Cotainr

**Released on March 28th, 2025**

[Documentation for this release](https://cotainr.readthedocs.org/en/2025.03.0/)

A major release focused on updating project packaging and tooling, as well as adding ARM support.

New features

- Cotainr is now available through pip and all documentation has been changed to reflect this. Additionally, the publishing has been automated through GitHub actions, [PR117](https://github.com/DeiC-HPC/cotainr/pull/117) [PR#118](https://github.com/DeiC-HPC/cotainr/pull/118) [PR#120](https://github.com/DeiC-HPC/cotainr/pull/120) [PR#127](https://github.com/DeiC-HPC/cotainr/pull/127).
- ARM is now officially supported on a smaller test matrix that includes Apptainer but excludes Singularity-CE and Singularity, [PR115](https://github.com/DeiC-HPC/cotainr/pull/115).

Bug fixes

- Ensure files are created in the context of the container to get correct file permissions, [PR80](https://github.com/DeiC-HPC/cotainr/pull/80) [PR#79](https://github.com/DeiC-HPC/cotainr/pull/79).

Maintenance updates

- The Python packaging now follows the pyproject.toml standard, as well as implementing the dependency-groups of PEP735, [PR107](https://github.com/DeiC-HPC/cotainr/pull/107), [PR#57](https://github.com/DeiC-HPC/cotainr/pull/57), [PR#65](https://github.com/DeiC-HPC/cotainr/pull/65), [PR#126](https://github.com/DeiC-HPC/cotainr/pull/126) [PR#112](https://github.com/DeiC-HPC/cotainr/pull/112) [PR#113](https://github.com/DeiC-HPC/cotainr/pull/113).
- Deprecate Python 3.8, [PR119](https://github.com/DeiC-HPC/cotainr/pull/119)
- The project has been updated to primarily use the `uv` tool for dependency resolution, installation and publishing.

**Released on October 21th, 2024**

[Documentation for this release](https://cotainr.readthedocs.org/en/2024.10.0/)

A major release mainly focused on synchronizing with the LUMI 09/2024 upgrade.

New features

- Create an example of using cotainr for building a container with a mpi4py/MPICH conda/pip environment for use on LUMI-C. It includes simple "MPI hello world" examples of how to correctly use such a container on LUMI. [PR50](https://github.com/DeiC-HPC/cotainr/pull/50)
- Cotainr now features the standard `--version` CLI command, [PR59](https://github.com/DeiC-HPC/cotainr/pull/59)

Bug fixes

- Flush the input buffer before prompting the user for permissions, [61](https://github.com/DeiC-HPC/cotainr/pull/61)
- Fix when base image files have restricted permissions, [PR49](https://github.com/DeiC-HPC/cotainr/pull/49)
- Minor documentation fixes, [54](https://github.com/DeiC-HPC/cotainr/pull/54), [#60](https://github.com/DeiC-HPC/cotainr/pull/60), [#69](https://github.com/DeiC-HPC/cotainr/pull/69)

Maintenance updates

- The LUMI PyTorch example has been updated to match the state of LUMI as of 09/2024, [PR73](https://github.com/DeiC-HPC/cotainr/pull/73).
- Update documentation to reflect cotainr residing in the CrayEnv software stack on LUMI, [63](https://github.com/DeiC-HPC/cotainr/pull/63)
- Upgrade the Github Continuous Integration [PR71](https://github.com/DeiC-HPC/cotainr/pull/71), [PR#72](https://github.com/DeiC-HPC/cotainr/pull/72)
- Change from linting from using `black` to `ruff` [PR56](https://github.com/DeiC-HPC/cotainr/pull/56)

**Released on November 3rd, 2023**

[Documentation for this release](https://cotainr.readthedocs.org/en/2023.11.0/)

A major release introducing several new features to significantly improve the user experience.

New features

- Cotainr now features a much improved formatting of console output when running `cotainr build`, including options to set verbosity (`--verbose`/`--quiet`), directing the output to a log file (`--log-to-file`), colored output, filtering of output from Conda/Singularity to fix the console output being messed up by attempted line updates, and finally the introduction of a spinner prepended to the latest output message, when cotainr is doing work in the background, [PR45](https://github.com/DeiC-HPC/cotainr/pull/45).
- When including a conda environment in your container (via `cotainr build ... --conda-env=...`), you are now required to explicitly accept the [Miniforge license terms](https://github.com/conda-forge/miniforge/blob/main/LICENSE) as we use Miniforge to bootstrap the conda environment. This license (along with all other required licenses, if any) may be accepted on beforehand by specifying the option `--accept-licenses` when running `cotainr build ...`, [PR#40](https://github.com/DeiC-HPC/cotainr/pull/40).
- The cotainr release notes now include links to the corresponding HTML documentation hosted on Read the Docs, [PR38](https://github.com/DeiC-HPC/cotainr/pull/38).

Bug fixes

- Cotainr now correctly recognizes and supports the [SingularityCE](https://sylabs.io/singularity/) variant of the Singularity runtime, [PR#42](https://github.com/DeiC-HPC/cotainr/pull/42).
- Minor documentation fixes, [PR34](https://github.com/DeiC-HPC/cotainr/pull/34), [PR#39](https://github.com/DeiC-HPC/cotainr/pull/39), [PR#43](https://github.com/DeiC-HPC/cotainr/pull/43).

Maintenance updates

- Cotainr now implements [PEP394](https://peps.python.org/pep-0394/) to always invoke the python interpreter as "python3", [PR#33](https://github.com/DeiC-HPC/cotainr/pull/33).
- The LUMI PyTorch example has been updated to match the state of LUMI as of 09/2023, [PR41](https://github.com/DeiC-HPC/cotainr/pull/41).
- Minor internal additions/refactorizations/updates, [PR36](https://github.com/DeiC-HPC/cotainr/pull/36).

**Released on February 21st, 2023**

Minor release to fix an issue.

Bug fixes

- Changed shebang in cotainr script to always point to python3

**Released on January 17th, 2023**

Minor release to fix an issue.

Bug fixes

- Fixed a bug with `umask`, which could lead to permission problems with `conda` installation.

**Released on December 21st, 2022**

First public release.

New features

- `cotainr info` command for showing system info
- `cotainr build` command for building containers
- HPC systems integration via `system.json` and `--system` CLI flag
- Conda environment use case support via `--conda-env` CLI flag
- Documentation hosted on https://cotainr.readthedocs.io
- LUMI ROCm PyTorch cotainr build example