← back to pin list

MVT option A — one piece, built to the rulings — mvt_option_a.py

kerfmaster pin list · The spec — what we build to

recordStates what the code does today. Auditable against the code — this is the only type that can go stale.

Content last changed 2026-08-13 — computed from the item itself, not typed.

Contract — design stage

This contract describes the code as it runs today.

takes vector ?

makes contours ?

fails if A join placed by us silently where no ruling covers it; a sliver the laser deletes that nobody counted; polylined curves in any emitted cut file.

Contract last changed 2026-08-13 — computed, not typed. Dated separately from the text above, so neither date can speak for the other.

PARKED 2026-08-13, UNREVIEWED — Aristide: "park this for now… I did see some errors, will discuss later." The errors are NOT yet named. Nothing on this page is accepted; do not build the cut DXF or the B/C previews on top of this geometry until he names them and they are fixed. The build below is the honest state of what was made, not a settled result.

Built 2026-08-13 on Aristide's go, from the rulings on record: mvt_option_a.pyviewer/mvt/option-a/ (png + stats.json + page). The knockout is cloud.difference(outer_disc.buffer(clear)) with clear = 0.169 in, the MEASURED median off his agreed photo (min 0.034; a uniform buffer is the generalizable reading of a hand-drawn cutback). It opens 22.62 in2 of interior the cloud filled — the “open circle interior” ruling — and severs the cloud into body (225.7 in2) and lobe (71.0 in2), as the option-C plate count already recorded.

Joins, all 0.22 in (measured off the photo; ours were 0.202, inside the raster error): three triangle→ring just off the corners aimed radially, side per corner chosen as the closest match to his example's 121/117/121 (got 118.2/122.4/119.4); one ring→body at bearing 287.5 (centre of the 272–303 run where the body faces the ring across exactly the knockout gap) — his bottom-right ruling; one ring→lobe at bearing 178.5, the lobe's nearest approach — OURS, not ruled, flagged on the page for Aristide: without it option A is two pieces. Bars lap a bar's width into each plate (mvt_prep's kiss-vs-lap lesson).

Proofs run every build: union is ONE polygon (raises otherwise); sub-kerf air-gap scan over all boundary loops (clean at kerf 0.007); sub-kerf slivers removed deliberately and counted (none here); narrowest metal = the joins themselves at 0.220; area 395.74 in2, 14.05 lb at 1/8. Interior open area after joins 57.37 in2.

Not yet: the true-arc cut DXF (fitted from source Béziers + circle arcs, never polylined from these measurement polygons), and the B/C previews + customer letter (one dataset, two renderers). Next steps in order.