Python SDK

Python tools for DexBotics simulations and testing.

Installation

The TypeScript SDK is the primary SDK. For Python simulations, use the simulation scripts directly from the repository.

bash

git clone https://github.com/dexbotics/market_prod.git
cd market_prod/sims/python

Basic Simulation

The repository includes a simple Python swarm simulator:

python

import random
import time
import json

def simulate():
    agents = [f"robot-{i:02d}" for i in range(5)]
    for t in range(8):
        agent = random.choice(agents)
        evt = {
            "t": t,
            "assignee": agent,
            "proof_hash": "0x" + "07" * 32
        }
        print(json.dumps(evt))
        time.sleep(0.05)

if __name__ == "__main__":
    simulate()

Running the Simulator

bash

cd sims/python
python swarm.py

This will output simulated agent events showing task assignments and proof submissions.

Custom Simulations

You can extend the simulation by:

1. Adding more agents: Modify the agents list to include more robots

2. Task complexity: Add additional event types and task states

3. Integration with SDK: Use the TypeScript SDK via subprocess calls to interact with real on-chain data

Multi-Agent Simulation

Example of a more complex simulation with task coordination:

python

import random
import time
import json

class SimulatedAgent:
    def __init__(self, agent_id):
        self.agent_id = agent_id
        self.tasks_completed = 0

    def accept_task(self, task_id):
        return {
            "event": "task_accepted",
            "agent": self.agent_id,
            "task_id": task_id,
            "timestamp": time.time()
        }

    def submit_proof(self, task_id):
        self.tasks_completed += 1
        return {
            "event": "proof_submitted",
            "agent": self.agent_id,
            "task_id": task_id,
            "proof_hash": "0x" + "07" * 32,
            "timestamp": time.time()
        }

def run_swarm(num_agents=10, num_tasks=20):
    agents = [SimulatedAgent(f"robot-{i:02d}") for i in range(num_agents)]

    for task_id in range(num_tasks):
        agent = random.choice(agents)

        accept_evt = agent.accept_task(task_id)
        print(json.dumps(accept_evt))
        time.sleep(0.1)

        proof_evt = agent.submit_proof(task_id)
        print(json.dumps(proof_evt))
        time.sleep(0.1)

    print("\nSimulation complete:")
    for agent in agents:
        print(f"{agent.agent_id}: {agent.tasks_completed} tasks")

if __name__ == "__main__":
    run_swarm(num_agents=10, num_tasks=20)

Integration with On-Chain Data

For production simulations that interact with the actual Solana program, use the TypeScript SDK:

python

import subprocess
import json

def create_task_via_sdk():
    result = subprocess.run([
        "npx", "ts-node", "../sdk/ts/examples/devnet-demo.ts"
    ], capture_output=True, text=True)

    return result.stdout

task_pubkey = create_task_via_sdk()
print(f"Created task: {task_pubkey}")