Releases: jooh/neuroconda
Releases · jooh/neuroconda
Neuroconda 2.0 - shell wrappers, more sustainable development model
- New development model where we separate key package requirements (neuroconda_basepackages.yml) from the final neuroconda.yml, which also includes any dependencies of these packages. This makes maintaining and updating a lot easier and hopefully more sustainable (e.g., packages that are no longer a dependency for any base package will get pruned in the next update).
- Automated update and install with GNU make. This was a lot of work, but it should make it easier to push out minor releases frequently in the future.
- Introduce CBU-specific shell wrappers for adding non-conda packages to path (these can be used as templates for your local setup). This is necessary because the env_vars.{sh,csh} stuff in conda remains quite broken (see README for more on this). Maybe one day it will get fixed and we can get back to a conda-only setup.
- Thanks to above improvements in the development model, neuroconda packages are now bleeding edge. We include the latest available release of Altair, Nibabel, Nipype, Nilearn, etc etc... The only version left behind is python, which is still 3.7.3.
Neuroconda 1.4: fsleyes, theano, bids-validator
- Various package updates (Tensorflow most notably) to take care of nasty numpy incompatibility warnings
- Managed to move a few pips to conda (bleach, prov)
Neuroconda v1.3 - Project rename, heudiconv, dipy, transplant
- New packages: Heudiconv, Dipy, Transplant (matlab interface)
- Rename project to better represent scope (ie, not just python)
- Update all packages to the latest available conda-forge versions
- Pin Github-hosted pip dependencies for better reproducibility
- More work to ensure no quiet additional pip installs during build
R, datalad extensions
v1.2 Merge branch 'master' of github.com:MRC-CBU/cbu_nipy_env
First release
This is the initial release of the CBU Nipy environment. On the CBU imaging system, you can access it by doing
conda activate cbu_nipy_1_00