There are several ways to install Python and its packages, but here's how to install them using Anaconda.
I often see discussions around me about the pros and cons of building an environment with Anaconda and using the Python community standard method. You can choose the one that suits your purpose, but it can be difficult for beginners to determine which one suits their purpose.
Below is a summary of my personal views on the value that Anaconda (Continuum Analytics) offers as a distributor.
Meanwhile, wheel, a tool that provides the official binary distribution for the Python community, has also been enhanced.
I hope it will give you an opportunity to think about something that suits your purpose.
You can download Anaconda from the following sites.
You need to select the Python version when installing. There are 2.7 and 3.x versions of Python.
In conclusion, unless you have a specific reason, you can choose the latest version of the new version, 3.x (currently 3.7).
Python is incompatible with 2.x and 3.x, 2.x is in maintenance mode, and the Python community has PEP 373. -0373 /) will be supported until 2020.
Environments that have been using 2.x for some time still have a support period, but no new features will be provided to 2.x in the future. Therefore, if you are new to Python or are developing new ones, you should not hesitate to use the latest version of 3.x.
Installers are available for each OS for Windows, OS X, and Linux. Select and install the installer for the platform you are using.
You can download the installer for Windows from the following.
Only a graphical installer is available for Windows. Here, download the 64-bit version and proceed with the installation work.
Run * Anaconda3-5.3.0-Windows-x86_64.exe * to launch the installer.
Install with administrator privileges. If you are not particular about it, you can install it by default as it is. If you install with the default settings, it will be installed in the following location.
C:\Anaconda3
Start a command prompt and run python to see the Python interpreter installed in Anaconda launch as shown below.
C:\Users\user>python
Python 3.7.0 (default, Jun 28 2018, 08:04:48) [MSC v.1912 64bit (AMD64)] :: Anaconda, Inc. on win32
Type "help", "copyright", "credits" or "license" for more information.
>>>
Let's try importing numpy. You can see that the numpy in the Anaconda environment has been imported.
>>> import numpy
>>> numpy.__file__
'C:\\Anaconda3\\lib\\site-packages\\numpy\\__init__.py'
You can download the installer for OS X from:
Both graphical and command line installers are available for the OS. Here, download the graphical version and proceed with the installation work.
Run * Anaconda3-5.3.0-MacOSX-x86_64.pkg * to launch the installer.
If you are not particular about it, you can install it by default as it is. If you install it for your personal environment, it will be installed in the following location.
$ ls ~/anaconda/
Navigator.app bin conda-meta envs etc include lib pkgs python.app share ssl
Run python and check that the Python interpreter installed by Anaconda starts as follows.
$ python
Python 3.7.0 (default, Jun 28 2018, 07:39:16)
[Clang 4.0.1 (tags/RELEASE_401/final)] :: Anaconda, Inc. on darwin
Type "help", "copyright", "credits" or "license" for more information.
Let's try importing numpy. You can see that the numpy in the Anaconda environment has been imported.
>>> import numpy
>>> numpy.__file__
'/anaconda3/lib/python3.7/site-packages/numpy/__init__.py'
By default, the Anaconda installer sets the environment variable * PATH *. If you are concerned about conflicts between commands installed on your system and commands with the same name installed by Anaconda, you may want to disable this setting and set * PATH * only when using Anaconda.
$ vi ~/.bash_profile
# added by Anaconda3 5.3.0 installer
...
export PATH="/anaconda3/bin:$PATH"
You can download the installer for Linux from the following.
Only the command line installer is available for Linux. Here, download the 64-bit version and proceed with the installation work.
$ bash Anaconda3-5.3.0-Linux-x86_64.sh
...
Do you wish the installer to prepend the Anaconda3 install location
to PATH in your /home/vagrant/.bashrc ? [yes|no]
[no] >>>
You will be asked to confirm the license agreement and the installation location interactively, but if you are not particular about it, you can just install it by default. Finally, you will be asked if you want to include the Anaconda-installed executable in the environment variable * PATH * to preferentially use it. The default is * no *.
The pros and cons of changing * PATH * by default are controversial. For example, Anaconda installs the curl command, but it hides the command installed on your system.
$ which curl
/usr/bin/curl
$ export PATH=/path/to/anaconda3/bin:$PATH
$ which curl
~/anaconda3/bin/curl
If you are concerned about conflicts between other commands installed on your system and those installed by Anaconda, do not set * PATH * by default, and only temporarily when using Python (and its environment) installed by Anaconda. Make sure to set * PATH * on your device.
$ export PATH=/path/to/anaconda3/bin:$PATH
If you install with the default settings, it will be installed in the following location.
$ ls ~/anaconda3/
LICENSE.txt conda-meta etc libexec phrasebooks qml
ssl vscode_inst.py.log bin doc include man pkgs resources
translations x86_64-conda_cos6-linux-gnu compiler_compat envs
lib mkspecs plugins share var
After setting * PATH *, try starting Python.
$ python
Python 3.7.0 (default, Jun 28 2018, 13:15:42)
[GCC 7.2.0] :: Anaconda, Inc. on linux
Type "help", "copyright", "credits" or "license" for more information.
Let's try importing numpy. You can see that the numpy in the Anaconda environment has been imported.
>>> import numpy
>>> numpy.__file__
'/path/to/anaconda3/lib/python3.7/site-packages/numpy/__init__.py'
A Python distribution that includes Python provided by Continuum Analytics and a package manager called conda. The site states that it includes R and Scala packages in addition to Python, but you can think of them as mainly Python packages.
Anaconda seems to want to provide a platform dedicated to data science. That's why it features a set of packages that are commonly used in Python for those areas.
In general, in the world of Python, tools and libraries that analyze large amounts of data are important for processing performance, so extension modules developed in languages such as C / C ++ ([Extending and embedding the Python interpreter]([Python interpreter extension and embedding]) There are many packages that include http://docs.python.jp/3.5/extending/extending.html)). To install these packages, you need to compile and install for the OS you are using in your local environment, but Anaconda compiles in your local environment by providing a pre-compiled package (binary distribution). You can install it without any hassle.
It's not difficult for the average developer to build an environment for compiling Python C extensions. However, the task of creating such an environment can be painful for researchers and data scientists. A binary distribution called Anaconda is popular because it eliminates the complexity of building such a development environment.
A tool called pip is provided as a standard Python package manager, but if you are familiar with the Anaconda environment, [conda](http: //: //) You can also use conda.pydata.org) as your package manager.
Anaconda is a distribution that includes a set of packages for data science, while you can also install only Python and conda with a minimal configuration. The minimum configuration distribution is called Miniconda.
Miniconda can be downloaded and installed from:
Let's use * conda * as a package manager.
$ conda list
# packages in environment at /path/to/anaconda3:
#
_license 1.1 py36_1
alabaster 0.7.9 py36_0
anaconda 4.3.0 np111py36_0
anaconda-client 1.6.0 py36_0
anaconda-navigator 1.4.3 py36_0
...
$ conda search django
Using Anaconda Cloud api site https://api.anaconda.org
Fetching package metadata .......
django 1.6.5 py26_0 defaults
1.6.5 py27_0 defaults
1.6.5 py33_0 defaults
1.6.5 py34_0 defaults
...
$ conda install django
Using Anaconda Cloud api site https://api.anaconda.org
Fetching package metadata .......
Solving package specifications: ..........
Package plan for installation in environment /path/to/anaconda3:
The following NEW packages will be INSTALLED:
django: 1.10.5-py36_0
Proceed ([y]/n)? y
$ conda update lxml
Using Anaconda Cloud api site https://api.anaconda.org
Fetching package metadata .......
Solving package specifications: ..........
Package plan for installation in environment /path/to/anaconda3:
The following packages will be UPDATED:
lxml: 3.7.1-py36_0 --> 3.7.2-py36_0
Proceed ([y]/n)? y
$ conda remove django
Fetching package metadata .......
Using Anaconda Cloud api site https://api.anaconda.org
Solving package specifications: ..........
Package plan for package removal in environment /path/to/anaconda3:
The following packages will be REMOVED:
django: 1.10.5-py36_0
Proceed ([y]/n)? y
Thanks to @python_ufo and @y__sama for their comments. Thank you very much.
As a function of conda, in addition to being able to build a virtual environment for Python, it seems that it can handle switching between multiple versions of Python. If you don't use Anaconda, use virtualenv to build a virtual environment, and pyenv to switch between multiple versions of Python. I think that com / yyuu / pyenv) is often used. It seems that it can be used as an alternative to these with the function of conda.
I am building a virtual environment called * py27 * that uses Python 2.7.
$ conda create -n py27 python=2.7 anaconda
You can check the environment you are currently using as follows.
$ conda info -e
# conda environments:
#
py27 /path/to/anaconda3/envs/py27
root * /path/to/anaconda3
To switch to the * py27 * environment on Linux or OS X, use the shell's built-in command * source * and run: Windows does not have a * source * command, so execute the * activate * command directly.
$ source activate py27 # $ activate py27 (on windows)
(py27) $
Make sure that the environment is switched with the conda command.
(py27) $ conda info -e
# conda environments:
#
py27 * /path/to/anaconda3/envs/py27
root /path/to/anaconda3
(py27) $ python
Python 2.7.13 |Anaconda 4.3.0 (64-bit)| (default, Dec 20 2016, 23:09:15)
[GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
Anaconda is brought to you by Continuum Analytics.
Please check out: http://continuum.io/thanks and https://anaconda.org
>>>
To exit the * py27 * environment, execute as follows.
(py27) $ source deactivate # $ deactivate (on windows)
$ python
Python 3.6.0 |Anaconda custom (64-bit)| (default, Dec 23 2016, 12:22:00)
[GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>>
For details on how to use conda, refer to the official document Using conda.
A repository is provided to manage the package called Anaconda Cloud. It seems that you can also upload arbitrary packages to the personal area of this repository and manage packages for personal use. It's free for public project plans and $ 7 / month for private plans.
Anaconda-client (https://github.com/Anaconda-Platform/anaconda-client) is provided as a command line interface for interacting with Anaconda Cloud. You can work with Anaconda Cloud using the command * anaconda * by installing anaconda-client (if you installed the Anaconda distribution, you also have anaconda-client installed).
Here, the package LittleHTTPServer is used as an example.
Searching for packages not provided by conda will not show anything.
$ conda search littlehttpserver
Using Anaconda Cloud api site https://api.anaconda.org
Fetching package metadata .......
Try searching the repository on Anaconda Cloud using anaconda-client. It seems that you can specify either * conda * or * pypi * package type with * -t (--package-type) * option. Here, specify * conda *.
$ anaconda search -t conda littlehttpserver
Using Anaconda API: https://api.anaconda.org
Run 'anaconda show <USER/PACKAGE>' to get more details:
Packages:
Name | Version | Package Types | Platforms
------------------------- | ------ | --------------- | ---------------
auto/littlehttpserver | 0.1.4 | conda | linux-64
: http://bitbucket.org/t2y/littlehttpserver
t2y/LittleHTTPServer | 0.5.0 | pypi, conda | linux-64
: Little bit extended SimpleHTTPServer
Found 2 packages
I found two packages of littlehttpserver that someone has published. Select 0.5.0, which is the latest version at the moment. Here, select the package published by the user * t2y *.
Execute the installation command with the * -c (--channel) * option specified as follows.
$ conda install -c https://conda.anaconda.org/t2y littlehttpserver
Fetching package metadata ...........
Solving package specifications: .
Package plan for installation in environment /path/to/anaconda3:
The following NEW packages will be INSTALLED:
littlehttpserver: 0.5.0-py36_0 t2y
Proceed ([y]/n)? y
You can see that littlehttpserver keras is installed from the package list.
$ conda list | grep littlehttpserver
littlehttpserver 0.5.0 py36_0 t2y
$ littlehttpserver --help
usage: littlehttpserver [-h] [-d DOCUMENT_DIR] [-i INDEX_DIRECTORY]
[-p PORT_NUMBER] [-v] [--protocol PROTOCOL]
[--servertype {process,thread}] [--version]
optional arguments:
-h, --help show this help message and exit
-d DOCUMENT_DIR, --dir DOCUMENT_DIR
set some document directories
-i INDEX_DIRECTORY, --indexdir INDEX_DIRECTORY
set arbitrary top directory
-p PORT_NUMBER, --port PORT_NUMBER
set server port number
-v, --verbose set verbose mode
--protocol PROTOCOL set protocol
--servertype {process,thread}
set server type
--version show program version
For details on how to use anaconda-client, refer to the official document Anaconda Cloud.
Recommended Posts