dojo.market_agents.uniswapV3

Uniswap market agents.

class HistoricReplayAgent(dojo.agents.base_agent.BaseAgent[dojo.observations.uniswapV3.UniswapV3Observation]):

Agent that replays all historic actions exactly.

HistoricReplayAgent( chain: dojo.common.Chain, pools: list[str], block_range: tuple[int, int], mode: Literal['standard', 'swaps_only'] = 'standard', name: str = 'MarketAgent')

Initialize the policy.

Parameters
  • chain: Agent is for a specific chain.
  • pools: List of pools that the agent should be active on.
  • mode: In standard mode, all events are replayed. In swaps_only mode, only swaps are replayed
  • block_range: Range of blocks to replay.
def reward(self, obs: Any) -> float:

Reward is not relevant for the market agent.