PyPI page
Author:
Sean O'Connell
Summary:
Review-ready data from AEC documents for LLMs: page maps, located text, tables, review markups, hidden CAD text, the quantities the text states and the measurement scale the PDF itself stores, searched exactly or forgivingly, plus deterministic drawing geometry carrying CAD layers and fill with confidence-scored annotation constructs (leaders, dimensions, title blocks, callouts) from PDF and DXF sheets, as framework-neutral LLM tools or over the Model Context Protocol
Latest version:
0.13.0
Required dependencies:
ezdxf
|
numpy
|
opencv-python-headless
|
pymupdf
|
rapidfuzz
Optional dependencies:
mcp
|
rapidocr-onnxruntime
Downloads last day:
56
Downloads last week:
409
Downloads last month:
1,652
7-day smoothing