refactor: extract shared apiClient and vectorMath utilities

[devin-ai-integration]

Summary

Two verbatim-duplicated code patterns were factored into shared utilities. Pure refactor — no behavior change.

1. Frontend HTTP client — the exact same request<T>() helper (plus API_BASE constant and phyverse-token localStorage/Bearer header handling) was copy-pasted in authApi.ts, aiConfigApi.ts, and cloudApi.ts, with the token/header block duplicated again in sendAiChat. Extracted into src/shared/utils/apiClient.ts:

export const API_BASE = import.meta.env.VITE_API_BASE_URL ?? '/api'
export const TOKEN_KEY = 'phyverse-token'
export function getStoredToken(): string | null
export function buildAuthHeaders(base?): Record<string,string>   // Content-Type + Bearer
export async function apiRequest<T>(path, options?): Promise<T>   // fetch + json + !ok -> throw

The three api modules now just call apiRequest(...); sendAiChat uses buildAuthHeaders(). AuthProvider.tsx now imports TOKEN_KEY / getStoredToken instead of redefining the token key literal.

2. Vector math in experiment data collectors — Math.sqrt(v.x*v.x + v.y*v.y + v.z*v.z) and its horizontal sqrt(x²+z²) / sqrt(dx²+dz²) variants were repeated across ~8 mechanics experiments. Extracted into src/shared/utils/vectorMath.ts:

magnitude(v: Vec3Like): number            // sqrt(x²+y²+z²)
magnitudeXZ(v: {x,z}): number             // sqrt(x²+z²)
distanceXZ(a, b): number                  // horizontal distance between two points

Applied to centripetalForce, curvilinearMotion, energyConservation, motionComposition, freeFall, tickerTimer, newtonSecondLaw, galileoIncline. Rapier's translation()/linvel() return {x,y,z} objects that satisfy the helper signatures, so call sites collapse e.g. to return magnitude(ball.rigidBody.linvel()).

Notes

• tsc -b passes; ESLint is clean on all touched files.
• Preexisting (not touched here): ESLint errors in api/ai/chat.ts, AiAgentPanel.tsx, sandboxStore.ts, and a broken vitest/jsdom setup in this environment (native binding / ESM resolution) — both unrelated to this refactor.

Link to Devin session: https://app.devin.ai/sessions/cee3bacf2df6421f8934c4bafc857a60
Requested by: @Zyz555444

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment, CI, and merge conflict monitoring

[vercel]

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

Project	Deployment	Actions	Updated (UTC)
phyverse	Ready	Preview, Comment	Jul 18, 2026 12:05pm

[devin-ai-integration]

✅ Runtime test — PR #4 (shared apiClient + vectorMath refactor)

Tested locally (Vite dev server). This is a behavior-preserving refactor; verified both extracted utilities at runtime.

apiClient (auth flow) — verified in-browser
Submitting the Sign In modal produced a correctly-formed request via the shared client:

{ "url": "/api/auth/login", "method": "POST",
  "headers": { "Content-Type": "application/json" },
  "body": "{\"email\":\"test@example.com\",\"password\":\"password123\"}" }

With a token in localStorage (phyverse-token), buildAuthHeaders adds Authorization: Bearer <token>. Backend absent → 404 (expected/acceptable; request shape is correct).

✅ vectorMath collectors — verified against a real headless Rapier sim (expand)

Ran the actual MECH-03/MECH-15 setup() + collectors against a stepped PhysicsWorld:

• Free Fall: speed = magnitude(linvel) exactly equals |vy|; height 5.0→4.21, vy 0→−3.92.
• Centripetal: radius=1.200 (distanceXZ), speed=1.999 (magnitude), ω=1.666 (magnitudeXZ/distanceXZ), F=1.665 = m·ω²·r ✓.

⚠️ Environment caveat + landing page

Rapier WASM is unstable in this test env (getObject(...).now is not a function during Vite pre-bundle; dispose() crash on reload; Node 20.18.1 < Vite's 20.19+). This blocked capturing live readouts in the 3D UI, so vectorMath was verified via the headless sim above. The scene does render and the ball physically falls on a good load.

Tested by Devin — see session for details.