solsim

Discord Chat

solsim is the Solana complex systems simulator. It simulates behavior of dynamical systems—DeFi protocols, DAO governance, cryptocurrencies, and more—built on the Solana blockchain.

Philosophy

Define your system how you see fit.

solsim will simulate its behavior and collect its results in a straightforward, structured manner.

Usage

  1. Implement initialStep and step methods.
  2. From each, return the current state, i.e. a dictionary mapping variables to current values.
  3. Specify the variables you’d like to “watch.”
  4. Instantiate a Simulation, call .run().
  5. Receive a pandas DataFrame containing values of “watched” variables at each step in time.

Installation

pip install solsim

Development setup

First, install poetry.

Then:

git clone https://github.com/cavaunpeu/solsim.git
cd solsim
poetry install
poetry shell

Detailed usage

Systems using Solana

First, write the Solana programs in Rust or Anchor that comprise your system.

Next, copy the generated idl.json for each into a directory (e.g. named workspace) built as such:

workspace
└── target
    └── idl
        ├── program1.json
        ├── program2.json
        └── program3.json

Then,

  1. Build a system class that inherits from BaseSolanaSystem.
  2. Implement initialStep and step methods.
  3. Call super().__init__("workspace") in its __init__.

In 3, solsim exposes the following attributes to your system:

  • self.workspace: IDL clients for the Solana programs that comprise your system (via anchorpy).

For example, these clients let you interact with your respective programs’ RPC endpoints.

  • self.client: a general Solana client (via solana-py).

This client lets you interact with Solana’s RPC endpoints. Documentation here.

Finally,

  1. Define a watchlist: variables (returned in initialStep and step) you’d like to “watch.”
  2. Instantiate and run your simulation, e.g. Simulation(MySystem(), watchlist, n_steps=10).run().

Example

See the drunken escrow system.

Systems not using Solana

  1. Build a system class that inherits from BaseSystem.
  2. Implement initialStep and step methods.
  3. Define a watchlist: variables (returned in initialStep and step) you’d like to “watch.”
  4. Instantiate and run your simulation, e.g. Simulation(MySystem(), watchlist, n_steps=10).run().

Example

See the Lotka-Volterra system, inspired by cadCAD Edu.

Inspiration

solsim humbly builds on the shoulders of the giants that are and tokenspice, among others.

GitHub

View Github