James Leslie James-Leslie

Dev Environment Setup

My baseline setup for code-first projects: mise manages language runtimes (Python, Node) and CLI tools (prek, just), uv manages Python packages and the virtualenv, pnpm manages Node packages. Tool-per-concern, no overlap.

Tool	Purpose
git	Version control, and the source of Git Bash (the shell used here)
scoop	Windows package manager (installs mise, gh)

Example conda environment file

Conda documentation

1. File structure

name: <environment-name>
channels:
 - conda-forge

How to add conda environments to JupyterLab

Credit for this goes to the Stack Overflow user Statistic Dean for his post on this question

1. Install `nb_conda_kernels` into your base environment

(base)$ conda install -c conda-forge nb_conda_kernels

2. Activate desired environment and install `ipykernel`

Managing conda virtual environments

Conda documentation

1. Getting started

1.1. Create env

Create env with python 3.7 and pip

conda create --name whatwhale python=3.7 pip

	[project]
	name = "example-project"
	version = "0.1.0"
	description = "Add your description here"
	readme = "README.md"
	requires-python = ">=3.11"
	dependencies = [
	]

	[dependency-groups]

	# kimi alias for Claude Code via Vercel AI Gateway
	# Requires $AI_GATEWAY_API_KEY to be set in your environment

	alias kimi='ANTHROPIC_BASE_URL=https://ai-gateway.vercel.sh ANTHROPIC_AUTH_TOKEN="$AI_GATEWAY_API_KEY" ANTHROPIC_API_KEY="" ANTHROPIC_MODEL="moonshotai/kimi-k2.6" claude'