Skip to content

Marimo Notebooks

This repository contains interactive Marimo notebooks.

Features

Notebooks live in book/marimo/notebooks/ (configured via MARIMO_FOLDER in .rhiza/.env):

Notebook Description
Experiment1.py Basic CTA strategy implementation using moving averages
Experiment2.py Improved CTA strategy with volatility scaling
Experiment3.py Advanced CTA strategy with price filtering and oscillators
Experiment4.py CTA strategy with optimization and risk scaling
Experiment5.py Advanced CTA strategy with correlation-based optimization

Running the Notebooks

Using the Makefile

From the repository root:

make marimo

This will start the Marimo server and open all notebooks in the notebooks folder as specified in .rhiza/.env.

Validating Notebooks

To validate that all notebooks run without errors:

make marimo-validate

Running a Specific Notebook

To run a single notebook:

marimo edit book/marimo/notebooks/Experiment1.py

The notebooks include inline dependency metadata, making them self-contained:

uv run book/marimo/notebooks/Experiment1.py

This will automatically install the required dependencies and run the notebook.

Notebook Structure

Marimo notebooks are pure Python files (.py), not JSON. This means:

  • ✅ Easy version control with Git
  • ✅ Standard code review workflows
  • ✅ No hidden metadata
  • ✅ Compatible with all Python tools

Each notebook includes inline metadata that specifies its dependencies:

# /// script
# requires-python = ">=3.12"
# dependencies = [
#     "marimo==0.13.15",
#     "numpy==2.3.0",
#     "pandas==2.3.0",
#     "plotly==6.1.2",
#     "cvxsimulator==1.4.3"
# ]
# ///

Configuration

Marimo is configured in pyproject.toml to properly import the local package:

[tool.marimo.runtime]
pythonpath = ["src"]

The notebook folder is configured in .rhiza/.env:

MARIMO_FOLDER=book/marimo/notebooks

CI/CD Integration

The .github/workflows/rhiza_marimo.yml workflow automatically:

  1. Discovers all .py files in book/marimo/notebooks/
  2. Runs each notebook in a fresh environment
  3. Verifies that notebooks can bootstrap themselves
  4. Ensures reproducibility

This guarantees that all notebooks remain functional and up-to-date.

Creating New Notebooks

To create a new Marimo notebook:

  1. Create a new .py file in the notebooks directory: 

    marimo edit book/marimo/notebooks/my_notebook.py

  2. Add inline metadata at the top: 

    # /// script
# requires-python = ">=3.12"
# dependencies = [
#     "marimo==0.13.15",
#     # ... other dependencies
# ]
# ///

  3. Start building your notebook with cells

  4. Test it runs in a clean environment: 

    uv run book/marimo/notebooks/my_notebook.py

  5. Commit and push — the CI will validate it automatically

Learn More

Tips

  • Reactivity: Remember that cells automatically re-run when their dependencies change
  • Pure Python: Edit notebooks in any text editor, not just Marimo's UI
  • Git-Friendly: Notebooks diff and merge like regular Python files
  • Self-Contained: Use inline metadata to make notebooks reproducible
  • Interactive: Take advantage of Marimo's rich UI components for better user experience