Is it good idea to store python package eggs in ar

问题:

Currently I am developing automated test framework. This test-framework has different packages. These packages will be refer in different project and these may be modified locally by the developer. I want to manage the python package eggs. I am thinking of using Artifactory. I tried to look for Artifactory help for Python,But I couldn't get anything useful.
should I use Artifactory or PIP ?

Edit:
Is there any way or command in python which can help me to put the eggs in artifactory?

回答1:

There are numerous reasons to prefer a binary repository manager over a simple shared directory/SCM binary storage:

1. Fine grained security.
2. Ability to proxy and cache remote repositories.
3. More efficient handling of binaries (because it's a tool that's tailored to do so).
4. Sharing the binaries with other teams and the world is a lot safer and easier.
5. Integration with many tools in the ecosystem.
6. Search and manipulation facilities.
7. Administration tools.

Artifactory exposes a very rich REST API and the deployment of any artifact can be achieved by a simple HTTP PUT request.

回答2:

Take a look at the Defend Against Fruit project. It provides the previously missing glue between Python and Artifactory. http://teamfruit.github.io/defend_against_fruit/

回答3:

You can use "in house" PyPi (either with easy_install -f ... or pip -f ...).

For a server you can have just Apache serving a directory with all the eggs or something like http://pypi.python.org/pypi/pypiserver