- Previous work on cycle arbitrage: https://github.com/ccyanxyz/uniswap-arbitrage-analysis
- General structure:
.
├── abi // smart contract ABI file
├── data // output data
├── go-ethereum // modified go-ethereum
├── reports // data analysis results
├── template // latex figure templates
├── test // test scripts
├── uniswap-v2-py // uniswap v2 python client
└── utils // helper scripts
- Go-ethereum: this is a modified go-ethereum client allowing us to export all transactions and receipts related to uniswap/sushiswap, the output is a 22959997 line file, each line contains a transaction and the corresponding receipt, key modifications:
- go-ethereum/core/blockchain.go: start from line 792, function ExportReceiptsN
- go-ethereum/core/transaction.go: start from line 49, Txdata structure
- Key Python scripts:
- get_all_pairs_thegraph.py: get all uniswap pairs via TheGraph API
- get_all_cycle_arbs.py: identify and extract all successful cyclic arbitrages
- count_cycles.py: count the cycle lengths
- get_start_token_stats.py: count the start tokens for cyclic arbitrages
- get_daily_revenue.py: get daily revenue of all cyclic arbitrages
- get_gas_trend.py: get daily gas cost for cyclic arbitrages
- get_profit_rev_dist.py: get profit/revenue distribution for cyclic arbitrages
- get_bot_stats.py: get stats of cyclic arbitrage contracts
- get_tx_list.py: get all transactions related to those arbitrage contracts we identified, including the UniswapV2Router contract
- get_failed_stats.py: get stats of failed cyclic arbitrages excluding those initiated from the UniswapV2Router contract
- get_failed_router_stats.py: get stats of all failed cyclic arbitrage transactions initiated from the UniswapV2Router contract
- get_top10bot_daily_revenue.py: get the daily revenue stats of the top 10 bots in terms of total revenue
- get_blockwise_stats.py: get blockwise reserves info of all uniswap pairs
- get_toppair_blockwise_reserves.py: get the blockwise reserves of those pairs in the top 10 arbitrage path
- draw_xxx.py: draw the analysis results with python matplotlib
- tex_xxx.py: draw the analysis results with Latex templates
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent