Elegant use of Python software management, python Management

Elegant use of Python software management, python Management

In the previous article, "elegant environment management with python" (the rich python software package is also one of its advantages, so you do not need to repeat the wheel ). This article describes how to manage Python software packages elegantly.

Common Software Package management tools

Generally, python software package management tools mainly include:

Figure common python package management tools

We can see that distribute is an alternative to setuptools (because the Setuptools package is no longer maintained), and pip is an alternative to easy_install. This article focuses on pip and others interested in it.

Pip's most popular software package management tool

Donald Stufft, one of pip's core developers, wrote a blog last year describing the 4-day traffic of the PyPI CDN supplier. One latitude is displayed. about 75% of the users who use the software package management tool to install the software package from PyPI use pip. This tool has proved to be very useful and popular,In2013Year 10Monthly pipWill be removed from 3.4Starting with pythonBuilt-in release.

Figure: PyPI Installation tool usage Statistical Chart (https://caremad.io/2013/10/a-look-at-pypi-downloads)

Install the recommended software package on the PyPI homepage:

Install

Method 1:

L download get-pip.py files, https://bootstrap.pypa.io/get-pip.py

L python get-pip.py

D: \ Python33> python get-pip.py

Downloading/unpacking pip

Downloading/unpacking setuptools

Installing collected packages: pip, setuptools

Successfully installed pip setuptools

Cleaning up...

Method 2: source code Installation

L download source package: https://pypi.python.org/packages/source/p/pip/

L decompress

L cd to the corresponding directory

L python setup. py install

Use

L install the software package from PyPI:

$ pip install SomePackage
 

  [...]
 

  Successfully installed SomePackage
 

L view the files and paths installed in the software package:

$ pip show --files SomePackage
 

  Name: SomePackage
 

  Version: 1.0
 

  Location: /my/env/lib/pythonx.x/site-packages
 

  Files:
 

   ../somepackage/__init__.py
 

   [...]
 

L check which software packages have updated:

$ pip list --outdated
 

  SomePackage (Current: 1.0 Latest: 2.0)
 

L upgrade software package:

$ pip install --upgrade SomePackage
 

  [...]
 

  Found existing installation: SomePackage 1.0
 

  Uninstalling SomePackage:
 

    Successfully uninstalled SomePackage
 

  Running setup.py install for SomePackage
 

  Successfully installed SomePackage
 

L uninstall the software package:

$ pip uninstall SomePackage
 

  Uninstalling SomePackage:
 

    /my/env/lib/pythonx.x/site-packages/somepackage
 

  Proceed (y/n)? y
 

  Successfully uninstalled SomePackage
 

Pip advanced usage

The previous section describes the basic functions of pip, which have already met most of the requirements. The following describes some advanced usage of pip:

L query software packages

$ pip search "query"
 

L list all installed software packages

$ pip list
 

L specified version number of the installation package

Use =, >=, <=, >,< to specify a version number.

$ pip install SomePackage            # latest version
 

$ pip install SomePackage==1.0.4     # specific version
 

$ pip install 'SomePackage>=1.0.4'     # minimum version
 

L install software packages based on dependent files

Imagine a scenario where "elegant environment management using python" is used to create a virtual environment that feels like it, install some dependent software packages, and develop an application APP. In this case, do you need to deploy it on the server?

In this case, you can use pip to export the dependency file list, and then automatically install the corresponding software package on the server based on the dependency file list. Is it convenient!

Pip freeze> requirements.txt

Pip install-r requirements.txt

Related connections:

Https://pip.pypa.io/en/latest/installing.html

Https://pip.pypa.io/en/latest/quickstart.html