Morph MORPH
SIGN IN SIGN UP
opendatalab / MinerU UNCLAIMED

Transforms complex documents like PDFs into LLM-ready markdown/JSON for your Agentic workflows.

0 0 99 Python

COMMITS

master
May 26, 2026
M
refactor: rename index noise detection function to clarify its purpose in visual gap evaluation
myhloli committed
2e75bde
M
refactor: update mineru-vl-utils version constraint in pyproject.toml
myhloli committed
28dae4f
M
refactor: update hybrid OCR detection logic and streamline title splitting process
myhloli committed
40ee80d
M
refactor: improve index noise detection and vertical gap calculations between visual blocks
myhloli committed
56a9c8b
M
refactor: enhance local metric line resolution and cross-page handling in text merging
myhloli committed
4cfa9e8
X
Merge pull request #5021 from myhloli/dev
Xiaomeng Zhao committed
59b3013
May 25, 2026
M
refactor: replace omegaconf with yaml for configuration loading
myhloli committed
e9e89ef
M
refactor: remove albumentations dependency and update image processing logic
myhloli committed
935287d
X
Merge pull request #5018 from myhloli/dev
Xiaomeng Zhao committed
fd3b5d9
M
feat: implement style caching and retrieval for improved DOCX processing
myhloli committed
9a9ee38
M
refactor: remove cross-page merging option and enhance text merging logic
myhloli committed
84995a2
M
feat: add function to build rendered cell segments for target row in table merging
myhloli committed
8b26f7a
M
feat: update Dockerfiles to conditionally install qwen-vl-utils based on backend
myhloli committed
7c4ccdd
M
fix: restore qwen-vl-utils dependency for lmdeploy section in pyproject.toml
myhloli committed
53ccaea
M
feat: implement lazy loading for S3 IO classes to avoid mandatory boto3 dependency
myhloli committed
3a49633
M
refactor: simplify PDF classification by removing hybrid method and enhancing error logging
myhloli committed
d08dbea
M
feat: enhance Excel to HTML conversion by extracting non-empty rows and rendering compact tables
myhloli committed
7680a71
M
feat: enhance Excel to HTML conversion by extracting non-empty rows and rendering compact tables
myhloli committed
10f2b9b
M
refactor: update comments for clarity and maintain historical context in connected component logic
myhloli committed
9581530
M
feat: add functions to improve OOXML suffix guessing from zip packages and file paths
myhloli committed
e92376f
May 22, 2026
M
feat: add function to determine HTTP client backend visibility for server address configuration
myhloli committed
5d1a790
M
feat: add function to retrieve layout source blocks for improved page rendering
myhloli committed
7804621
M
feat: enhance text block merging logic with new functions for handling reference text lists
myhloli committed
2094cc5
M
feat: enhance OCR detection by adding support for hybrid text blocks and automatic merging based on OCR line hints
myhloli committed
4611b0b
M
refactor: consolidate block type checks for bounding box calculations in draw_bbox.py
myhloli committed
faef8a4
May 21, 2026
X
Merge pull request #5007 from myhloli/dev
Xiaomeng Zhao committed
fb0f6fa
M
feat: add function to identify transparent visual relation blocks in visual_magic_model_utils.py
myhloli committed
072a717
M
fix: remove hardcoded model path in vlm_analyze.py for better flexibility
myhloli committed
6b36b2c
M
fix: update model paths to version 2605 in enum_class and vlm_analyze
myhloli committed
01493c6
M
fix: update mineru-vl-utils version constraint to allow 1.0.0
myhloli committed
06da629