{"name": "ArchaeoTrace", "package_name": "ai_vectorizer", "description": "Local-first contour tracing for historical maps with optional AI assistance and experimental DEM reconstruction.", "about": "ArchaeoTrace helps users digitize elevation contours from scanned historical maps and review locally generated terrain hypotheses.\n\nFeatures:\n- Multi-scale 9/15/31px Ink Centerline with a bounded 320px direction-aware Live-Wire\n- Literal 0-100% cursor/path coordinate blending; green preview shows the path accepted by the next click\n- Optional Smart Recovery for weak Ink segments; default OFF with an explicitly installed EfficientSAM-Ti model\n- Freehand plus LSD, HED, MobileSAM, SAM (ViT-B), and Legacy Canny under Advanced / Legacy methods\n- Elevation attributes and optional spot-height capture\n- Experimental linear-TIN DEM and hillshade generation from saved elevation data\n- Developer-only reproducible benchmarks with publication ranking disabled until independent review\n\nExternal dependencies: default Ink and Freehand use NumPy supplied by QGIS.\nSciPy and scikit-image are optional Ink optimizations. Smart Recovery requires\nONNX Runtime and separately installed fixed-hash weights. LSD, HED, and SAM\nrequire OpenCV; MobileSAM and full SAM additionally require requests,\nPyTorch, their pinned backend package, and separately downloaded weights.\nInstallation commands and platform notes are in the linked homepage README.\nQGIS 3.22/Python 3.8 remains a source and no-pip contract target; optional\npip stacks require Python 3.10 or newer. See the release-readiness record\nfor the exact locally executed and configured-but-not-run QGIS matrix.\n\nProcessing is local-first and map data is not sent to a remote inference service.\nModel downloads and missing-model source checks are user-initiated network actions.\nTerrain outputs are reviewable hypotheses rather than archaeological ground truth.", "homepage": "https://github.com/lzpxilfe/AI-Vectorizer-for-Archaeology", "repository": "https://github.com/lzpxilfe/AI-Vectorizer-for-Archaeology", "tracker": "https://github.com/lzpxilfe/AI-Vectorizer-for-Archaeology/issues", "author": "lzpxilfe(balguljang2)", "tags": ["contour", "ai", "vectorization", "archaeology", "digitizing", "historical maps"], "downloads": 1196, "latest_version": "0.1.6", "versions": [{"version": "0.1.6", "experimental": true, "qgis_min": "3.22.0", "qgis_max": "4.99.0", "downloads": 115, "uploaded_by": "lzpxilfe", "upload_datetime": "2026-09-09T04:23:39.208612"}, {"version": "0.1.5", "experimental": true, "qgis_min": "3.22.0", "qgis_max": "4.99.0", "downloads": 159, "uploaded_by": "lzpxilfe", "upload_datetime": "2026-08-25T21:54:22.325184"}, {"version": "0.1.4", "experimental": true, "qgis_min": "3.22.0", "qgis_max": "3.99.0", "downloads": 531, "uploaded_by": "lzpxilfe", "upload_datetime": "2026-03-20T06:26:23.794994"}, {"version": "0.1.3", "experimental": true, "qgis_min": "3.22.0", "qgis_max": "3.99.0", "downloads": 82, "uploaded_by": "lzpxilfe", "upload_datetime": "2026-03-18T07:23:36.870918"}, {"version": "0.1.2", "experimental": true, "qgis_min": "3.22.0", "qgis_max": "3.99.0", "downloads": 156, "uploaded_by": "lzpxilfe", "upload_datetime": "2026-02-20T03:15:22.478797"}, {"version": "0.1.1", "experimental": true, "qgis_min": "3.22.0", "qgis_max": "3.99.0", "downloads": 153, "uploaded_by": "lzpxilfe", "upload_datetime": "2026-02-08T03:54:33.583042"}]}