GalloDaSballo / fair-selling

Work towards a trust-minimized (one day trustless) token swapper. Currently: OnChainPricing & Cowswap Contract Swapper. One day: Automatic Limit Order Swap Protocol. Moved to Badger Repo
https://github.com/Badger-Finance/on-chain-pricer
18 stars 6 forks source link

Fair Selling

A BadgerDAO sponsored repo of Open Source Contracts for:

Why bother

We understand that we cannot prove a optimal price because at any time a new source of liquidity may be available and the contract cannot adapt.

However we believe that given a set of constraints (available Dexes, handpicked), we can efficiently compute the best trade available to us

In exploring this issue we aim to:

Current Release V0.3 - Pricer - BribesProcessor - CowswapSeller

Notable Contracts

CowswapSeller

OnChain Integration with Cowswap, all the functions you want to:

BribesProcessor

Anti-rug technplogy, allows a Multi-sig to rapidly process CowSwap orders, without allowing the Multi to rug Allows tokens to be rescued without the need for governance via the ragequit function

OnChainPricingMainnet

Given a tokenIn, tokenOut and AmountIn, returns a Quote from the most popular dexes

Dexes Support

Covering >80% TVL on Mainnet. (Prob even more)

Example Usage

BREAKING CHANGE: V3 is back to view even for Balancer and UniV3 functions

isPairSupported

Returns true if the pricer will return a non-zero quote NOTE: This is not proof of optimality

    /// @dev Given tokenIn, out and amountIn, returns true if a quote will be non-zero
    /// @notice Doesn't guarantee optimality, just non-zero
    function isPairSupported(address tokenIn, address tokenOut, uint256 amountIn) external returns (bool)

In Brownie

quote = pricer.isPairSupported(t_in, t_out, amt_in)

findOptimalSwap

Returns the best quote given the various Dexes, used Heuristics to save gas (V0.3 will focus on this) NOTE: While the function says optimal, this is not optimal, just best of the bunch, optimality may never be achieved fully on-chain

    function findOptimalSwap(address tokenIn, address tokenOut, uint256 amountIn) external virtual returns (Quote memory)

In Brownie

quote = pricer.findOptimalSwap(t_in, t_out, amt_in)

Mainnet Pricing Lenient

Variation of Pricer with a slippage tollerance

Notable Tests

Proof that the math is accurate with gas savings

These tests compare the PricerV3 (150k per quote) against V2 (1.6MLN per quote)

brownie test tests/heuristic_equivalency/test_heuristic_equivalency.py

Benchmark specific AMM quotes

TODO: Improve to just use the specific quote

brownie test tests/gas_benchmark/benchmark_pricer_gas.py --gas

Benchmark coverage of top DeFi Tokens

TODO: Add like 200 tokens TODO: Compare against Coingecko API or smth

brownie test tests/gas_benchmark/benchmark_token_coverage.py --gas

Notable Test from V2

Run V3 Pricer against V2, to confirm results are correct, but with gas savings

brownie test  tests/heuristic_equivalency/test_heuristic_equivalency.py

Deployments

WARNING: This list is not maintained and may be out of date or incorrect. DYOR.

bveCVX Bribes Processor: https://etherscan.io/address/0xb2bf1d48f2c2132913278672e6924efda3385de2

bveAURA Bribes Processor: https://etherscan.io/address/0x0b6198b324e12a002b60162f8a130d6aedabd04c

Pricers can be found by checking processor.pricer()