Skip to content

SETUP.md

Latest commit

 

History

History
104 lines (70 loc) · 2.38 KB

SETUP.md

File metadata and controls

104 lines (70 loc) · 2.38 KB
Raw

Setup Guide

This project is packaged as pypsa-bc and managed with uv, a fast Python package/project manager. This guide walks through installing uv and setting up the project from a fresh clone.

1. Install uv

macOS / Linux:

curl -LsSf https://astral.sh/uv/install.sh | sh

Windows (PowerShell):

powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

Alternatively, if you already have Python and pip/pipx:

pipx install uv
# or
pip install uv

Verify the install:

uv --version

2. Clone the repository

git clone https://github.com/DeltaE/PyPSA_BC.git
cd PyPSA_BC

3. Set up the environment

uv reads pyproject.toml and .python-version in the repo root. It will automatically download the pinned Python version (3.12) if it isn't already installed, create a .venv, and install all project dependencies plus the pypsa-bc package itself in editable mode:

uv sync

This produces:

  • .venv/ — the project's virtual environment
  • uv.lock — a locked, reproducible set of dependency versions (commit this file)

To include development dependencies (pytest, ipykernel, etc.), uv sync already installs the dev dependency group by default. To skip them:

uv sync --no-dev

4. Run project code

Use uv run to execute anything inside the managed environment without manually activating it:

uv run python run_pypsa_bc.py
uv run snakemake -s workflow/snakefile
uv run jupyter lab

Or activate the virtual environment directly:

source .venv/bin/activate   # macOS/Linux
.venv\Scripts\activate      # Windows

5. Adding or updating dependencies

Add a new runtime dependency:

uv add <package>

Add a development-only dependency:

uv add --dev <package>

Update the lockfile after changing pyproject.toml by hand:

uv lock

Upgrade all dependencies to their latest allowed versions:

uv lock --upgrade

Notes

  • The project previously used a conda environment (env/environment.yaml) and a setup.py. These have been superseded by pyproject.toml + uv.lock. The conda environment file is kept for reference only.
  • Python version is pinned via .python-version (3.12) to match the geospatial/scientific stack (geopandas, rasterio, atlite, etc.) most reliably.