GitLab

refactor(magic-pdf): optimize model initialization and concurrency control

- Remove concurrency limit logic from app.py
- Update model initialization process in various modules
- Remove unused VRAM check for concurrency limit
- Refactor OCR model initialization in pdf_extract_kit.py
- Update txt_spans_extract_v2 function to use lang parameter instead of ocr_model

refactor(ocr): replace AtomModelSingleton with ocr_model_init for OCR model instantiation

- Remove usage of AtomModelSingleton for OCR model creation
- Add ocr_model_init function to initialize OCR model
- Update OCR model initialization in pdf_extract_kit.py and pdf_parse_union_core_v2.py
- Modify txt_spans_extract_v2 function to accept ocr_model as a parameter
- Update parse_page_core function to use ocr_model instead of lang for OCR processing

refactor(model): implement thread-safe OCR model initialization

- Add threading support for OCR model initialization
- Modify AtomModelSingleton to handle thread-specific instances
- Update PDFExtractKit and PDFParseUnionCoreV2 to use new thread-safe OCR initialization

build(deps): specify minimum version for ultralytics

- Update `ultralytics` dependency to version >= 8.3.43
- This change ensures compatibility with yolov8 for formula detection

refactor(magic_pdf): remove unused threading lock and model initialization code

- Remove threading.Lock import and usage
- Delete unused model initialization comments and code- Simplify OCR model initialization in both pdf_extract_kit.py and pdf_parse_union_core_v2.py

feat: update test case

fix(multi-threading ):Enable multi-threading support for PaddleOCR.

- Remove usage of AtomModelSingleton for OCR model initialization- Use ocr_model_init function for creating OCR model instance
- Update import statement to include ocr_model_init- Comment out old OCR model initialization code

- Remove usage of AtomModelSingleton for OCR model initialization
- Add import of ocr_model_init from model_init module
- Update OCR model initialization process to use ocr_model_init function
- Remove lock for OCR processing as it's no longer needed

- Remove usage of ModelSingleton class
- Initialize model directly using custom_model_init function
- Add self._lock attribute to PDFExtractKit class for thread safety- Replace local lock with self._lock for OCR processing

fix(model): simplify model initialization logic

fix: update notify

- Remove unnecessary threading.Lock in AtomModelSingleton
- Add threading.Lock to CustomPEKModel for OCR processing
- Simplify model initialization logic in AtomModelSingleton

perf(model): optimize model initialization

- Add condition to return existing model if already initialized
- Improve efficiency by avoiding redundant model creation

perf(model): add threading lock for OCR model initialization

- Introduce a lock to synchronize access to OCR model initialization- This change improves thread safety when multiple threads access the OCR model concurrently
- The lock ensures that the OCR model is initialized only once, even in multi-threaded scenarios

feat(gradio_app): implement dynamic concurrency limit based on VRAM

- Update VRAM checking logic in app.py and model_utils.py
- Add None and type checks for VRAM values
- Adjust concurrency limit calculation in app.py
- Modify clean_vram function to handle cases with no VRAM information

- Add get_concurrency_limit function to calculate concurrency limit based on VRAM
- Update clean_vram function and rename to get_vram for better clarity
- Apply concurrency limit to the to_markdown function in the Gradio app

master->dev

Release 0.10.5

Dev -> 0.10.5

fix(pre_proc): prevent errors when imageWriter is None