By Ray Donnelly & Crystal Soja
We are pleased to announce that Python 3.7 packages for all supported platforms and packages of the Anaconda Distribution Repository (repo.anaconda.com) are now available. There are 865 packages built for Linux, 864 packages built for macOS, and 779 packages built for Windows.
Python 3.7, released June 27, 2018, represents the combined effort from the core Python developers and includes a long list of PEPs and other features and improvements. The PEPs range from new syntax features (PEP 563) to new library modules (PEP 567 & PEP 557) to new built-in features (PEP 553) to documentation improvements (PEP 545). To review the full list, check out: https://docs.python.org/3/whatsnew/3.7.html
Apart from the numerous performance improvements, we at Anaconda especially are looking forward to using breakpoint() and testing deterministic .pyc files from a “reproducible builds” perspective.
If you have a Python 3 environment with Anaconda installed, you can now easily update it to Python 3.7:
conda install python=3.7 anaconda=custom
You can also create and activate a new conda environment with Python 3.7 and your favorite packages.
conda create -n example_env numpy scipy pandas scikit-learn notebook anaconda-navigator conda activate example_env
Another way to install Python 3.7 is with the newly released Miniconda3 v4.5.11.
Miniconda3 v4.5.11 includes Python 3.7 instead of Python 3.6. The Windows Miniconda installers now check for write permissions before proceeding and no longer allow a comma (,) in the installation path. It also includes conda version 4.5.11, which has over 20 bug fixes and improvements as compared to the previous conda version included in Miniconda v4.5.4.
Stay tuned for a new Python 3.7 release of the Anaconda installers in September!
Note: For open source libraries that do not yet support Python 3.7, a Python 3.7 version will not be available.