Crate evm_amm_state 

A real-time AMM state engine built on a forked-EVM state cache (evm_fork_cache).

evm-amm-state composes a forked-EVM state cache (evm_fork_cache) with a set of protocol adapters into a pipeline that tracks a working set of AMMs, cold-starts their on-chain state into the cache, keeps exact-write protocols current straight from chain log events, and emits bounded repair requests for protocols whose events don’t carry final storage values. Once a pool’s quote read-set is warmed and current, swap simulations run fast and offline against it.

The pieces, roughly in pipeline order:

• adapters — per-protocol adapters (Uniswap V2, the Uniswap V3 family, Balancer V2, Solidly V2, and Curve — StableSwap/NG + CryptoSwap/Tricrypto-NG) over a single adapters::AmmAdapter trait. Each adapter knows how to cold-start a pool’s storage into an evm_fork_cache::cache::EvmCache, which log events to subscribe to, how to apply those events reactively, and how to simulate_swap against the cached state. The adapters::AdapterRegistry dispatches by pool key, and adapters::AmmSyncEngine drives the resync-capable live path on top of adapters::AmmReactiveHandler and the evm_fork_cache reactive runtime.

See the crate’s examples/adapter_pipeline.rs for an end-to-end demo that cold-starts a pool, subscribes to its events over a WebSocket endpoint, applies them reactively, and simulates a swap against the live-synced state.

Re-exports

pub use alloy_primitives;

pub use evm_fork_cache;

Modules

adapters

Protocol-neutral AMM adapter boundary.