How reinhardtmarta/hdm works
HDM is a Python library for deterministic high-dimensional vector projection, search, and adaptive DCT-II compression. Its main workflows are either a caller using HDMCompressor to reduce and reconstruct vectors or serialize sparse coefficients, or an API client submitting vectors for signatures and nearest-neighbor queries. Separate motion-analysis components derive trajectory metrics from scanner signatures, while the trajectory tensor computes angular coherence. The API, scanner, compressor, and motion modules are shown in source; other documented integration details were not independently inspected.
Main entry points: scanner.py HDM scanner · api.py FastAPI endpoints · compressor.py DCT compressor · kinematics.py Motion kinematics
Components 8
Projection and compression
- HDM scanner
highdimensionsmap/scanner.py - DCT compressor
highdimensionsmap/compressor.py
Trajectory analysis
- Motion kinematics
highdimensionsmap/kinematics.py - Robust tracker
highdimensionsmap/tracker.py - Coherence tensor
highdimensionsmap/tensor.py
HTTP service
- FastAPI endpoints
highdimensionsmap/api.py
Other components
- Library caller
- HTTP client
Connections 6
- Library caller → DCT compressor: compresses vectors
README.md - Library caller → DCT compressor: reconstructs vectors
README.md - HTTP client → FastAPI endpoints: submits requests
highdimensionsmap/api.py - FastAPI endpoints → HDM scanner: transforms and queries
highdimensionsmap/api.py - Motion kinematics → HDM scanner: transforms trajectories
highdimensionsmap/kinematics.py - Robust tracker → HDM scanner: transforms trajectories
highdimensionsmap/tracker.py