dca3188a48
    hydrate documents lazily during rebuild Yiorgis Gozadinos 2026-05-05 12:47:45 +0300
  • ff82d36c2f
    add rebuild --descriptions: run VLM over stored picture bytes only Yiorgis Gozadinos 2026-05-05 12:02:46 +0300
  • 5fad0aa5c6
    warn when picture descriptions silently fail to come back Yiorgis Gozadinos 2026-05-05 11:44:10 +0300
  • 2038d43435
    collapse pictures enum to picture_description.enabled boolean Yiorgis Gozadinos 2026-05-05 11:16:08 +0300
  • e88321767d
    always store picture bytes; collapse converter to a single zip path Yiorgis Gozadinos 2026-05-05 10:40:40 +0300
  • 17e36fc069
    Add first openrag results Yiorgis Gozadinos 2026-05-05 09:50:22 +0300
  • 35d5f4416e
    fill in vLLM error-path and MCP image-query coverage Yiorgis Gozadinos 2026-05-05 09:39:20 +0300
  • ee41bc676f
    remove defensive code Yiorgis Gozadinos 2026-05-04 15:12:27 +0300
  • e8d89aa035
    dedup picture-only chunks at the search-result layer Yiorgis Gozadinos 2026-05-04 13:31:15 +0300
  • 33e462d987
    add real vLLM integration test Yiorgis Gozadinos 2026-05-04 13:16:31 +0300
  • aa3e9406cf
    auto-append /v1 to per-model Ollama base_url, fix flaky tests Yiorgis Gozadinos 2026-05-04 13:12:19 +0300
  • 65d9c74224
    expose image-as-query through MCP and the CLI. Yiorgis Gozadinos 2026-05-04 11:46:45 +0300
  • ff656504d3
    make client.search() polymorphic on query type: str | bytes | PIL.Image.Image. Bytes/PIL queries embed via embed_image_query and run vector-only against the chunks table Yiorgis Gozadinos 2026-05-04 11:22:22 +0300
  • 21f12cd770
    add vision: bool flag on ModelConfig (default False). Gate the agent search tool's BinaryContent attachment on qa.model.vision so picture bytes are only sent to vision-capable QA models. Yiorgis Gozadinos 2026-05-04 10:47:21 +0300
  • 6037f26673
    drop mlx Yiorgis Gozadinos 2026-05-03 19:05:49 +0300
  • c9fdad15a9
    bump docling-core to >=2.74.1 so that local docling-local and docling-serve chunkers run the same MarkdownTableSerializer Yiorgis Gozadinos 2026-05-03 18:40:44 +0300
  • c9227c649f
    emit synthetic picture chunks at ingest under multimodal embedders. Yiorgis Gozadinos 2026-05-03 17:03:56 +0300