GitLab

- Remove PIL usage across multiple files
- Convert image processing functions to use NumPy arrays
- Update crop_img function to work with NumPy arrays
- Modify image loading and resizing to use NumPy and OpenCV
- Clean up unused imports and comments related to PIL

- Move environment variable settings for NPU, MPS, and other configurations to the global scope in doc_analyze_by_custom_model.py
- Remove redundant environment variable settings in pdf_extract_kit.py
- This change ensures consistent configuration across the application and avoids potential conflicts or duplicate settings

- Update RapidTable dependency to version 1.0.3
- Add support for sub-models in RapidTable
- Update magic-pdf configuration to include table sub-model
- Modify table model initialization to support sub-models
- Update table prediction logic to handle new output format

- Add support for NPU (Neural Processing Unit) when available
- Implement batch analysis for GPU and NPU devices
- Optimize memory usage and improve performance
- Update logging and error handling

- Merge title blocks that are close to each other horizontally
- Adjust line insertion logic for title blocks- Increase image size and decrease confidence threshold for layout detection
- Update DocLayoutYOLO model weights
- Refactor drawing of bounding boxes for different block types

- Add MPS support for Apple Silicon devices
- Implement empty_cache() for MPS devices
- Set PYTORCH_ENABLE_MPS_FALLBACK environment variable
- Adjust MFR model device allocation for MPS

- Update clean_memory function to support both CUDA and NPU devices
- Implement get_device function to centralize device selection logic
- Modify model initialization and memory cleaning to use the selected device
- Update RapidTableModel to support both RapidOCR and PaddleOCR engines

- Add NPU support for memory cleaning and model initialization
- Optimize table model initialization and prediction process
- Update memory utils to support NPU
- Add language parameter for table model

- Remove YOLO_VERBOSE environment variable from multiple files
- Set verbose=False in YOLOv8 prediction method to suppress logger output

- Add initial setup for layout detection
- Implement conditional cropping for tall images
- Skip cropping for wide images to improve performance
- Reuse Image object across layout detection steps

- Implement image cropping and pasting technique to enhance layout detection
- Adjust detected polygons to original image coordinates
- Add comments for better code readability

- 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

- 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

- 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

- 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

- 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

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

- Move page total time logging to doc_analyze_by_custom_model.py
- Remove page total time logging from pdf_extract_kit.py
- Add page_start timing variable to custom model analysis
- Update logger output format for page total time

- Update OCR utils to handle different box formats and improve angle calculation
- Modify PDF extraction kit to support OCR option and optimize processing flow
- Enhance PPOCR model to sort and filter detection boxes, improving text splitting accuracy

- Integrate RapidOCR with RapidTable model for table recognition
- Improve memory management for devices with <= 8GB VRAM
- Update table recognition process to use RapidOCR for RapidTable
- Add rapidocr-paddle dependency in setup.py

- Add RapidTable model support for table recognition
- Update table model configuration and initialization
- Modify table recognition process to use RapidTable when specified
- Add RapidTable dependency to setup.py

- Remove unused code for copying detection and recognition models
- Simplify OCR model initialization using atom_model_manager
- Delete unnecessary comments and empty lines

- Update StructTableModel to use the latest struct-eqtable library
- Add support for HTML table extraction in PDF Extract Kit
- Improve error handling and model initialization
- Update dependencies in setup.py for struct-eqtable

Modify line 397 to ensure compatibility with CPU execution, addressing the issue where specifying 'cpu' in config.json still results in a ValueError for expecting a cuda device but getting 'cpu' during demo execution.

- Remove import and usage of StructTableModel- Add support for TableMaster model- Update table model initialization logic to support TableMaster
- Log error and exit if StructEqTable is selected, as it's under upgrade
- Update README files to reflect changes in table parsing capabilities

- Lower the Y-axis overlap threshold for merging spans into lines from0.6 to 0.5
- Reduce the unclip ratio for OCR detection from 2.4 to 1.8

- Changed the confidence threshold from0.15 to 0.25 in the DocLayout_YOLO model prediction
- This adjustment aims to improve the accuracy of layout detection by filtering out low-confidence predictions

- Add new layout model option: DocLayout-YOLO
- Implement model initialization and prediction for DocLayout-YOLO
- Update configuration options to include new model- Modify existing code to support both LayoutLMv3 and DocLayout-YOLO models
- Update Gradio app to support more Custom Switch

- Remove unused functions such as split_long_words, ocr_mk_mm_markdown_with_para, etc.
- Simplify ocr_mk_markdown_with_para_core_v2 by removing unnecessary language detection and word splitting logic- Remove wordninja dependency from requirements
- Update ocr_model_init to include additional parameters for OCR model configuration

- Add detection for list and index blocks in OCR processing- Implement merging of list and index blocks across pages
- Update block types to include list and index categories
- Adjust text merging logic to handle new block types
- Modify layout drawing to distinguish list and index blocks

- Introduce a conditional memory cleanup step in the PDF extraction process
- Assess available GPU memory before deciding to perform memory cleanup- Log the time taken for garbage collection when it occurs
- This optimization helps to balance performance and resource utilization

- Enhance timing output precision to two decimal places for better readability- Calculate and log document analysis speed in pages per second
- Optimize logging for YOLO and table recognition processes
- Remove unnecessary comments and improve code efficiency

The clean_memory function has been removed from pdf_parse_union_core_v2.py due to it not being used.
This change streamlines the code and prevents potential confusion regarding its purpose.