James Sweeney · Resume (PDF) · LinkedIn · jmsweene@gmail.com
A set of small, focused, well-tested tools for gdsfactory chip layouts: quality-assurance checks, geometry and routing analysis, layout diffing, reproducible hashing, and layer transforms. Each is independent, dependency-light (numpy / gdstk, gdsfactory optional), MIT-licensed, and covered by a CI test suite.
| Repo | What it does |
|---|---|
| gdsqa | QA suite: port/connection, LayerStack, hierarchy, and label checks behind one CLI and one issue format. |
| denscheck | Per-layer area and a windowed CMP density DRC (min/max density + neighbour gradient). Complements gf.fill. |
| routecheck | Multi-route audit: route-to-route crossings, minimum spacing, length budget, and manhattan fraction. |
| gdsnorm | A canonical layout hash, invariant to polygon order, vertex rotation, winding, and sub-grid noise. Reproducible CI regression keys. |
| gdsreview | Semantic, hierarchy-aware layout diff with a PR-style review report (beyond pixel XOR). |
| curvecheck | Curvature profile, curvature-DRC, and a bend-loss proxy for waveguide paths. |
| gdsstats | A layout metrics report card for CI: counts, hierarchy depth, and a flatten-cost / reuse-factor metric. |
| layerops | Spec-driven layer remap / keep-drop / per-layer boolean-merge, written back to GDS. |
Each repo has a runnable examples/quickstart.py, a pytest suite (python run_checks.py suite), and a README
with a sample report.
Python, numpy, gdstk for GDS I/O and boolean geometry, matplotlib for figures. Validated against analytic / ground-truth references rather than golden files.