I'm wondering how to install two different versions of a module for python. I host some scripts on a server and some particular ones require older version of the module because some things just doesn't work in the newer version and vice versa, some things doesn't work on the older version.
I know i can use pkg_resources and require a certain version, but if i try to install and use another version it just replaces the other one.
Easiest solution would be to use virtual environments for one or both of your scripts. This is however not recommended in the long run if you want to scale and manage. You should switch up the script using the older version or move to something like a docker based solution.
I'm looking for a virtual environment solution for Python applications and I would like something that respects these requirements:
Windows and Linux
works with x86/x64 Python versions
easy to use/maintain
Python 2.6-2.7 compatible and preferably even 3.x
source control friendly - I want to keep the packages in SCM.
So far I identified virtual-python and zc.buildout, but at least buildout does not supports py3. I'm not looking for a list of solutions, only one you recommend, preferably after you used more than one.
Either virtualenv or zc.buildout will work. Virtualenv is easier to learn and use; buildout is more powerful. I personally use buildout for development/deployment of packages I develop, and virtualenv for deployment of 3rd-party applications (like Trac).
Disclaimer: I've never attempted to use either on Windows, or on Python 3. I see that there's a version of virtualenv that works in Python 3, and there's a branch of zc.buildout that attempts to add Python 3 support (but it seems to be abandoned).
I'm currently toying with python at home and I'm planning to switch to python 3.1. The fact is that I have some scripts that use python 2.6 and I can't convert them since they use some modules that aren't available for python 3.1 atm. So I'm considering installing python 3.1 along with my python 2.6. I only found people on the internet that achieve that by compiling python from the source and use make altinstall instead of the classic make install. Anyway, I think compiling from the source is a bit complicated. I thought running two different versions of a program is easy on Linux (I run fedora 11 for the record). Any hint?
Thanks for reading.
On my Linux system (Ubuntu Jaunty), I have Python 2.5, 2.6 and 3.0 installed, just by installing the binary (deb) packages 'python2.5', 'python2.6' and 'python3.0' using apt-get. Perhaps Fedora packages them and names them as RPMs in a similar way.
I can run the one I need from the command line just by typing e.g. python2.6. So I can also specify the one I want at the top of my script by putting e.g.:
#!/usr/bin/python2.6
Download the python version you want to have as an alternative, untar it, and when you configure it, use --prefix=/my/alt/dir
Cheers
Nik
You're not supposed to need to run them together.
2.6 already has all of the 3.0 features. You can enable those features with from __future__ import statements.
It's much simpler run 2.6 (with some from __future__ import) until everything you need is in 3.x, then switch.
Why do you need to use make install at all? After having done make to compile python 3.x, just move the python folder somewhere, and create a symlink to the python executable in your ~/bin directory. Add that directory to your path if it isn't already, and you'll have a working python development version ready to be used. As long as the symlink itself is not named python (I've named mine py), you'll never experience any clashes.
An added benefit is that if you want to change to a new release of python 3.x, for example if you're following the beta releases, you simply download, compile and replace the folder with the new one.
It's slightly messy, but the messiness is confined to one directory, and I find it much more convenient than thinking about altinstalls and the like.
I bought a low-end MacBook about a month ago and am finally getting around to configuring it for Python. I've done most of my Python work in Windows up until now, and am finding the choices for OS X a little daunting. It looks like there are at least five options to use for Python development:
"Stock" Apple Python
MacPython
Fink
MacPorts
roll-your-own-from-source
I'm still primarily developing for 2.5, so the stock Python is fine from a functionality standpoint. What I want to know is: why should I choose one over the other?
Update:
To clarify, I am looking for a discussion of the various options, not links to the documentation. I've marked this as a Community Wiki question, as I don't feel there is a "correct" answer. Thanks to everyone who has already commented for their insight.
One advantage I see in using the "stock" Python that's included with Mac OS X is that it makes deployment to other Macs a piece of cake. I don't know what your deployment scenario is, but for me this is important. My code has to run on any number of Macs at work, and I try to minimize the amount of work it takes to run my code on all of those systems.
I would highly recommend using MacPorts with Porticus for managing your Python installation. It takes a while to build everything, but the advantage is that whatever you build yourself will be built against the same libraries, so you won't have to futz around with statically linked shared objects, etc. if you want your Python stuff to work with Apache, PostgreSQL, etc.
If you choose to go this way, remember to install the python_select port and use it to make your system use the Python installed from MacPorts.
As an added bonus, MacPorts has packages for most main-stream Python eggs, so if you should be able to have MacPorts keep you up-to-date with the latest versions of all that stuff :)
Here's some helpful info to get you started. http://www.python.org/download/mac/
Depends what you are using python for. If you are using MacOS funitionality and things like PyObjC you are probably best of with MacPython or the python provided by Apple.
I use Python on my Mac mostly for development of server side applications which later will run on FreeBSD & Linux boxes. For that I have used fink python for a few years and ever since MacPorts python. With mac ports it is simple to add required c modules (like database driver etc). It's also easy to keep two python Versions (2.5 & 2.6 in my case) around.
I used "compile your own" python to test pre-3.0 python but generally I find managing dependencies to c modules painfull if done by hand.
Thanks to easy_install installing pure python modules is fast and easy for all the options mentioned above.
I was never very much an IDE person. For development I use command line subversion installed by MacPorts, Textmate and occasionaly Expandrive do directly access files on servers. I personally are very dependent on Bicyclerepairman for Textmade to handle my refactoring needs.
Others seem to be very happy with Eclipse & Pydev.
How about EPD from Enthought? Yes, it's large but it is a framework build and includes things like wxPython, vtk, numpy, scipy, and ipython built-in.
I recommend using Python Virtual environments, especially if you use a Timecapsule because Timecapsule will back everything up, except modules you added to Python!
Based on the number of bugs and omissions people have been encountering in Leopard python (just here on SO!), I couldn't recommend that version. e.g., see:
Why do I get wrong results for hmac in Python but not Perl?
Problems on select module on Python 2.5
I would choose MacPorts.
It does not eliminate your existing python supplied by Apple since it installs by default in /opt/local/bin (plays nice with it) and plus it is easy to download and install additional python modules (even binary modules that you need to compile!). I use Porticus GUI to maintain my MacPorts installed list of packages, including python.
In my windows environment I use Eclipse and PyDev, which works quite well together, even if it's a bit sparse. Apparently the exact same environment is available for the Mac as well, so I suggest downloading Eclipse and using the internal update software function to update PyDev with the URL http://pydev.sourceforge.net/updates/. To look further into PyDev, look here.
Apple's supplied python is quite old – my tiger install has 2.3.5. This may not be a problem for you, but you would be missing out on a lot. Also, there is a risk that Apple will update it. I'm not sure if moving from 2.3.5 to (say) 2.4 would cause code to break, but I guess it's possible. This happened to perl people recently: http://developers.slashdot.org/article.pl?sid=09/02/18/1435227
Macpython is a framework build (as is Apple's, I believe). To be honest, I'm not sure exactly what that means, but it's a prerequisite for some modules, in particular wxPython. If you get python from macports or fink, you will not be able to run wxPython (unless you run it through X11).
And guess what was forgotten by every answer here ... ActivePython.
No compilation required, even for third-party modules such as numpy, lxml, pyqt and thousands of others.
I recommend python (any python?) plus the ipython shell. My most recent experience with MacPython was MacPython 2.5, and I found IDLE frustrating to use as an editor. It's not very featureful, and its' very slow to scroll large quantities of output.
I have a linux VPS that uses an older version of python (2.4.3). This version doesn't include the UUID module, but I need it for a project. My options are to upgrade to python2.6 or find a way to make uuid work with the older version. I am a complete linux newbie. I don't know how to upgrade python safely or how I could get the UUID modules working with the already installed version. What is a better option and how would I go about doing it?
The safest way to upgrading Python is to install it to a different location (away from the default system path).
To do this, download the source of python and do a
./configure --prefix=/opt
(Assuming you want to install it to /opt which is where most install non system dependant stuff to)
The reason why I say this is because some other system libraries may depend on the current version of python.
Another reason is that as you are doing your own custom development, it is much better to have control over what version of the libraries (or interpreters) you are using rather than have a operating system patch break something that was working before. A controlled upgrade is better than having the application break on you all of a sudden.
The UUID module exists as a separate package for Python 2.3 and up:
http://pypi.python.org/pypi/uuid/1.30
So you can either install that in your Python2.4, or install Python2.6. If your distro doesn't have it, then Python is quite simple to compile from source. Look through the requirements to make sure all the libraries you need/want are installed before compiling Python. That's it.
The best solution will be installing python2.6 in the choosen directory - It will you give you access to many great features and better memory handling (infamous python=2.4 memory leak problem).
I have got several pythons installed onto my two computers, I found that the best solution for are two directories:
$HOME/usr-32
$HOME/usr-64
respectively to using operating system (I share $HOME between 32 and 64 bit versions of Linux).
In each I have one directory for every application/program, for example:
ls ~/usr-64/python-2.6.2/
bin include lib share
It leads completetely to avoiding conflicts between version and gives great portability (you can use usb pendrives etc).
Python 2.6.2 in previously example has been installed with option:
./configure --prefix=$HOME/usr-64/python-2.6.2