myhloli
31553d29ea
ci: add PR gates and separate optional backend tests
2026-10-24 01:16:43 +08:00
myhloli
412d2c9b86
refactor(gradio): require 6.21 and simplify event chains
2026-10-23 01:11:35 +08:00
myhloli
65ae44f833
test: align DocVortex integration with native-only 0.6
2026-10-22 23:24:00 +08:00
myhloli
be6e78a914
chore(deps): require docvortex >=0.6.0,<0.7
2026-10-22 21:52:04 +08:00
myhloli
92613d738d
chore: align Python 3.11-3.14 support and release checks
2026-10-22 21:06:05 +08:00
myhloli
7ce2f07302
fix(cli): preserve full-document output after parsing
...
Refresh completed non-PDF ranges, keep whole-document content and cursor reads, and require DocVortex 0.5.18 for cross-page table alignment.
Refs opendatalab/MinerU#5654 and opendatalab/MinerU#5652 .
2026-10-22 21:04:59 +08:00
myhloli
57485941fb
feat(api): process JSON tasks and deliver per-file outputs
2026-10-22 18:41:58 +08:00
Xiaomeng Zhao
a2a4960b08
feat(onnx): add Linux WebGPU and compact OCR recognition outputs
2026-10-21 18:56:45 +08:00
myhloli
48524bfabb
feat(onnx): automatically select CPU or DirectML devices
...
Add auto/cpu/gpu configuration, Windows adapter discovery, isolated rec calibration, atomic preference persistence, and vendor session presets. Route layout, rec, and full MFR batches through DirectML with CPU fallback.
Use mutually exclusive Windows x64 DirectML dependencies and guard ORT/Magika distribution identity. Include calibration fixtures, three-machine results, the 0.8 threshold with ten UHD 770 repeats, and installation documentation.
Validation: 762 MinerU tests and 71 DocVortex tests passed; dependency checks and Windows Python 3.10-3.14 resolution passed. The 19-page OCR result matches the CPU baseline. DocVortex PyPI acceptance remains pending a compatible release; pinned source acceptance passed.
2026-10-19 11:02:29 +08:00
myhloli
a87030c1f3
chore(deps): bump mineru-llama-cpp and mineru-vl-utils minimum versions
2026-10-11 18:23:52 +08:00
myhloli
193d193aa8
fix(ci): require DocVortex 0.5.16 and repair docs and coverage checks
2026-10-11 04:19:10 +08:00
myhloli
2ce1df8719
Serve API documentation assets offline with proxy-aware authorization; Refs #5341
2026-10-10 00:30:39 +08:00
myhloli
e6d89a6af4
refactor: remove OpenCV from base dependencies
2026-10-08 15:08:18 +08:00
myhloli
1aa76acf1c
chore: add .zcodeignore for build artifacts and temporary files
2026-09-28 22:49:54 +08:00
myhloli
f712c9fd7b
feat(pdf): implement code span spacing protection and update tight spacing logic
2026-09-28 20:43:31 +08:00
myhloli
9669c12111
deps: update docvortex to version 0.5.2
2026-09-27 22:26:39 +08:00
myhloli
96c73ae847
fix(dependencies): update docvortex version to 0.4.24
2026-09-23 18:30:27 +08:00
myhloli
8d6adb6ec6
feat(mhtml): add MHTML support with analysis and preview functionality
2026-09-23 14:41:07 +08:00
myhloli
eb7e904b8a
fix(dependencies): update docvortex version to 0.4.22
2026-09-22 20:00:14 +08:00
myhloli
4cc9ee4fd9
fix(dependencies): update docvortex version to 0.4.20
2026-09-21 00:19:10 +08:00
myhloli
3103e954a4
feat(artifacts): update image naming convention for materialized assets and improve image resolution handling
2026-09-20 22:20:44 +08:00
myhloli
2563c3e434
fix(dependencies): update mineru-vl-utils version to 2.0.5
2026-09-20 00:31:36 +08:00
myhloli
40af48de43
feat(pdf): add support for PDFPageVectorGeometry in content processing
2026-09-18 22:27:01 +08:00
myhloli
35a956bce3
refactor: remove language option from API server and related components
2026-09-18 14:34:34 +08:00
myhloli
2d016d626e
fix: never fetch remote HTML images and keep external links
...
Remove fetch_remote_images=True from the API server HTML job and the
SDK default source context so remote images are never downloaded
server-side, closing the SSRF path on untrusted uploads. Bump docvortex
to >=0.4.12, which removes the fetch capability and lets saved bundles
keep external image links instead of requiring materialized assets.
Add a regression test covering parse, API-server zip output and link
retention with zero network requests.
Refs #5524
2026-09-18 03:59:09 +08:00
myhloli
09744440b7
fix: update docvortex dependency version to 0.4.11
2026-09-18 02:18:04 +08:00
myhloli
366fadb1c9
Promote TSV to an independent file suffix (ADR-0034)
...
- Route tsv through the shared CSV flash pipeline with an independent
file_type and metadata.file_suffix; bump docvortex pin to 0.4.9
- Add ADR-0034 documenting the content-addressed cache identity decision,
annotate ADR-0028, and update docs/next tier-group wording, telemetry
and e2e specs
- Cover detection, sync/async analyze, API server jobs and doclib
lifecycle for tsv in tests
2026-09-16 17:55:07 +08:00
myhloli
e18edc23c7
Update dependencies in requirements.txt and pyproject.toml; enhance documentation in quick_usage.md and cli_tools.md
2026-09-16 15:48:22 +08:00
myhloli
42624a867a
Bump docvortex dependency version to 0.4.7 for compatibility
2026-09-16 03:31:36 +08:00
myhloli
0fdabfce70
Bump mineru-llama-cpp dependency version to 0.1.2 for compatibility
2026-09-15 15:01:39 +08:00
myhloli
8a1d663097
Bump docvortex dependency version to 0.4.5 for compatibility
2026-09-15 02:20:21 +08:00
myhloli
0bcfb5c33f
Bump docvortex dependency version to 0.4.4 for compatibility
2026-09-14 20:36:36 +08:00
myhloli
53a8bc20c8
Enhance image handling in API client: support HTML image references and externalize assets during save
2026-09-14 19:14:30 +08:00
myhloli
51a3415c7c
Fix API page geometry remapping and align CI dependencies
2026-09-14 02:13:38 +08:00
myhloli
f1f212f898
Enhance PDF layout handling and documentation for layout geometry integration
2026-09-14 00:43:59 +08:00
myhloli
5d314ed32d
Update mineru-vl-utils dependency version to 2.0.3 in pyproject.toml
2026-09-12 05:32:20 +08:00
myhloli
119fdeec86
Add typing_extensions for compatibility and update tomllib encoding in tests
2026-09-12 05:26:53 +08:00
myhloli
080cc79c23
Rename 'mineru-gradio' command to 'mineru-webui' in pyproject.toml
2026-09-11 19:40:01 +08:00
myhloli
37bc36e297
Refactor model configuration to replace 'stack' with 'small_backend' and 'vlm.engine' for improved backend selection
2026-09-11 18:38:41 +08:00
myhloli
74dc0fac70
Add shared OCR resources and update model stack references for ONNX and Torch
2026-09-11 00:56:45 +08:00
myhloli
e29a1e16ca
Remove 'six' dependency and simplify image input validation in operators
2026-09-09 16:24:22 +08:00
myhloli
da3b3d04d6
Refactor imports to align with DocVortex 0.4 structure and update dependency versions
2026-09-09 14:45:06 +08:00
myhloli
f7c7504c9b
Use the native MLX server with eight active sequences
2026-09-09 04:43:17 +08:00
myhloli
6e5e881827
Fix Python 3.10 CI compatibility and minimum Torch version
2026-09-09 02:45:03 +08:00
myhloli
e8a3fe1d37
Merge Transformers 5 migration and integrate Python 3.14 document and backend support
2026-09-08 23:05:45 +08:00
myhloli
5d04281ca6
Update dependencies and refactor image handling in server implementation
2026-09-08 21:04:52 +08:00
myhloli
81c7b7a407
Update dependencies and refactor image handling in server implementation
2026-09-08 19:57:38 +08:00
myhloli
4861ce08e9
Enhance document metadata extraction by integrating source properties and improving error handling
2026-09-08 19:44:36 +08:00
myhloli
e32be37bf4
Add OFD preview functionality with lazy conversion and error handling
2026-09-08 16:43:13 +08:00
myhloli
73d16e6be7
Add new SVG icons for toolbar buttons and update Gradio header styles
2026-09-08 10:27:44 +08:00
myhloli