KiCadRoutingTools

# Python
__pycache__/
*.py[cod]
*$py.class
*.so
.Python
build/
develop-eggs/
dist/
downloads/
eggs/
.eggs/
lib/
lib64/
parts/
sdist/
var/
wheels/
*.egg-info/
.installed.cfg
*.egg

# IDE
.vscode/
.idea/
*.swp
*.swo
*~

# OS
.DS_Store
Thumbs.db
/.claude/settings.local.json

# Rust build outputs
rust_*/target/
*.pyd

# render_placement writes <board>_placement.png (and _B/_F on a two-sided
# board) BESIDE THE BOARD when -o is omitted -- render_placement.py:2225. So
# grading the corpus leaves 44 PNGs in kicad_files/, and a `git add -A` ships
# ~21 MB of generated renders into the commit. Measured: it did, into a #936
# follow-up branch, and only reading the diffstat before pushing caught it.
#
# The COUNT and the SIZE were re-measured by rendering all 33 tracked boards
# (22 one-sided renders + 11 _B/_F pairs = 44 files, 22,407,483 bytes). The
# first figures here, 31 PNGs / ~14 MB, were recalled from that branch rather
# than measured and were low on both. Treat the byte figure as an order of
# magnitude: it moves with the renderer's default size/supersample and with
# Pillow's compression, so it is not a constant to pin. The three patterns
# below were checked against the 44 actual filenames -- all matched.
/kicad_files/*_placement.png
/kicad_files/*_placement_B.png
/kicad_files/*_placement_F.png

# KiCad outputs (routed files, backups, project files)
/kicad_files/test_diffpair*kicad_pcb
/kicad_files/fanout_output*.kicad_pcb
/kicad_files/~*.kicad_pcb.lck
/kicad_files/routed_output*.kicad_pcb
/kicad_files/qfn_fanned_out.kicad_pcb
/kicad_files/fanout_starting_point.kicad_pcb
/kicad_files/routed_output_diff12.pcb
/kicad_files/fp-info-cache
/kicad_files/kit-out*.kicad_pcb
/kicad_files/flat_hierarchy_*.kicad_pcb
/kicad_files/interf_u_*.kicad_pcb
/kicad_files/sonde_u_*.kicad_pcb
kicad_files/.history/
/kicad_files/*kicad_pro
/*.kicad_prl
/kicad_files/*.kicad_prl

# Test-fixture boards used as INPUT by tests/ - keep tracked despite the broad
# output patterns above (test_bga_fanout_summary, test_diff_pair_route).
!/kicad_files/interf_u_unrouted_placed.kicad_pcb
!/kicad_files/routed_output.kicad_pcb

#Other
/git_version*

# #411 perturbation working dirs -- boards and scoreboards, never committed
wk/

# Generated by py_tools/repo_metrics.py and deployed from the workflow artifact;
# regenerated on every run, so committing it is pure churn (and a daily conflict
# with the scheduled job). The ARCHIVE under metrics/data/ is the part that is
# tracked, because it cannot be regenerated -- GitHub discards traffic after 14 days.
docs/site/

# #1081: the stage3d renderer's npm install (package-lock.json is tracked)
py_router/stage3d/node_modules/
# ...but the vendored three.js keeps upstream's own build/ layout (#1081)
!py_router/stage3d/vendor/three/build/