dolo.py

Dolo is a Python package to describe and solve economic models. It provides a simple classification scheme to describe many types of models, lets you write models as simple text (YAML) files, and compiles these files into efficient Python objects representing them. It provides many reference solution algorithms to find the solution of these models under rational expectations: local perturbations up to third order, perfect foresight solutions, policy iteration, value iteration, and time iteration. Most algorithms are either parallelized or vectorized, and are written in pure Python, so they can easily be inspected or adapted.

Documentation: https://econforge.github.io/dolo.py/

Install

Dolo is distributed as a conda package on the econforge channel on prefix.dev. With pixi, in a new or existing project:

pixi init my-project && cd my-project            # skip for an existing project
pixi workspace channel add --prepend https://prefix.dev/econforge
pixi add dolo

--prepend gives the econforge channel priority over conda-forge, so you get the latest Dolo release. Dolo can of course also be installed from PyPI with pip install dolo.

Example

Write a model in rbc.yaml:

name: Real Business Cycle

symbols:
  exogenous: [e_z]
  states: [z, k]
  controls: [n, i]
  parameters: [beta, sigma, eta, chi, delta, alpha, rho, zbar, sig_z]

definitions: |
  y[t] = exp(z[t])*k[t]^alpha*n[t]^(1-alpha)
  c[t] = y[t] - i[t]

equations:
  arbitrage: |
    chi*n[t]^eta*c[t]^sigma - w[t]                     ⟂ 0.0 <= n[t] <= inf
    1 - beta*(c[t]/c[t+1])^(sigma)*(1-delta+rk[t+1])   ⟂ -inf <= i[t] <= inf

Then import it, solve it, and simulate:

from dolo import yaml_import, time_iteration, simulate

model = yaml_import("rbc.yaml")
dr = time_iteration(model)
sim = simulate(model, dr)

print(dr)

Contributing

Contributors work from a clone of this repository. The sections below describe the development setup.

Setup

Using pixi (production and development environment)

pixi can be installed on a *nix system using the following command:

curl -fsSL https://pixi.sh/install.sh | sh

The dev environment brings together the test, lint, docs, and release features:

pixi install -e dev

To run unit and coverage tests:

pixi run -e dev test
pixi run -e dev cov

To format the code and check the types:

pixi run -e dev black
pixi run -e dev typecheck

To build and serve the documentation locally with zensical:

pixi run -e dev docs

To launch JupyterLab with the package installed in editable mode:

pixi run -e dev jlab