Installation¶
cellpy is available on Windows, macOS, and Linux. Prefer conda when you want the scientific stack and awkward native deps handled for you; use pip when you want a lean install and can manage system packages yourself.
After installing, continue to Setup and configuration and Check your installation.
Install by platform¶
Conda
This pulls in the critical dependencies (and Jupyter, which is handy for the tutorials). Install into a virtual environment if you can.
New to Python? Teaspoon path
- Install Miniconda or Anaconda (64-bit, Python 3.11+).
- Open Anaconda Prompt and create an environment:
- Install cellpy:
Bitness matters for Arbin .res (Access) drivers: match the driver to
your Python build (32- vs 64-bit).
Pip
On Windows, packages such as tables (HDF5) can be awkward with pip. Prefer
conda, or create an env from the repo
environment.yml
first.
Arbin .res files
.res files are Microsoft Access databases. You need an Access /
ACE ODBC driver that matches your Python bitness
(Microsoft download).
If loading fails, try:
System packages
For Arbin .res support, cellpy uses mdbtools to export to temporary CSV
(there is no Access ODBC path like on Windows). Also install HDF5 libs if you
plan to build tables with pip:
Apple Silicon (Homebrew prefixes) — if tables fails to build under pip:
export HDF5_DIR=/opt/homebrew/opt/hdf5
export BLOSC_DIR=/opt/homebrew/opt/c-blosc
python -m pip install cython tables
Conda
Pip
Use a virtual environment (python -m venv .venv or conda env).
System packages
For Arbin .res support, cellpy uses mdbtools to export to temporary CSV
(there is no Access ODBC path like on Windows). Ubuntu/Debian example, plus
common pip build deps:
Other distros: install the equivalent packages with your package manager.
Conda
Pip
Prefer a virtual environment. Without sudo, try the same apt packages in a
user/container image, or stick to conda.
Installation from sources¶
Clone the public repository:
Recommended for contributors (see CONTRIBUTING.md):
Or with conda + editable install:
Dependencies¶
Conda and environment.yml / the project lockfile (uv.lock, driven by
pyproject.toml) install what you need for normal use. Highlights:
| Need | Packages / notes |
|---|---|
| Core science stack | numpy, scipy, pandas |
| cellpy files (HDF5 path) | tables (PyTables) |
| Fitting helpers | lmfit |
| Templates | jinja2-time, and git on PATH |
| Tutorials / notebooks | jupyter, seaborn, plotly |
Optional extras and plotting backends evolve with the release — prefer the conda-forge package or the repo env file over hand-picking versions.