Automating a Scientific Python Setup: A Gentle Walkthrough

A shell script to set up poliastro and its dependencies

The first in a series of posts on poliastro – a Python library for visualizing and analyzing orbits in Python.
orbital mechanics
scientific computing
poliastro
shell scripts
Author

Madhukara S Putty

Published

June 25, 2025

Introduction

Setting up scientific Python environments used to feel overwhelming to me—until I started collaborating with AI. With its help, I’ve learned how to automate my setup using shell scripts and gained confidence in managing the tools I use for orbital mechanics and data visualization.

In this post, I’ll walk you through a script that sets up a fully functional Python environment to work with Poliastro, a powerful library for orbital mechanics. Whether you’re a beginner or someone looking to streamline your workflow, this guide will give you both a working setup and an understanding of what each step does.

What We’ll Do

By the end of this post, you’ll know how to:

  • Write and run a basic shell script

  • Use conda to manage environments and packages

  • Install the correct versions of poliastro, astropy, and other tools

  • Understand why versions matter in scientific computing

The Full Script (Preview)

Here’s what you’ll build and understand step by step. To get started, copy the code below into a new file and save it as setup_orbitviz.sh. This script will automate the creation of your orbital mechanics Python environment.

#!/bin/bash

echo "🛰️  Creating conda environment 'orbitviz' with Python 3.10..."
conda create -y -n orbitviz python=3.10

echo "🚀 Activating environment (if this fails, activate manually)..."
conda activate orbitviz || source activate orbitviz

echo "🔭 Installing astropy 5.3..."
conda install -y -n orbitviz -c conda-forge astropy=5.3

echo "📈 Installing matplotlib 3.5..."
conda install -y -n orbitviz matplotlib=3.5

echo "🌌 Installing poliastro (with all extras)..."
conda run -n orbitviz pip install poliastro[all]

echo "📓 Installing JupyterLab..."
conda install -y -n orbitviz -c conda-forge jupyterlab

echo "🎨 Installing Plotly..."
conda install -y -n orbitviz -c plotly plotly

echo "🧮 Installing numpy, scipy, tqdm..."
conda install -y -n orbitviz numpy scipy tqdm

echo "✅ Setup complete. To activate the environment later, run: conda activate orbitviz"

Line-by-Line Explanation

  1. #!/bin/bash

This is called a shebang. It tells the operating system: “Run this file using the bash shell.”

Without this line, the system might not know how to run the script.

  1. echo

The echo command prints messages to the terminal. It’s a simple way to give feedback while the script runs:

echo "Creating environment..."

You’ll often use this to keep the user informed of progress.

  1. conda create -y -n orbitviz python=3.10

Let’s break this down:

conda is a package and environment manager, especially useful for scientific work.

create tells it to make a new environment.

-n orbitviz gives the environment a name.

python=3.10 installs Python 3.10 in that environment.

-y means “yes to all prompts” — it won’t ask for confirmation.

Why use environments?

They keep your projects isolated. This way, a change in one project (e.g. upgrading NumPy) won’t break another.

  1. conda activate orbitviz || source activate orbitviz This tries to activate the environment so all future commands run inside it.

conda activate orbitviz: Standard command for modern conda setups.

|| source activate orbitviz: This part runs only if the first one fails (older conda versions).

  1. conda install -n orbitviz -c conda-forge astropy=5.3 This installs a package called astropy from the conda-forge channel:

-n orbitviz: Install into this environment.

-c conda-forge: Use the conda-forge channel — a community-driven source of scientific packages.

astropy=5.3: Install exactly version 5.3.

Why a specific version? Not all versions of Astropy work well with Poliastro. Version pinning helps avoid compatibility issues.

This logic repeats for:

matplotlib=3.5

jupyterlab

plotly

numpy, scipy, and tqdm.

  1. conda run -n orbitviz pip install poliastro[all]

Sometimes, we need pip to install packages from PyPI (Python Package Index).

  • conda run -n orbitviz: Run a command inside the environment.

  • pip install poliastro[all]: Install Poliastro and all its optional features.

  1. chmod +x setup_orbitviz.sh

Before you run the script, you make it executable with:

chmod +x setup_orbitviz.sh

chmod changes file permissions.

+x means “make this file executable.”

Now you can run it as:

./setup_orbitviz.sh

Why Do Versions Matter?

Scientific libraries like astropy, matplotlib, and poliastro evolve constantly. But newer isn’t always better—especially when libraries depend on each other.

For example:

  • Poliastro might rely on internal behavior of Astropy 5.3 that’s changed in 5.4.

  • Matplotlib might break visualization formatting if you jump versions.

Version pinning ensures a stable and predictable environment—which is critical when you’re doing simulations or research.

Want to try?

Here’s how you can run the whole thing:

# Make the script executable
chmod +x setup_orbitviz.sh

# Run it!
./setup_orbitviz.sh

If something fails (e.g. activation), try this manually:

conda activate orbitviz

And then:

jupyterlab

Summary

  • You learned how to write and run a bash script

  • Understood conda environments, channels, and version pinning

  • Set up a working Python lab for orbital mechanics in just a few commands

In the next post, we’ll write our first orbit simulation using Poliastro. You’ll be able to compute orbital elements, propagate orbits, and visualize them in 3D.

AI Attribution: This blog post is part of an ongoing experiment in human–AI collaboration. The shell script presented here was generated by AI based on the author’s request and intent. The initial draft of this article was also AI-assisted, with light human edits and review.