capture-viewer: render extracted JSON preview artifacts

[Aymericr]

The manifest-first archive moved surface-mesh, point-cloud, and device-motion payloads out of inline stream data into JSON preview artifacts, but renderability and hydration still assumed inline — surface mesh reported "no data", point cloud and the motion trajectory silently vanished.

• streamHydratesJsonPayload is one predicate for both renderability and runtime hydration, so a stream can never be declared renderable without a hydration path.
• CaptureStreamLayer fetches the JSON artifact into the inline payload shape the layers already consume.
• The device-motion playback key is keyed on artifact identity and load state rather than the payload's contents, which can be megabytes.

Covered by stream-rendering.test.ts.

🤖 Generated with Claude Code

---

Note

Medium Risk
Changes capture viewer data loading and render gating for large JSON artifacts; failures surface as thrown fetch errors in the stream error boundary, but scope is limited to preview layers.

Overview
Fixes manifest-first archives where surface mesh, point cloud, and device motion no longer ship inline payloads—those layers were missing or empty because renderability and hydration still required stream.inline.

Adds streamHydratesJsonPayload so device motion, point cloud, and surface mesh streams with a JSON artifact (and no inline data) count as renderable and share the same hydration rule. CaptureStreamLayer fetches that artifact via useJsonArtifactPayload (loopback URL rewrite, abort on unmount) and feeds the result into the existing layer props as payload (inline ?? fetched).

For device motion, trajectory parsing and playback remounting use payload instead of inline only; the motion playback key uses artifact URL plus load state when the payload is fetched, avoiding JSON.stringify on multi‑MB JSON.

Tests assert isCaptureStreamRenderable is true for ready streams with JSON preview artifacts for all three layer kinds.

^{Reviewed by Cursor Bugbot for commit 80438fd. Bugbot is set up for automated code reviews on this repo. Configure here.}

[pascal]

I hit an error while handling your request (Model unavailable on AI Gateway free tier: Free tier users do not have access to this model. Upgrade to paid credits at https://vercel.com/d?to=%2F%5Bteam%5D%2F%7E%2Fai%3Fmodal%3Dtop-up for unrestricted…).

Please try again, rephrase, or reach out if it keeps failing.

Error id: 641e5cca-9915-462b-9a02-15eeb0a193da