Jorge Jesus Alonso Ortiz

ON YOUR COMPUTER

One pip install

The base install is the four-package numerical core, NumPy, SciPy, pandas and Matplotlib, plus requests; every one of them ships with Pyodide. It covers the estimators and the data fetchers, and the first check runs offline, with no data and no API key.

pip install puremacro
python -m puremacro.examples.sign_restrictions_uhlig

Optional extras: [io] (parquet and Excel files), [backend] (numba, Apple-Silicon mlx), [cuda], [data], [narrative], [local-llm], [notebooks].

IN YOUR BROWSER

A playground with nothing to install

The JupyterLite playground runs puremacro on a Pyodide kernel inside the browser tab, with 64 notebooks in English and Spanish.

ON AN IPAD

Juno: Python on the iPad itself

Native Juno and a browser’s Pyodide kernel are different runtimes. Installation depends on the app’s Python version and available packages; browser evidence does not establish native-iPad compatibility. Tablet use remains best effort, with a local workstation as the supported target.

• Install. Add puremacro with Juno’s package manager. Parquet support is optional (the [io] extra, which needs pyarrow); without it, move data with pocket cartridges.
• Check the runtime. Start with runtime.report() to inspect the current environment, then run a small example on the device. Native Juno and browser kernels can differ in package availability and file access.
• Browser evidence. The Pyodide gate installs the wheel using the playground’s dependency rules and exercises selected workflows. This is evidence for that browser runtime and test set, not native-Juno or every-tablet validation. See the runtime guide for supported paths and limits.

from puremacro import runtime
print(runtime.report())

# In a Pyodide kernel (juno.sh or the playground), install with micropip instead:
import micropip
await micropip.install("puremacro")

WHEN THE TABLET IS NOT ENOUGH

Offload to Google Colab

Package a heavy job, such as a 10,000-draw MCMC or a large bootstrap, as a self-contained notebook, run it on Colab and load the result back as a portable cartridge.

from puremacro.runtime.colab import generate_colab_notebook, load_colab_result

generate_colab_notebook(
    """
import puremacro as pm
result = pm.dsge.estimate_sw07(n_draws=10000, n_chains=4)
""",
    mount_drive=True,
    save_path="sw07_offload.ipynb",
    output_filename="sw07_posterior.pmz",
)
# posterior = load_colab_result("sw07_posterior.pmz")