forcesim, a "force-based" market simulator

Dependencies/requirements

• C++ (C++20)

  • Google's glog logging library (https://github.com/google/glog/)

  • crow HTTP server (https://github.com/CrowCpp/Crow)

  • json from (https://github.com/nlohmann/json)

  • boost::program_options shared library

  • asio shared library

• Python
  • Python >= 3.9
  • pyyaml, matplotlib, aiohttp

Building

1. In the project root, adjust cmake.sh based on your system as needed. Set the following environment variables if needed:

   • INCLUDE_PREFIX: /usr on Linux (defaults to /usr/local)
   • VENDOR: path to the the C++ dependency directory: we will have, for example, $VENDOR/Crow/include

2. Clone the crow, json and glog repositories into the dependency directory

3. Adjust CMakeLists.txt if needed and run cmake.sh

4. In the build directory, run make forcesim, or other targets as needed

Python setup

• Navigate to forcesim/py
• Run something like python3.9 -m pip install .

Starting the simulator instance

$ ./forcesim --interface-address 127.0.0.1 --interface-port 18080

Running a simulation using the basic.py test

• Navigate to py/tests/basic
• Copy config.yml.sample to config.yml and set the output_path key
• Adjust and/or create file(s) in info-json, subscribers-json (currently some examples are already present)
• Run something like export PYTHON=python3.9 so that run.sh knows how to run Python (in case there are multiple versions available and/or there is no link to python)
• ./run.sh to run the simulation, generating output

Notes

Demonstration of basic usage with the basic.py test

forcesim.webm

Resulting graph