Python Scientific Computing Environment recommendation--anaconda

Anaconda is a scientific computing environment similar to canopy, but it is more convenient to use. The Package Manager Conda is also very powerful.

The first is to download the installation. Anaconda provides two versions of Python2.7 and Python3.4, and can be created by Conda if additional versions are required. When the installation is complete, you can see that Anaconda provides Spyder,ipython and a command line. Here's a look at Conda.

Enter Conda list to see all the Python extensions that came with the installation. A cursory look, which includes commonly used Numpy, Scipy, Matplotlib and Networkx, as well as beautiful-soup, requests, flask, Tornado and other network-related extensions.

Strangely enough, there is no sklearn in it, so first install it.

Conda Install Scikit-learn
If you need to specify a version, you can also specify it directly using [package-name]=x.x].

Conda repo in the extension is not too new, if you want to update, you may want to use PyPI or download the source code yourself. Conda and Pip are well-correlated. Using PIP to install something can be managed using Conda, which is better than canopy. I used pip to install the NLTK, Jieba and Gensim.

Another requirement I have for this scientific computing environment is the ability to coexist with multiple Python versions, especially the coexistence of 2.x and 3.x. This can be done by virtualenv. Anaconda is also achieved through it.

The following is a Conda to create a python2 version of the python2.7 environment.

Conda create-n Python2 python=2.7
This will create the Python2 directory under the Envs directory under the Anaconda installation directory.

Installing the extension to it can:

Use the Conda install directly with the-N to indicate the installed environment, which is naturally python2.
Like Virtualenv, first activate and then installed in the virtual environment.
There is a sudden problem here, how to use the created environment in the IDE? If the IDE is Pycharm, set up the Python installation directory directly. What about the Spyder? In fact, the Spyder is a python extension, and you need to install a Spyder in a virtual environment.

The final is the Spyder main interface. Variable viewing window is very useful.

