Quantum Optimiser

# Create environment
python3 -m venv venv && source venv/bin/activate

# Install packages
pip install praisonaiagents praisonai

# Set API key
export OPENAI_API_KEY="your-key"

# Create quantum circuits
cat > circuits.json << 'EOF'
{
  "QFT-8": {"qubits": 8, "depth": 45, "single_qubit_gates": 64, "two_qubit_gates": 28, "algorithm": "QFT"},
  "VQE-4": {"qubits": 4, "depth": 120, "single_qubit_gates": 96, "two_qubit_gates": 48, "algorithm": "VQE"},
  "QAOA-6": {"qubits": 6, "depth": 80, "single_qubit_gates": 72, "two_qubit_gates": 36, "algorithm": "QAOA"}
}
EOF

# Create optimization rules
cat > optimizations.json << 'EOF'
{
  "gate_cancellation": {"applicable": ["consecutive_inverse"], "reduction": 0.15},
  "gate_commutation": {"applicable": ["non_overlapping_qubits"], "reduction": 0.10},
  "template_matching": {"applicable": ["known_patterns"], "reduction": 0.20},
  "depth_reduction": {"applicable": ["parallel_gates"], "reduction": 0.25}
}
EOF

from praisonaiagents import Agent, AgentTeam, Task, tool
from pydantic import BaseModel
from typing import List
import json

# Structured output
class OptimizationResult(BaseModel):
    circuit_id: str
    original_depth: int
    optimized_depth: int
    original_gates: int
    optimized_gates: int
    depth_reduction_pct: float
    gate_reduction_pct: float
    estimated_fidelity: float
    optimizations_applied: List[str]

# Database persistence is configured via memory={} parameter

# Tools
@tool
def get_circuit(circuit_id: str) -> str:
    """Get quantum circuit details."""
    with open("circuits.json") as f:
        circuits = json.load(f)
    return json.dumps(circuits.get(circuit_id, {}))

@tool
def get_optimization_rules() -> str:
    """Get available optimization rules."""
    with open("optimizations.json") as f:
        return f.read()

@tool
def apply_optimization(circuit_id: str, depth: int, gates: int, optimization_type: str) -> str:
    """Apply optimization to circuit."""
    with open("optimizations.json") as f:
        rules = json.load(f)
    
    rule = rules.get(optimization_type, {})
    reduction = rule.get("reduction", 0)
    
    new_depth = int(depth * (1 - reduction * 0.8))  # Depth reduction
    new_gates = int(gates * (1 - reduction))  # Gate reduction
    
    # Fidelity estimation (simplified)
    fidelity = 0.99 ** (new_depth / 10)  # Decreases with depth
    
    return json.dumps({
        "optimization": optimization_type,
        "new_depth": new_depth,
        "new_gates": new_gates,
        "estimated_fidelity": round(fidelity, 4)
    })

@tool
def benchmark_circuit(original_depth: int, optimized_depth: int, original_gates: int, optimized_gates: int) -> str:
    """Benchmark optimization results."""
    depth_reduction = (original_depth - optimized_depth) / original_depth * 100
    gate_reduction = (original_gates - optimized_gates) / original_gates * 100
    
    return json.dumps({
        "depth_reduction_pct": round(depth_reduction, 2),
        "gate_reduction_pct": round(gate_reduction, 2),
        "speedup_estimate": round(original_depth / optimized_depth, 2)
    })

# Agents
analyzer = Agent(
    name="CircuitAnalyzer",
    instructions="Analyze quantum circuits. Use get_circuit tool.",
    tools=[get_circuit],
    memory={
        "db": "sqlite:///quantum_opt.db",
        "session_id": "quantum-optimization"
    }
)

optimizer = Agent(
    name="CircuitOptimizer",
    instructions="Apply optimizations. Use get_optimization_rules and apply_optimization tools.",
    tools=[get_optimization_rules, apply_optimization]
)

benchmarker = Agent(
    name="Benchmarker",
    instructions="Benchmark results. Use benchmark_circuit tool.",
    tools=[benchmark_circuit]
)

# Tasks
analyze_task = Task(
    description="Analyze circuit: {circuit_id}",
    agent=analyzer,
    expected_output="Circuit metrics"
)

optimize_task = Task(
    description="Apply optimizations to reduce depth and gates",
    agent=optimizer,
    expected_output="Optimized circuit metrics"
)

benchmark_task = Task(
    description="Benchmark optimization results",
    agent=benchmarker,
    expected_output="Optimization report",
    output_pydantic=OptimizationResult
)

# Run
agents = AgentTeam(agents=[analyzer, optimizer, benchmarker], tasks=[analyze_task, optimize_task, benchmark_task])
result = agents.start(circuit_id="VQE-4")
print(result)

# Optimize single circuit
praisonai "Optimize QFT-8 quantum circuit for depth reduction" --verbose

# With persistence
praisonai "Compare optimization strategies for VQE circuit" --memory --user-id quantum_team

# Batch optimization
praisonai "Optimize all circuits in circuits.json" --telemetry

framework: praisonai
topic: "quantum circuit optimization"
roles:
  analyzer:
    role: Circuit Analyst
    goal: Analyze quantum circuits
    backstory: Expert at quantum computing
    tasks:
      analyze:
        description: |
          Analyze circuit:
          - Qubit count
          - Circuit depth
          - Gate counts (1Q, 2Q)
          - Algorithm type
        expected_output: Circuit analysis
        
  optimizer:
    role: Circuit Optimizer
    goal: Reduce circuit complexity
    backstory: Expert at quantum compilation
    tasks:
      optimize:
        description: |
          Apply optimizations:
          - Gate cancellation
          - Gate commutation
          - Template matching
          - Depth reduction
        expected_output: Optimized circuit
        
  benchmarker:
    role: Performance Analyst
    goal: Measure optimization quality
    backstory: Expert at quantum benchmarking
    tasks:
      benchmark:
        description: |
          Benchmark results:
          - Depth reduction %
          - Gate reduction %
          - Fidelity estimate
          - Speedup factor
        expected_output: Optimization report

praisonai agents.yaml --verbose

# View optimization history
praisonai --history 10 --user-id quantum_team

# Check metrics
praisonai --metrics

# Export results
praisonai --save quantum_optimization_report

from praisonaiagents import Agent, tool
import json

@tool
def quick_circuit_estimate(qubits: int, depth: int, two_qubit_gates: int) -> str:
    """Quick circuit optimization estimate."""
    # Estimate potential reductions
    depth_reduction = min(0.4, depth / 200)  # Up to 40%
    gate_reduction = min(0.3, two_qubit_gates / 100)  # Up to 30%
    
    optimized_depth = int(depth * (1 - depth_reduction))
    optimized_2q = int(two_qubit_gates * (1 - gate_reduction))
    
    # Fidelity estimate
    fidelity = 0.99 ** (optimized_depth / 10) * 0.995 ** optimized_2q
    
    return json.dumps({
        "original": {"depth": depth, "2q_gates": two_qubit_gates},
        "optimized": {"depth": optimized_depth, "2q_gates": optimized_2q},
        "estimated_fidelity": round(fidelity, 4),
        "recommendation": "optimize" if fidelity > 0.9 else "simplify_algorithm"
    })

agent = Agent(
    name="QuantumAPI",
    instructions="Estimate quantum circuit optimization potential.",
    tools=[quick_circuit_estimate]
)

agent.launch(path="/optimize-circuit", port=8000)

curl -X POST http://localhost:8000/optimize-circuit \
  -H "Content-Type: application/json" \
  -d '{"message": "Estimate optimization for 8 qubits, depth 100, 40 two-qubit gates"}'

rm -f quantum_opt.db circuits.json optimizations.json
deactivate