myhloli
|
b4c57116c1
refactor: move PDF byte conversion logic to pdf_page_id and simplify image conversion process
|
3 weeks ago |
myhloli
|
ace7f76869
refactor: move PDF byte conversion functions to pdf_page_tools and simplify logic
|
3 weeks ago |
myhloli
|
5999f6664f
refactor: simplify PDF byte preparation by removing multiprocessing and enhancing direct conversion
|
3 weeks ago |
myhloli
|
c32ff88400
refactor: rename check_mac_env to check_sys_env and add Windows environment detection
|
3 weeks ago |
myhloli
|
4214634de8
feat: add timing logs for PDF byte preparation to improve performance monitoring
|
3 weeks ago |
myhloli
|
bffc6aff53
fix: streamline PDF conversion process by restructuring try-except block and ensuring proper resource management
|
3 weeks ago |
myhloli
|
05e114f8b9
feat: implement multiprocessing for PDF conversion to enhance performance
|
3 weeks ago |
myhloli
|
66d5f3dfd2
feat: refactor PDF image conversion to use get_end_page_id utility function and add multi-threading support
|
3 weeks ago |
myhloli
|
305e3a61e8
fix: disable tokenizers parallelism to prevent potential issues
|
4 weeks ago |
myhloli
|
28b278508f
fix: add error handling for PDF conversion in common.py
|
1 month ago |
myhloli
|
cf59949ba9
add tiff
|
1 month ago |
myhloli
|
c9315b8e10
Refactor suffix guessing to handle PDF extensions for AI files
|
2 months ago |
myhloli
|
907099762f
Normalize PDF suffix handling for AI files to be case-insensitive
|
2 months ago |
myhloli
|
2c356cccee
Fix suffix identification for AI files to correctly handle PDF extensions
|
2 months ago |
myhloli
|
bba31191a4
fix: update backend handling to enforce correct usage of vlm engines in sync and async modes
|
2 months ago |
myhloli
|
e60da65cca
feat: enhance file type detection using Magika for improved suffix guessing
|
2 months ago |
myhloli
|
55eaad224d
feat: add support for vlm 2.5
|
2 months ago |
myhloli
|
c3db578247
fix: add line sorting logic for bounding boxes in PDF processing
|
3 months ago |
myhloli
|
9a3a314916
refactor: add support for additional keyword arguments in client and analysis functions
|
4 months ago |
myhloli
|
67715001cf
fix: update environment variable names for formula and table enable in common.py and vlm_middle_json_mkcontent.py
|
5 months ago |
myhloli
|
677a301a9f
fix: standardize parameter names for formula and table options across modules
|
5 months ago |
myhloli
|
c29327fde6
feat: add support for additional image formats in file upload
|
5 months ago |
myhloli
|
86391acf28
fix: prevent server URL assignment for non-client backends in VLM processing
|
5 months ago |
myhloli
|
e7649e432c
refactor: modularize PDF processing and enhance async support for VLM backend
|
5 months ago |
myhloli
|
bfaf07c69f
fix: refactor device mode and virtual VRAM size handling in client.py and common.py
|
5 months ago |
myhloli
|
1383787bad
fix: refactor formula and table enable handling to use environment variables
|
5 months ago |
myhloli
|
98b8c4a967
refactor: streamline formula and table enable configurations in the pipeline
|
5 months ago |
myhloli
|
8737ebb2e2
feat: remove model path parameter from vlm_doc_analyze and streamline model loading
|
5 months ago |
myhloli
|
4eaa85fd31
refactor: update make mode constants to improve content list handling
|
5 months ago |
myhloli
|
02898cdd81
refactor: simplify file reading function and improve input validation
|
5 months ago |