The current popular Python scientific computing distribution

Often have the side of the alumni asked to use what Python distribution is better?

In fact, the current more popular Python scientific computing distribution, mainly have so few:

Python (x, y)

Based on PYQT, the GUI was once the most versatile and powerful, and is the perfect choice for the science free python distribution in Windows. But it's different now! Many of the packages inside the Pythonxy are not able to use the latest packages for compatibility issues. In particular, the MinGW is still an antique-grade version 4.5, and TDM-GCC is now 4.8.1-3. However, this package after you installed, in addition to occupy a larger disk space, basically you do not have to pay what God to find, for the basic scientific calculation has: Numpy,sicpy,matplotlib,spyder ... The current version is 2.7.6.0. From installation to use, play "Crash" from time to time! The stability of the version is far less than that of 2.7.3.1. I think it may be the author of the main focus on the development and maintenance of winpython above the reason it!

Winpython

Winpython features are also relatively complete, the software package is relatively new, the GUI is based on PYQT, but relative to Python (x, y), it focuses on the portable installation experience: You can put it in the U disk. I am now no longer using PYTHONXY in Windows, changed to Winpython, for MinGW, I directly installed QT5.20 (including MinGW4.8.0), plus boost 1.54.0, for scientific research, has not been visual Studio can be the same! The current version is 2.7.6.3. Stability has improved considerably. The published version is also relatively full:

[Plain]View Plaincopy

1. Winpython-32bit-2.7.6.3.exe
2. Winpython-32bit-3.3.2.3.exe
3. Winpython-64bit-2.7.6.3.exe
4. Winpython-64bit-3.3.2.3.exe

If you're still using PYTHONXY, consider changing to Winpython or the anaconda below.

Anaconda

This is a new show, it has been updated many times this year, the current version is 1.9.1. Package management using Conda,gui based on pyside, all packages are basically the latest version, no PYQT and Wxpython, etc., the capacity is moderate, but the scientific calculation package has: Numpy,sicpy,matplotlib,spyder ... This is the current installation of my CentOS system server. Linux system, Anaconda installation, update and delete are very convenient, and all things are installed in a directory only/home/wxp/anaconda/, This is much better than the canopy below. Anaconda's development and maintenance is a key member of the Python founders and community, and you can imagine that this package will definitely "come from behind"! Anaconda currently offers Python 2.6.9,python 2.7.X and Python 3.3.X three series packages, which is a legacy of other distributions. So in various operating systems, whether it's Linux, Windows, or Mac, I highly recommend anaconda!

It is convenient to upgrade and install in Anaconda, but instead of providing a graphical environment like Winpython, use the command line:

[Plain]View Plaincopy

1. Conda List #查看所有的可安装包
2. Conda Install Wxpython #安装
3. Conda Install PYQT #安装
4. Conda Update Ipython #升级

Enthought Canopy (enthought Python distribution)

The current version of Enthought is 1.3.0. GUI based Wxpython, including Pyside, but excluding Pyqt.wxpython is more convenient to use, but far from PyQt and pyside popular, need to use PYQT can be installed on their own. Canopy has its own integrated development environment (IDE), the inside of the Code smart hints and auto-complete function is not worse than Ipython! Canopy also integrates the online upgrade and management system of the Pyhton package, which is very convenient. Due to its commercial level, canopy's performance and stability are strong! Free version and academic version are also available (for educational research is also free). Formerly called EPD, now changed to name Canopy. Canopy is the first release to upgrade Ipython to 1.1.0, Matplotlib has been upgraded to 1.3.1;numpy 1.8.0;scipy 1.2.0. But it is mainly the pursuit of performance and stability, so can not expect all the installation package is up to date, for example, Mingw,canopy is 4.8.1, the other version of the release can be 4.7 Ah!basemap official has been replaced by 1.0.7, This is a boon for users who are still forced to use grads and NCL! If you have a school mailbox, you can register on Enthought's website and choose the distribution of academic +full, which will make your work more powerful.

Sage

The focus of the release, the goal is to kill: MATLAB, MAPLE, Matcad, Mathematica and so on. At present I am also in the study.

Ps:

Due to the famous matplotlib (Stable 1.3.1), IPython (Stable 1.2.1), Numpy (Stable 1.8.1) and scipy (0.13.3) have released the latest stable version, not a few months, The above distribution will usher in a larger update, let us wait and see!

There is also a website for Windows customers to install the extension library, provide *.exe release for download, click the link below to

Unofficial Windows Binaries for Python Extension Packages

The current popular Python scientific computing distribution