Appendix B — Python Setup and Code Reference

All code in this book runs inside a pixi-managed Python environment. This appendix explains how to set it up and reproduce any example from the text.

Installing pixi

pixi is a fast cross-platform package manager. Install it with:

# macOS / Linux
curl -fsSL https://pixi.sh/install.sh | bash

# Windows (PowerShell)
iwr -useb https://pixi.sh/install.ps1 | iex

Setting Up the Environment

From the root of the repository:

pixi install          # download and install all packages
pixi run setup        # register the Jupyter kernel

This creates a self-contained environment in .pixi/envs/default/. It does not touch your system Python or any other environments you have installed.

Rendering the Book

pixi run render       # build HTML → docs/
pixi run render-pdf   # build PDF → docs/
pixi run preview      # live preview in browser (auto-reloads on changes)

The render task sets QUARTO_PYTHON automatically so Quarto uses the pixi environment’s Python.

Key Libraries

Library	Purpose	First used
`numpy`	Numerical arrays, linear algebra	Ch 1
`scipy`	Statistical distributions, optimization	Ch 2
`matplotlib`	Static plots	Ch 1
`plotly`	Interactive plots (HTML output)	Ch 3
`pandas`	Tabular data	Ch 4
`scikit-learn`	ML algorithms	Ch 11
`pymc`	Bayesian modeling, MCMC	Ch 3
`networkx`	Graph structures	Ch 5
`scikit-fuzzy`	Fuzzy sets and inference	Ch 9–10
`pgmpy`	Probabilistic graphical models	Ch 5
`dowhy`	Causal effect estimation	Ch 13

Common Patterns

Random number generation (reproducible):

import numpy as np
rng = np.random.default_rng(seed=42)   # seed for reproducibility
x = rng.normal(0, 1, 100)

Plotting (matplotlib):

import matplotlib.pyplot as plt
fig, ax = plt.subplots(figsize=(8, 4))
ax.plot(x, y, color="#4e79a7", linewidth=2)
ax.set_xlabel("x"); ax.set_ylabel("y")
plt.tight_layout()
plt.show()

Bayesian model (PyMC):

import pymc as pm
with pm.Model() as model:
    mu = pm.Normal("mu", mu=0, sigma=1)
    obs = pm.Normal("obs", mu=mu, sigma=0.5, observed=data)
    trace = pm.sample(1000, tune=500, progressbar=False)

Fuzzy set (scikit-fuzzy):

import numpy as np
import skfuzzy as fuzz
x = np.arange(0, 101, 1)
membership = fuzz.trimf(x, [20, 50, 80])   # triangular MF

Troubleshooting

No module named yaml when running quarto render directly: Always use pixi run render, not quarto render directly. The task sets QUARTO_PYTHON to the correct interpreter.

Slow first render: PyMC compiles models with pytensor on first use; subsequent renders are faster.

pixi install fails on Windows for pymc: pymc is available on conda-forge for win-64; if the lock file is causing issues, delete pixi.lock and run pixi install again.

--- title: "Python Setup and Code Reference" --- All code in this book runs inside a `pixi`-managed Python environment. This appendix explains how to set it up and reproduce any example from the text. ## Installing pixi {.unnumbered} pixi is a fast cross-platform package manager. Install it with: ```bash # macOS / Linux curl -fsSL https://pixi.sh/install.sh | bash # Windows (PowerShell) iwr -useb https://pixi.sh/install.ps1 | iex ``` ## Setting Up the Environment {.unnumbered} From the root of the repository: ```bash pixi install # download and install all packages pixi run setup # register the Jupyter kernel ``` This creates a self-contained environment in `.pixi/envs/default/`. It does not touch your system Python or any other environments you have installed. ## Rendering the Book {.unnumbered} ```bash pixi run render # build HTML → docs/ pixi run render-pdf # build PDF → docs/ pixi run preview # live preview in browser (auto-reloads on changes) ``` The `render` task sets `QUARTO_PYTHON` automatically so Quarto uses the pixi environment's Python. ## Key Libraries {.unnumbered} | Library | Purpose | First used | |---|---|---| | `numpy` | Numerical arrays, linear algebra | Ch 1 | | `scipy` | Statistical distributions, optimization | Ch 2 | | `matplotlib` | Static plots | Ch 1 | | `plotly` | Interactive plots (HTML output) | Ch 3 | | `pandas` | Tabular data | Ch 4 | | `scikit-learn` | ML algorithms | Ch 11 | | `pymc` | Bayesian modeling, MCMC | Ch 3 | | `networkx` | Graph structures | Ch 5 | | `scikit-fuzzy` | Fuzzy sets and inference | Ch 9–10 | | `pgmpy` | Probabilistic graphical models | Ch 5 | | `dowhy` | Causal effect estimation | Ch 13 | ## Common Patterns {.unnumbered} **Random number generation (reproducible):** ```python import numpy as np rng = np.random.default_rng(seed=42) # seed for reproducibility x = rng.normal(0, 1, 100) ``` **Plotting (matplotlib):** ```python import matplotlib.pyplot as plt fig, ax = plt.subplots(figsize=(8, 4)) ax.plot(x, y, color="#4e79a7", linewidth=2) ax.set_xlabel("x"); ax.set_ylabel("y") plt.tight_layout() plt.show() ``` **Bayesian model (PyMC):** ```python import pymc as pm with pm.Model() as model: mu = pm.Normal("mu", mu=0, sigma=1) obs = pm.Normal("obs", mu=mu, sigma=0.5, observed=data) trace = pm.sample(1000, tune=500, progressbar=False) ``` **Fuzzy set (scikit-fuzzy):** ```python import numpy as np import skfuzzy as fuzz x = np.arange(0, 101, 1) membership = fuzz.trimf(x, [20, 50, 80]) # triangular MF ``` ## Troubleshooting {.unnumbered} **`No module named yaml` when running `quarto render` directly:** Always use `pixi run render`, not `quarto render` directly. The task sets `QUARTO_PYTHON` to the correct interpreter. **Slow first render:** PyMC compiles models with `pytensor` on first use; subsequent renders are faster. **`pixi install` fails on Windows for `pymc`:** `pymc` is available on conda-forge for `win-64`; if the lock file is causing issues, delete `pixi.lock` and run `pixi install` again.