Caching

Forked States

Titanoboa automatically caches states when running in fork mode. It uses python's builtin sqlite support This allows forking to take less time and use less memory.

The cache file is by default located at ~/.cache/titanoboa/fork/chainid_<0x_chainid>-sqlite.db (one sqlite db file is generated per chain id). To customize this folder, pass the cache_dir argument to the fork function (see fork). In case cache_dir is set to None, the cache will be disabled. To debug cache hits and misses, pass debug=True to the fork function.

Warning

Caching a fresh block might lead to incorrect results and stale cache files.

Warning

When running boa in parallel (e.g. with pytest-xdist), the cache file will be shared between all processes. This can lead to more requests being made than necessary if multiple processes are requesting the same block.

Compilation results

By default, Titanoboa caches compilation results on Disk. The location of this cache is by default ~/.cache/titanoboa and the files are called {sha256_digest}.pickle.

To change the cache location, call set_cache_dir with the desired path. In case the path is None, caching will be disabled. Alternatively, call disable_cache to disable caching.

Etherscan

The utility from_etherscan fetches the ABI for a contract at a given address from Etherscan and returns an ABIContract instance.

Given Etherscan is rate-limited, it is recommended to cache the results. In order to enable this, Titanoboa uses the requests_cache package.

If the package is available in the environment, all requests to Etherscan will be cached.