mach-cctt 0.0.59

Test automated trades on the Mach exchange.

Cross Chain Trade Test (CCTT)

Test automated trades between multiple chains on Mach.

PyPI

Test PyPI

Overview

Test automated trades on Mach. Specify a start chain and symbol, and a policy determining how the destination token will be chosen (randomize chain and symbol, randomize chain and fix symbol). In each trade, the test wallet's entire balance of the source token will be sold for the destination token, and then the destination token because the new source token for the next trade. This continues until the script is killed by the user.

Usage

1. Install

   python -m pip install mach-cctt
   

2. Usage

   cctt --help
   

3. Example

   # The script will create log files and an account data file. Make a working directory for it.
   mkdir cctt/ && cd cctt/
   
   # Set password to avoid being prompted
   # export CCTT_PASSWORD="abc"
   
   cctt import # Import account
   cctt decrypt # Show public/private key
   cctt balances # Show account balances
   
   # Trade USDC on random chains, starting from optimism-USDC
   cctt run --source optimism-USDC --destination-policy fixed:USDC
   
   # Trade between random tokens on random chains, starting from arbitrum-USDT
   cctt run --source arbitrum-USDT --destination-policy random
   
   # If the --source is empty, then an appropriate source with a non-zero balance is chosen for you.
   cctt run --source --destination-policy random
   

   Notes:

   • You need to have a gas on every chain, or a gas of exactly 0 on chains that you do not wish the testing script to trade on, which will cause those chains to be disabled.
   • The --source token should be one that you hold a non-zero balance of in your wallet. If you set it to empty, it will be automatically chosen as any token with a non-zero balance in the wallet.
   • There is a known issue with trades filling for a single tick less than the order was placed at. As a preventative measure, any token balance of 1 tick in the wallet is treated as an empty balance.

Log Files

There are 4 log files created by the application:

• app.log - everything
• delayed_transactions.log - the source funds where not withdrawn in time
• stuck_transactions.log - the source funds were withdrawn but the destination funds were not received in time
• improper_fill.log - the trade filled on the source chain for 1 tick less than the order was placed for. This has been observed to happen primarily to USDM (on multiple chains)

Account File

A plaintext JSON file storing your account encrypted with your password will be created once you first generate or import a wallet. By default this file is called account.json but you can give a path to your own file via the --file PATH flag. This file is required to be present for any operation other than importing or generating a new wallet.

Development

Requirements

You need the just command runner and either Python 3.12.5+ with virtualenv, or pyenv.

Setup

1. Clone the repository

   git clone https://github.com/tristeroresearch/cross-chain-trade-test.git
   cd cross-chain-trade-test
   

2. Create and activate a virtual environment

   Using venv:

   python -m venv .venv
   source .venv/bin/activate
   

   Or using pyenv:

   pyenv virtualenv 3.12.5 cctt
   pyenv activate
   

3. Install dependencies

   just init
   

4. View available recipes

   just -l
   

   Note that there are recipes that allow you to run the application in-source without needed to build and package to PyPI first:

   just account_file=account.json import <private key> # Import a new account
   just decrypt # Show public and private key
   just balances # List balances of account
   just source=arbitrum-USDC policy=random run # Run the trade tester, starting from arbitrum-USDC and trading to random tokens
   just wallet=<public key> withdraw # Withdraw funds from the stored account to the given wallet
   

5. Building

   just build
   

Packaging

Make sure to bump the version number in pyproject.toml.

To upload to the test PyPI:

just upload-test

To upload to the main PyPI:

just upload

TODO

• Special logging for high-slippage orders
• Put all the AAVE stuff behind a feature flag (adds too many big dependencies)
• Switch over to new transactions module
• drain_eth is being skipped during withdrawals as we aren't calculating L1 gas fees, making it impossible to determine the amount to withdraw
• Use type 2 (EIP 1559) txs
• Resolve duplication between transfers.py and transactions.py
• Use Quote class from backend and deserialize with pydantic
• Unit tests