conda_setup: The Ultimate Conda Environment Manager

Overview

If you use Conda for Python environments (especially for data science, ML, or scientific computing), you know the commands can get verbose. I built conda_setup to streamline everything into one powerful CLI tool.

Download: conda_setup.py on GitHub

---

Why Conda?

Virtual environments (venv) are great for pure Python projects. But when you need:

• NumPy, SciPy, TensorFlow with optimized binaries
• Non-Python dependencies (like CUDA, MKL, OpenBLAS)
• Reproducible environments across machines

…Conda is the answer. But the commands are lengthy:

# The old way
conda create -n myenv python=3.11
conda activate myenv
conda install numpy pandas scikit-learn
conda env export > environment.yml
conda env list
conda env remove -n oldenv

With conda_setup, it’s all in one place with an interactive menu AND CLI shortcuts.

---

Features

Interactive Menu (20 Options!)

Run python3 conda_setup.py for the full menu:

Option	Feature
1	Create new environment
2	Activate environment
3	Delete environment
4	List all environments
5	Clone environment
6	Export to environment.yml
7	Import from environment.yml
8	Install package
9	Uninstall package
10	Update all packages
11	Search packages (conda-forge)
12	Show environment info
13	Auto-detect current environment
14	Show installed packages
15	Clean cache
16	Check conda health
17	Update conda itself
18	Show environment stats
19	Settings
20	Exit

CLI Quick Commands

# List all environments
python3 conda_setup.py --list

# Auto-detect active environment
python3 conda_setup.py --detect

# Show environment info
python3 conda_setup.py --info myenv

# Search for a package
python3 conda_setup.py --search pytorch

# Show packages in environment
python3 conda_setup.py --packages myenv

# Full help with all options
python3 conda_setup.py --help

Auto-Install Dependencies

When switching environments, the script checks for:

• environment.yml
• requirements.txt

And offers to install them automatically:

Switching to environment 'ml-project'...

Found dependency files:
  - environment.yml (45 packages)
  - requirements.txt (12 pip packages)

Install dependencies now? (y/n):

---

My Mistakes Building This

Mistake #1: The External Drive Shebang

Same issue as venv_setup - I hardcoded my external drive path:

#!/Volumes/KaliPro/Applications/miniconda3/envs/ranger/bin/python

When the drive wasn’t mounted? Script wouldn’t run at all.

The fix:

#!/usr/bin/env python3

Mistake #2: Duplicate Function Definitions

I had parse_args() defined TWICE in the file. Python just used the second one, but it caused confusion and the first definition’s features were lost.

Lesson: Always search your file for duplicate function names before adding new ones!

Mistake #3: Sudo for Everything

The original script asked for sudo password on EVERY run, even for:

• --list (just reads conda info)
• --help (just shows help text)
• --search (just queries conda-forge)

The fix:

def needs_sudo():
    read_only_args = ['--list', '--help', '-h', '--no-color',
                      '--info', '--detect', '--packages', '--search', '--bunny']
    for arg in sys.argv[1:]:
        if arg in read_only_args:
            return False
    return True

Now read-only commands are instant - no password prompt.

Mistake #4: Version Number Mismatch

The script said “v5.3” at the top but “v5.4” in the menu header. Small thing, but unprofessional.

Lesson: Keep version numbers in ONE place, or use a constant:

VERSION = "5.6"

---

Installation

# Download the script
curl -O https://raw.githubusercontent.com/davidtkeane/Rangers-Scripts/main/Programming-Scripts/System-Scripts/Environment-Managers/conda_setup.py

# Make executable
chmod +x conda_setup.py

# Optional: Create alias
echo 'alias cenv="python3 ~/bin/conda_setup.py"' >> ~/.zshrc
source ~/.zshrc

# Now use anywhere
cenv --list
cenv --detect
cenv --search numpy

Requirements

• Python 3.6+
• Anaconda or Miniconda installed
• colorama (pip install colorama)

---

Pro Tips

Quick Environment Switching

# See what's available
cenv --list

# Check current environment
cenv --detect

# Get detailed info
cenv --info ml-project

Package Discovery

# Search conda-forge for packages
cenv --search tensorflow

# See what's installed
cenv --packages myenv

Environment Backup

The export feature creates a complete environment.yml:

name: ml-project
channels:
  - conda-forge
  - defaults
dependencies:
  - python=3.11
  - numpy=1.24.0
  - pandas=2.0.0
  - scikit-learn=1.3.0
  - pip:
    - custom-package==1.0.0

Recreate on any machine with:

conda env create -f environment.yml

---

Easter Egg

Try python3 conda_setup.py --bunny for a hopping good time!

---

What I Learned

1. Shebang portability is crucial - External drive paths break everything
2. Check for duplicate functions - Easy to miss, hard to debug
3. Read-only commands shouldn’t need privileges - UX matters
4. Version numbers need a single source of truth
5. Auto-detection saves mental overhead - Users shouldn’t have to remember state

---

Download

GitHub: conda_setup.py

Version: 5.6 Enhanced Edition

---

Rangers lead the way!