Installation

uv is recommended for installation:

uv pip install fiqci-ems
# or
uv add fiqci-ems

Requires Python 3.11 or 3.12.

Verifying the installation

To test your installation clone the repository and run the tests against the installed package.

git clone https://github.com/fiqci/fiqci-ems.git
cd fiqci-ems
uv run --group dev pytest

HPC systems (Apptainer / Singularity)

On shared HPC systems it is recommended to install Python environments inside a container using Apptainer/Singularity. The definition file below builds on the official uv image, so uv is already available in the container.

Save it as fiqci-ems.def:

Bootstrap: docker
From: ghcr.io/astral-sh/uv:python3.12-bookworm-slim

%post
    uv pip install --system --no-cache fiqci-ems

And to build the image

apptainer build --fakeroot fiqci-ems.sif fiqci-ems.def
apptainer exec fiqci-ems.sif python3 my_script.py

LUMI

On LUMI fiqci-ems comes pre-installed in the fiqci-vtt-qiskit module which can be used via

module load Local-quantum
module load fiqci-vtt-qiskit

Otherwise you can use it in your own containers. It is highly recommended to use one of the two container-based workflows below. See the LUMI documentation on installing Python, the container wrapper and Singularity/Apptainer for background.

Define the environment once in an env.yml file.

channels:
  - conda-forge
dependencies:
  - python=3.12
  - pip
  - pip:
      - fiqci-ems
      - pytest

Option 1: LUMI container wrapper (recommended)

The lumi-container-wrapper builds a container behind the scenes but exposes the executables (python, pip, …) on your PATH as if the environment were activated normally. This is the simplest option when you just want to run Python scripts.

module load LUMI
module load lumi-container-wrapper

mkdir -p $PWD/fiqci-ems
conda-containerize new --prefix $PWD/fiqci-ems env.yml

export PATH="$PWD/fiqci-ems/bin:$PATH"

To add or change packages later, use a post-install script:

conda-containerize update $HOME/fiqci-ems --post-install post.sh

where post.sh contains commands run inside the environment, e.g. pip install <package>.

Verify the installation on a compute node via srun, from a clone of the repository so the tests directory is available:

git clone https://github.com/fiqci/fiqci-ems.git
cd fiqci-ems
srun --account project_xxx --partition interactive python3 -m pytest tests/

Option 2: Build a SIF image with cotainr

If you prefer an explicit, portable .sif image (for example to share it or manage container execution yourself), build it with cotainr:

module load CrayEnv
module load cotainr

cotainr build fiqci-ems.sif --system=lumi-c --conda-env=env.yml

Verify the installation on a compute node via srun.

git clone https://github.com/fiqci/fiqci-ems.git
srun --account=<project> --partition=interactive \
    singularity exec -B "$PWD:/mnt" fiqci-ems.sif \
    python3 -m pytest /mnt/fiqci-ems/tests -p no:cacheprovider