Map auto-hole regions into KCL artifacts

[Irev-Dev]

Summary

• Consumes the new hole_region_mappings data returned by the engine for region(...).
• Rebuilds auto-created hole loops as inner region paths in KCL and the artifact graph.
• Lets tags from cutout sketch entities, like plateRegion.tags.hole, resolve to the generated side face after extrusion.
• Fixes point-and-click edge selection for auto-created cutout edges, including when the selection resolves to the original sketch segment instead of the generated region segment.
• Points kittycad-modeling-cmds at KittyCAD/modeling-api#1272 so this branch can build against the response shape it needs.

Why

Take the following part:

When this part is made by extruding a single region with holes inside it, those holes are automatically cut out by the engine. Before this change, modeling app only rebuilt the selected outer region loop from region_mapping.

That meant the app did not have KCL paths or artifact graph paths for auto-created hole loops. So even once the engine knew that the hole side face came from the original sketch circle, the modeling app could not preserve that relationship all the way through to face tags or later refactors.

This PR wires that data through. The selected outer loop still uses region_mapping, while each auto-created cutout loop is represented from hole_region_mappings.

It also wires the point-and-click path through the same relationship. In practice, selecting a cutout edge can hand the command code either the generated region segment or the original sketch segment that produced it. The generated segment has the sweep and adjacency data needed for tagging, while the original sketch segment often does not, so selection helpers now follow originalSegId back to the generated region segment when needed.

Implementation

• region(...) now reads hole_region_mappings from CreateRegion and CreateRegionFromQueryPoint.
• The KCL region sketch keeps the selected outer loop in paths and stores auto-created cutout loops in inner_paths.
• Region tag remapping now considers both the outer region mapping and all hole mappings, so tags on cutout sketch entities can become region tags.
• The artifact graph creates synthetic region path artifacts for auto-created holes and links them back to the outer region through outer_path_id.
• Extrude and sweep artifact updates now consume all inner paths linked to the outer region, not just the single legacy inner_path_id.
• Point-and-click sweep lookup now resolves inner region paths through outerPathId.
• Point-and-click edge tagging now follows originalSegId from a source sketch segment to the generated region segment so it can use the generated segment's common faces.

Alternative considered

• We could have left this only in the engine response and relied on later face API or Z0006 work to recover the relationship.
• That is not enough because the modeling app artifact graph still needs concrete path and segment artifacts for the auto-hole loop before downstream selection/refactor code can reason about it.
• Keeping this PR focused on the KCL/artifact graph mapping also keeps it separable from the larger Z0006 refactor branch. The Z0006 branch can use this data once it is based on a branch that has the refactor implementation.

Point-and-click behavior

• The first selection case is straightforward: if the selected edge is a generated inner region segment, the helper can follow outerPathId to the swept outer region path.
• The second case is less obvious: sometimes the selection is the original sketch segment. That segment is useful for source identity, but it does not have the sweep or extrusion adjacency data needed by getCommonEdge.
• For that case, the selection helpers look for generated region segments whose originalSegId points at the selected source segment, then use that generated segment for sweep and common-face lookup.

Why the Mermaid snapshot changes

• This PR also starts populating Path.seg_ids for region paths.
• Before this, region segment artifacts pointed back to their region path through Segment.path_id, but the region path itself did not list those segments.
• That made the artifact graph only half-wired for regions. Downstream code needs to traverse from the region path to its boundary segments, including auto-created inner cutout loops.
• The Mermaid renderer shows both Path.seg_ids and Segment.path_id. Once both directions exist for the same path and segment pair, it deduplicates the relationship into an undirected-looking --- edge.
• So the broad Mermaid snapshot churn is expected: it is mostly Path <--x Segment becoming Path --- Segment, reflecting that region paths now own their segments in both directions.

Relies on https://github.com/KittyCAD/engine/pull/4777, can't merge until it is (and kittycad.ts is updated).

Screenshare.-.2026-07-13.1_05_11.PM.mp4

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
modeling-app	Ready	Preview, Comment	Jul 13, 2026 3:23am

[codspeed-hq]

Merging this PR will not alter performance

✅ 126 untouched benchmarks
⏩ 151 skipped benchmarks¹

---

_{Comparing kurt-auto-hole-region-artifact-mapping (7249e91) with main (29f9e2e)²}

Footnotes

1. 151 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

2. No successful run was found on main (6f87f4e) during the generation of this report, so 29f9e2e was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩

[Irev-Dev]

superseded by #12467