We are happy to announce that Anaconda 4.3 has been released. Anaconda is the leading open data science platform powered by Python.
Python 3.6.0 was released on December 23, 2016, and on the same day, we made conda binaries available for all major platforms. However, that did not mean that all Python packages had been created and tested against Python 3.6 on the same day. We are now happy to have full Python 3.6 support, which means that the Anaconda3 installers are based on Python 3.6. In all, Anaconda 4.3 supports Python 2.7, 3.4, 3.5 and 3.6. Anaconda 4.3 will also be the last release which supports Python 3.4. We will discontinue regular Python 3.4 package updates in the public Anaconda.
Long term support for older versions of Python is available by purchasing an Anaconda Subscription. Please contact sales@continuum.io with your needs for more information.
Important updates:
The Intel Math Kernel Library (MKL) has been updated to 2017.0.1 (from 11.3.3)
Over all, 90 package have been updated (within the Anaconda installer), see complete changelog.
Other changes:
In order to increase compatibility with the conda-forge community, we aligned the
jpegandlibpngversions Anaconda uses.We added a warning if a user is attempting to install on Windows into an install path with spaces, and we will not allow the installation to proceed if Unicode characters are in the in the install path.
We fixed a lot of the Windows menu uninstallation issues and some other often reported uninstallation issues on Windows.
Going forward Mac OS 10.9 will be the oldest supported version of Mac OS (dropping free support for 10.7 and 10.8) as of Anaconda 4.3. This means that 4.2 will be the last version that supports 10.7 or 10.8.
We added
seabornto the Anaconda installer.We removed
conda-buildandanaconda-clean, as well as the Jupyter Notebook extensions from the installers. All these packages are still available on the default conda repository, and can easily be installed usingconda.