Sync Notation
Estimate the dominant resonance phase of movement notation so dancers lock onto what is really driving the piece.
Phase estimationΒ· resonance
Section Β· Quantum
full primer βThe hook.
Movement notation is matched against a target resonance using quantum phase estimation; choreographers see a single dial that snaps when alignment is strongest.
Why this primitivePhase estimation is the right primitive here because movement notation reduces to a resonance problem; the kernel returns a result you can drop straight into the UI.
Kernel
a Quantum Phase Estimation kernel returning the dominant eigenphase of a problem operator
Drives the UI as
a phase dial (0 to 2Ο) tied to a resonance/match indicator
Appendix A
The mega-prompt.
This prompt is engineered to ship in a single Lovable build. Real Quantinuum Guppy/Selene circuit runs in the Linux sandbox at build time and the results are baked in as JSON. read the build strategy β
~14.2 KB292 lines1 msg Β· ~5 credits
Build "Sync Notation"βa one-page quantum phase estimation app for choreographers to discover dominant resonance phases in movement notation.
**CONSTRAINTS (read first):**
- FREE Lovable plan (~5 credits). Build in ONE message only.
- No accounts, auth, Cloud, or database. Static app only.
- One workspace page + small "About the quantum" footer strip.
**BUILD ORDER (follow exactly to ship in 1 message):**
1. In sandbox: `pip install guppylang selene-sim`
2. Create `quantum/kernel.py`: real .py file with @guppy Quantum Phase Estimation kernel (Guppy reads via inspect.getsource). Kernel estimates dominant eigenphase of a problem operator, β€8 qubits.
3. Create `quantum/run.py`: import kernel, compile via guppylang, run on Selene emulator (256 shots) over 16 unitaries β 16 phase estimates. Write `src/data/quantum-results.json` with array of `{input: string, output: {phase: number (0-2Ο), shots: number}}` records.
4. Execute `quantum/run.py` once during build. Commit real Selene output to JSON. No Python at runtime.
5. Frontend: TanStack Start, route `/`. Import JSON statically.
**FRONTEND COMPONENTS (shadcn/Tailwind):**
- **Card**: left panel (input pickerβdropdown or Tabs listing 16 movement notations)
- **Slider**: center panel (phase dial, 0β2Ο, read-only, tied to selected record)
- **Badge**: resonance match indicator (color: accent, text: phase angle + "lock" status)
- **Tabs**: right panel ("Visualization" tab shows phase as radial gauge; "Quantum trace" tab shows raw phase + collapsible code block with kernel.py source)
- Dark theme (bg-slate-950), one accent color (e.g., rose, violet, or amber matched to dance mood), one typeface (e.g., Inter).
**KEY DETAILS:**
- Selene API: `selene_sim.build()`, `Quest`, `run_shots()`
- Grid: 16 movement notations β 16 precomputed phases in JSON
- Quantum trace: display phase angle (0β2Ο) + inline kernel.py source (toggle visibility) so judges verify real Phase Estimation on Selene
- Tactile micro-interaction: fade/scale on input selection
- Every number on screen traces to real Selene shot
**DELIVERABLE:** Single deploy. All quantum results precomputed. Ship in one message.
--- QUANTINUUM APPENDIX (self-contained reference β paste verbatim) ---
[1] INSTALL (Lovable Linux sandbox):
python -m pip install --no-cache-dir guppylang selene-sim
[2] IMPORTS (use exactly these β no other quantum libs):
# quantum/kernel.py and quantum/run.py top
from guppylang import guppy
from guppylang.std.builtins import result
from guppylang.std.quantum import qubit, h, cx, rx, ry, rz, measure, discard, t as tgate, tdg
from guppylang.std.angles import angle, pi
from selene_sim import build, Quest
import math, json, sys, tempfile, importlib.util, uuid
from pathlib import Path
[3] HARD RULES (violating any breaks the build):
- @guppy reads source via inspect.getsource β kernels MUST live in a real .py file on disk. No exec(), no REPL strings, no inline templates.
- Allowed gate set ONLY: h, rx, ry, rz, cx, tgate, tdg. There is NO native ccx/toffoli, cswap, cphase, or crz β decompose using the snippets in [7].
- Qubit ownership: a qubit passed to a function is moved. You MUST measure() or discard() every qubit exactly once; never reuse after measure.
- Angle hygiene before baking a float into generated source:
theta = ((theta + math.pi) % (2.0 * math.pi)) - math.pi
and write it with repr: f"... {theta!r} ..." (str(float) can truncate).
[4] SELENE SHOT LOOP (canonical):
compiled = my_kernel.compile()
runner = build(compiled)
shots = []
for shot in runner.run_shots(Quest(), n_qubits=N, n_shots=S):
shots.append({str(lbl): int(v) for lbl, v in shot})
# N = MAX number of qubits simultaneously LIVE in the kernel.
# measure(q) releases the slot, so one ancilla reused across k windows still counts as 1.
[5] DRIVER PATTERN β sweep a kernel over many inputs (closures do NOT work):
ROOT = Path(__file__).resolve().parent.parent
if str(ROOT) not in sys.path: sys.path.insert(0, str(ROOT))
def run_one(params: dict, shots: int = 256):
# Bake params as literals into a fresh .py file that imports your kernel helpers.
src = (
"from quantum.kernel import guppy, my_helper\n"
"@guppy\n"
"def program() -> None:\n"
f" my_helper({params['a']!r}, {params['b']!r})\n"
)
tmp = Path(tempfile.gettempdir()) / "qprogs"; tmp.mkdir(exist_ok=True)
name = f"prog_{uuid.uuid4().hex[:8]}"
path = tmp / f"{name}.py"; path.write_text(src)
spec = importlib.util.spec_from_file_location(name, path)
mod = importlib.util.module_from_spec(spec)
sys.modules[name] = mod # register BEFORE exec_module
spec.loader.exec_module(mod)
runner = build(mod.program.compile())
out = []
for shot in runner.run_shots(Quest(), n_qubits=5, n_shots=shots):
out.append({str(l): int(v) for l, v in shot})
return out
[6] PER-QUBIT INTEGER DECODE (host-side):
# kernel emits: for j in range(n): result(f"x{j}", measure(q[j]))
def decode(rec, n):
x = 0
for j in range(n): x |= (rec.get(f"x{j}", 0) & 1) << j
return x
[7] DECOMPOSITION LIBRARY (copy verbatim into quantum/kernel.py):
# ---- Toffoli (CCX) from H, CX, T, Tdg β 6-T standard decomposition ----
@guppy
def toffoli(c1: qubit, c2: qubit, tgt: qubit) -> None:
h(tgt)
cx(c2, tgt); tdg(tgt)
cx(c1, tgt); tgate(tgt)
cx(c2, tgt); tdg(tgt)
cx(c1, tgt); tgate(c2); tgate(tgt)
h(tgt)
cx(c1, c2); tgate(c1); tdg(c2)
cx(c1, c2)
# ---- CSWAP (Fredkin) from CX + Toffoli ----
@guppy
def cswap(c: qubit, a: qubit, b: qubit) -> None:
cx(b, a)
toffoli(c, a, b)
cx(b, a)
# ---- Controlled phase exp(i*theta) on |11> from rz + cx ----
@guppy
def cphase(c: qubit, d: qubit, theta: float) -> None:
rz(d, angle(theta / 2.0))
cx(c, d)
rz(d, angle(-theta / 2.0))
cx(c, d)
# ---- Amplitude-encoded feature state (3 floats in [0,1] β 2-qubit state) ----
@guppy
def prep_features(q0: qubit, q1: qubit, a: float, b: float, c: float) -> None:
ry(q0, angle(a))
ry(q1, angle(b))
cx(q0, q1)
rz(q1, angle(c))
# ---- SWAP test kernel; HOST inverts: F = clamp(2*P(anc=0) - 1, 0, 1) ----
@guppy
def swap_test(ai: float, bi: float, ci: float,
aj: float, bj: float, cj: float) -> None:
anc = qubit()
pi0 = qubit(); pi1 = qubit()
pj0 = qubit(); pj1 = qubit()
prep_features(pi0, pi1, ai, bi, ci)
prep_features(pj0, pj1, aj, bj, cj)
h(anc)
cswap(anc, pi0, pj0)
cswap(anc, pi1, pj1)
h(anc)
result("anc", measure(anc))
discard(pi0); discard(pi1); discard(pj0); discard(pj1)
# n_qubits = 5 for swap_test above.
[8] CLASSICAL CROSS-CHECK (NumPy reference β commit alongside quantum result):
import numpy as np
I = np.eye(2); X = np.array([[0,1],[1,0]])
def RY(t): c,s = math.cos(t/2), math.sin(t/2); return np.array([[c,-s],[s,c]])
def RZ(t): return np.array([[np.exp(-1j*t/2),0],[0,np.exp(1j*t/2)]])
CX = np.array([[1,0,0,0],[0,1,0,0],[0,0,0,1],[0,0,1,0]])
def prep_psi(a,b,c):
s = np.kron(RY(a)@np.array([1,0]), RY(b)@np.array([1,0]))
s = CX @ s
s = np.kron(I, RZ(c)) @ s
return s
def classical_fidelity(p, q):
return float(abs(np.vdot(prep_psi(*p), prep_psi(*q)))**2)
# Commit BOTH quantum and classical values per record:
# {"input": [...], "quantum": 0.873, "classical": 0.881, "shots": 256}
[9] FRONTEND HANDOFF (TanStack Start):
# quantum/run.py last step:
Path("src/data/quantum-results.json").write_text(json.dumps({
"records": records, # list of {input, quantum, classical, ...}
"circuit": {"n_qubits": 5, "shots": 256},
"kernel_source": Path("quantum/kernel.py").read_text(), # show on /code tab
}, indent=2))
# In React: import data from "@/data/quantum-results.json" (resolveJsonModule:true).
# Render data.records; render data.kernel_source in a <pre> inside the Quantum trace tab.
[10] BUILD ORDER (do not deviate β keeps you under 5 credits):
1. Write quantum/kernel.py with the helpers from [7] you actually need.
2. Write quantum/run.py: driver pattern from [5], Selene loop from [4], cross-check from [8], JSON write from [9].
3. Execute once: python quantum/run.py (commit src/data/quantum-results.json).
4. Build the React route: static JSON import, shadcn Card + Tabs + Slider + Badge wiring.
5. Deploy. No Python at runtime; every on-screen number traces to committed Selene output.
[11] OUTPUT CONTRACT β selene_run v1 (write this exact shape to src/data/quantum-results.json):
{
"schemaVersion": 1,
"experiment": "<idea-slug>",
"title": "<idea title>",
"description": "<one plain-English sentence β what a non-quantum dev sees>",
"kernel": { "snippet": "<first 40 lines of quantum/kernel.py>", "qubits": <int>, "shotsPerRow": <int> },
"verdict": { "text": "<one-sentence headline of the result>", "good": true },
"metrics": [ { "name": "<label>", "value": <number>, "unit": "<%|bits|Hz|...>", "good": true } ],
"series": [ { "id": "<id>", "kind": "histogram"|"bar"|"line", "title": "...",
"xLabel": "...", "yLabel": "...", "yKeys": ["count"],
"points": [ { "label": "<x>", "values": { "count": <num> } } ] } ]
}
Reason: one schema = one renderer = a polished demo without bespoke React per idea.
All metric / series VALUES must come from real Selene shot data committed in step 3 β no mocks.
[12] RENDERER β create src/components/SeleneRunView.tsx with EXACTLY this body (no deps beyond react + tailwind):
import * as React from "react";
type Point = { label: string; values: Record<string, number> };
type Series = { id: string; kind: "histogram"|"bar"|"line"; title: string;
xLabel?: string; yLabel?: string; yKeys: string[]; points: Point[] };
type Metric = { name: string; value: number; unit?: string; good?: boolean };
export type SeleneRun = {
schemaVersion: 1; experiment: string; title: string; description: string;
kernel: { snippet: string; qubits: number; shotsPerRow: number };
verdict: { text: string; good: boolean };
metrics: Metric[]; series: Series[]; notes?: string;
};
const fmt = (n: number) => Math.abs(n) >= 100 ? n.toFixed(0) : Math.abs(n) >= 1 ? n.toFixed(2) : n.toFixed(3);
function Bars({ s }: { s: Series }) {
const max = Math.max(1, ...s.points.flatMap(p => s.yKeys.map(k => p.values[k] ?? 0)));
return (
<div className="space-y-1">
{s.points.map((p, i) => (
<div key={i} className="flex items-center gap-2 text-xs">
<div className="w-20 truncate text-muted-foreground">{p.label}</div>
<div className="flex-1 h-3 bg-muted rounded-sm overflow-hidden">
<div className="h-full bg-primary" style={{ width: `${(100*(p.values[s.yKeys[0]]??0))/max}%` }} />
</div>
<div className="w-12 text-right tabular-nums">{fmt(p.values[s.yKeys[0]]??0)}</div>
</div>
))}
</div>
);
}
function Line({ s }: { s: Series }) {
const W=320, H=120, P=20;
const ys = s.points.map(p => p.values[s.yKeys[0]] ?? 0);
const min = Math.min(...ys), max = Math.max(...ys), span = max - min || 1;
const pts = ys.map((y, i) => {
const x = P + (i*(W-2*P))/Math.max(1, ys.length-1);
const yy = H - P - ((y - min)/span)*(H - 2*P);
return `${x},${yy}`;
}).join(" ");
return (
<svg viewBox={`0 0 ${W} ${H}`} className="w-full h-32">
<polyline fill="none" stroke="currentColor" strokeWidth="2" points={pts} className="text-primary" />
</svg>
);
}
export function SeleneRunView({ run }: { run: SeleneRun }) {
return (
<div className="space-y-6">
<header>
<div className="text-xs uppercase tracking-wider text-muted-foreground">{run.experiment}</div>
<h2 className="text-2xl font-semibold">{run.title}</h2>
<p className="text-sm text-muted-foreground">{run.description}</p>
<div className={`mt-2 inline-block px-3 py-1 rounded-full text-xs ${run.verdict.good?"bg-emerald-500/15 text-emerald-400":"bg-amber-500/15 text-amber-400"}`}>
{run.verdict.text}
</div>
</header>
<section className="grid grid-cols-2 md:grid-cols-4 gap-3">
{run.metrics.map((m, i) => (
<div key={i} className="rounded-lg border border-border p-3">
<div className="text-[10px] uppercase tracking-wider text-muted-foreground">{m.name}</div>
<div className="text-xl font-semibold tabular-nums">{fmt(m.value)}<span className="text-xs text-muted-foreground ml-1">{m.unit}</span></div>
</div>
))}
</section>
<section className="space-y-6">
{run.series.map(s => (
<div key={s.id} className="rounded-lg border border-border p-4">
<div className="flex items-baseline justify-between mb-3">
<div className="text-sm font-medium">{s.title}</div>
<div className="text-[10px] text-muted-foreground">{s.xLabel} / {s.yLabel}</div>
</div>
{s.kind === "line" ? <Line s={s} /> : <Bars s={s} />}
</div>
))}
</section>
<footer className="text-[11px] text-muted-foreground">
kernel: {run.kernel.qubits} qubits Β· {run.kernel.shotsPerRow} shots/row
</footer>
</div>
);
}
Then in the route: import data from "@/data/quantum-results.json"; <SeleneRunView run={data as any} />.
Quantum trace tab: <pre>{data.kernel.snippet}</pre>.
[HOOK] PHASE ESTIMATION β eigenphase readout.
k control qubits + eigenstate register. Apply h on each control, then controlled-U^{2^j}
(decompose controlled-U with cphase from [7] when U is a Z-rotation).
Inverse QFT on controls (run the QFT primitive from the qft hook with negated angles), then measure.
Decoded integer y β phase Ο = y / (2**k).
selene_run mapping:
metrics: [ {"name":"phase","value":phi,"unit":"rad"},
{"name":"control qubits","value":k,"unit":""} ]
series: [ {"id":"phase-bins","kind":"bar","title":"Phase-bin counts",
"xLabel":"y/2^k","yLabel":"shots","yKeys":["count"],
"points":[{"label":f"{y/(2**k):.2f}","values":{"count":c}} for y,c in sorted(hist.items())]} ]Market sizing.
TAM
$8.0B
the global dance industry (~$5B; >2M studios worldwide).
SAM
$1.7B
the 21% of that market actively buying movement notation-adjacent software.
SOM
$50M
a realistic 3% capture of the serviceable slice in years 1β3 via the hackathon launch and creator-led distribution.
Indicative figures for hackathon pitches β refine with your own research before raising.
Adjacent entries.
choreography drafting
Tuner Drafting
Estimate the dominant resonance phase of choreography drafting so dancers lock onto what is really driving the piece.
ensemble synchronizationLock Synchronization
Estimate the dominant resonance phase of ensemble synchronization so dancers lock onto what is really driving the piece.
improvisation promptsAnchor Prompt
Estimate the dominant resonance phase of improvisation prompts so dancers lock onto what is really driving the piece.
rehearsal schedulingPivot Scheduling
Estimate the dominant resonance phase of rehearsal scheduling so dancers lock onto what is really driving the piece.