Develop Reference

How to use Python packages that were installed on Python 3.5 in Python 3.7? [duplicate]

I have upraded my Linux distro recently. Python 3.5 was replaced by Python 3.6.
All site packages I have installed with pip3 are still in the /usr/lib/python3.5/site-packages directory and Python does not find them there now, because it looks in .../python3.6/site-packages obviously.
I see the directory contents and I could manually install them again, but that does not look to me like the right way to do it. I could move the contents to the new directory, but again, this seems to me incorrect either.
How am I supposed to handle it properly?
Should I have prepared a pip3 freeze list before the upgrade?
I tried to search, but the keywords are probably too general and got many unrelated answers.

Python 3.5 was replaced by Python 3.6. But you still have the backup option of using python 3.5.
If you want to use python 3.6 you will have to reinstall all pip packages again for python 3.6. And it makes sense.
Say you were changing from 2.7 to 3.5. You would want to preserve both the environments separately. Hence 3.6 environment is different from 3.5.
A quick way to do this would be to pip freeze for 3.5 and then install those dependencies for 3.6.
pip freeze > reqs.txt
upgrade
pip install -r reqs.txt
Since you don't have this option anymore, first try and list all packages in your python3.5
for that you can install pip3.5 as answered by #kabanus.
sudo apt-get install python3=3.5.1*
sudo python3.5 easy_install.py pip
Also it's advised to use virtual environment per project so you can maintain separate environments for each of them.

I just hit the same problem upgrading from Python 3.6 to Python 3.7, I forgot to run pip freeze before I upgraded to Python 3.7. The solution that worked is to specify the --path option as the old site-packages/ directory (which was not deleted):
pip3 freeze --path /usr/local/lib/python3.6/site-packages/ > python3.6_requirements.txt
pip3 install -r python3.6_requirements.txt

This would have made things simpler for you to reinstall. Checkout the description. Using freeze you could have done something like:
$ env1/bin/pip3 freeze > requirements.txt
$ env2/bin/pip3 install -r requirements.txt
Generally the recommended method is you use a virtualenv for site packages, so you don't litter your installation areas, but TBH it never broke something for me. Another option is to check if the linux distribution has the package available for proper retrieval, as in:
sudo apt-get install python3-<somemodule>
This is what I prefer - and could have been upgraded with the distro. As for what to do now, If you really don't want to re-install everything properly you could try to cp /usr/lib/python3.5/site-packages/* /usr/lib/python3.6/site-packages. The differences between versions are not so great such that I believe most packages would work off the bat. You may have to sed to replace python3.5 with python3.6 in all files there though. Forgot delete all pyc files if you do this.
Python modules are self contained enough that if something is broken it can be handled per package, and the site packages are self contained completely, so you could always just remove everything and re-install.
A final note - you can try and install pyton3.5/pip3.5 for your linux, and then do the freeze thing. If there is no package you could install manually (whl or such) or compile a stand alone and configure the site path properly. If you want to keep things on a global site package directory or migrate to virtualenv this may be the safest option.

How to update macOS Python packages

I've been getting the following notice every time I fire up Terminal recently:
/usr/local/lib/python3.5/site-packages/powerline/bindings/zsh/powerline.zsh:200: /usr/local/bin/powerline-config: bad interpreter: /usr/local/opt/python3/bin/python3.5: no such file or directory
/usr/local/lib/python3.5/site-packages/powerline/bindings/zsh/powerline.zsh:204: /usr/local/bin/powerline-config: bad interpreter: /usr/local/opt/python3/bin/python3.5: no such file or directory'
I posted this to the powerline GitHub page as an issue and was advised that I have probably updated my Python version on macOS without me knowing (through homebrew or perhaps through a system update, since I'm a macOS Beta user).
Is there a way to update Python packages en-masse in macOS as in Gentoo, as pointed out in the linked GitHub issue? The dev mentioned that there is a python-updater script in Gentoo that is used to update packages after updating Python, but no such script exists for macOS.
OS: macOS 10.12.3 Beta
Any help would be appreciated!
-- paanvaannd

I think the diagnosis from the powerline GitHub page is correct: you updated Python 3.5 to 3.6 via Homebrew, so things expecting the 3.5 interpreter are broken. I've done this a couple of times.
The best way I've found to keep things sane is to store a list of the installed packages, remove them, update Python via Homebrew, and then reinstall the list of packages. In the shell this would be:
$ pip3 list | cut -d " " -f 1 > package-list.txt # Store package names without versions
$ pip3 uninstall -y $(cat package-list.txt) # Cannot use redirection
$ brew update && brew upgrade python3
$ pip3 install $(cat package-list.txt)
This isn't very helpful for you now, because you've already upgraded with out keeping this list. One option is to roll back your Python installation via Homebrew. If you've not yet done brew cleanup, you can do brew switch python3 3.5.x (where x is the latest version you had). After this, you can do the procedure above, swapping brew upgrade python3 with brew switch python3 3.6.0.
If you've cleaned up the previous installation, you can try to reinstall it, using this answer, and then do the above.
If both those fail, you can just manually re-install your packages. Look at what's in /usr/local/lib/python3.5/site-packages/ for a list of the packages you had installed via Pip, and install them manually. The best thing is to pick one with lots of dependencies, so that you install many packages at once.

beginner installing nosetests package

I'm trying to install the following as per the learnpythonthehardway tutorial:
pip from http://pypi.python.org/pypi/pip
distribute from http://pypi.python.org/pypi/distribute
nose from http://pypi.python.org/pypi/nose/
virtualenv from http://pypi.python.org/pypi/virtualenv
I've visited these links and clicked the download button; each file is in my downloads folder now, and I unarchive/unzipped them- not sure what this means but it seemed required. Are they "installed"? If not, what does it mean to really install them? I've tried typing nosetests in the terminal (as the book says you should), as well as tried easy_install but that doesn't seem to work. It appears my understanding is limited in a number of ways here.
I get the following -bash: nosetests: command not found, but am trying to get:
nosetests . ---------------------------------------------------------------------- Ran 1 test in 0.007s
OK

This works on mac, it may work on linux
1)Open terminal (Be prepared to enter your password)
2)Type: sudo easy_install pip
3)Type: sudo pip install distribute
4)Type: sudo pip install nose
5)Type: sudo pip install virtualenv
Hope that helps, cheers!

They are not yet installed.
Each has its own vagaries about how exactly the install process works and in fact some of those packages will include the other packages with them. I.e.
"If you use virtualenv, a copy of pip will be automatically be
installed in each virtual environment you create."
http://pypi.python.org/pypi/pip
As a relatively new python user myself, (and having gone thru a similar process not long ago) I am eagerly awaiting the more complete answers that come back for this one.

Find out where the nosetests script is. On OSX:
/usr/local/share/nosetests
Execute directly, or set up a bash alias, perhaps in .bash_profile:
alias nosetests='/usr/local/share/python/nosetests'
alias nosetests3='/usr/local/share/python3/nosetests'

You were on the right track, but the package is actually called "nose". The documentation for nose (nosetests is the script) can be found here:
https://nose.readthedocs.org/en/latest/
The short answer is that you have 2 basic options: easy_install or pip. Install using setuptools/distribute: easy_install nose. Install using pip: pip install nose

Install Python 2.7 and after just do this (only on Windows - Windows PowerShell):
Setup PATH env:
C:\Python27\;C:\Python27\Scripts\", "User"
pip is already installed
3.
.\pip install distribute
4.
.\pip install nose
5.
.\pip install virtualenv

How to properly install Python on OSX for use with OpenCV?

I spent the past couple of days trying to get opencv to work with my Python 2.7 install. I kept getting an error saying that opencv module was not found whenever I try "import cv".
I then decided to try installing opencv using Macports, but that didn't work.
Next, I tried Homebrew, but that didn't work either.
Eventually, I discovered I should modify the PYTHONPATH as such:
export PYTHONPATH="/usr/local/lib/python2.6/site-packages/:$PYTHONPATH"
My problem is that I didn't find /usr/local/lib/python2.*...etc
The folder simply doesn't exist
So my question is this:
How do I properly install Python on OS X Snow Leopard for it to work with opencv?
Thanks a lot,

I spent a couple days on this myself. For me, the problem was that that OpenCV installer was not finding the right python installation. It was defaulting to the MacOS-installed version despite the fact that I had upgraded python with homebrew and was using a virtualenv for python. I have collected most of my setup in a gist here:
https://gist.github.com/4150916
Use homebrew to get all the dependencies, but then download the OpenCV tarball and compile yourself being sure to specify all the python related configuration options.
Assuming a virtualenv named 'opencv'...
cd OpenCV-2.4.3/
mkdir release
cd release
cmake -D PYTHON_EXECUTABLE=$WORKON_HOME/opencv/bin/python \
-D PYTHON_PACKAGES_PATH=$WORKON_HOME/opencv/lib/python2.7/site-packages \
-D INSTALL_PYTHON_EXAMPLES=ON\
-D PYTHON_INCLUDE_DIR=/usr/local/Cellar/python/2.7.3/Frameworks/Python.framework/Headers\
-D PYTHON_LIBRARY=/usr/local/Cellar/python/2.7.3/Frameworks/Python.framework/Versions/2.7/lib/libpython2.7.dylib\
..
make -j8
make install

You need to install the module using your python2.7 installation. Pointing your PYTHONPATH at stuff installed under 2.6 to run under 2.7 is a Bad Idea.
Depending on how you want to install it, do something like python2.7 setup.py or easy_install-2.7 opencv to install.
fwiw, on OS X the modules are usually installed under /System/Library/Frameworks/Python.framework/ but you should almost never need to know where anything installed in your site packages is physically located; if Python can't find them without help you've installed them wrong.

Installing OpenCV with Homebrew
brew tap homebrew/homebrew-science
brew install opencv
Setting up Python
Depending on your install location - OS X Default
cd /Library/Python/2.7/site-packages/
or - Homebrew Python
cd /usr/local/lib/python2.7
Then create the symbolic link
ln -s /usr/local/Cellar/opencv/2.4.9/lib/python2.7/site-packages/cv.py cv.py
ln -s /usr/local/Cellar/opencv/2.4.9/lib/python2.7/site-packages/cv2.so cv2.so
The above method sourced from a blog post.

I searched and tried installing opencv3 with python3 for 3 days. Some links suggest for Brew and some virtual env, some say install xcode but all failed in my case.
Dont use linux steps to instal opencv-python on Mac. Problem with Mac is Python 2.7 is already installed by Mac. On top of that installing and linking all site-packages is little problematic and we end up with errors.
I'll share what I did: easy steps to install complete package opencv3, numpy, matplotlib, notebook, spyder etc.. on Mac.
Install anaconda, it creates a directory and install everything inside that
use this link -> https://www.continuum.io/downloads
download command-line-install
After download, goto terminal and download location of anaconda.
$ bash Anaconda3-4.3.0-MacOSX-x86_64.sh
Installation will ask you to append path to .bash_profile >> say yes
Goto home directory, run .bash_profile
$ source .bash_profile
check python, should be pointing to
$ which python
$ /.../anaconda/bin/python
Last step
$ pip install opencv-pyhton
$ python
$ import cv2
if no errors, we are good to go.

Problem with virtualenv in Mac OS X

I've installed virtualenv via pip and get this error after creating a new environment:
selenium:~ auser$ virtualenv new
New python executable in new/bin/python
ERROR: The executable new/bin/python is not functioning
ERROR: It thinks sys.prefix is u'/System/Library/Frameworks/Python.framework/ Versions/2.6' (should be '/Users/user/new')
ERROR: virtualenv is not compatible with this system or executable
In my environment:
PYTHONPATH=/System/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/site-packages
PATH=/System/Library/Frameworks/Python.framework/Versions/2.6/bin:/Library/Frameworks/Python.framework/Versions/2.6/bin:/Library/Frameworks/Python.framework/Versions/2.6/bin:/usr/bin:/bin:/usr/sbin:/sbin:/usr/local/bin:/usr/X11/bin
How can I repair this?
Thanks.

Just in case there's someone still seeking for the answer.
I ran into this same problem just today and realized since I already have Anaconda installed, I should not have used pip install virtualenv to install virtual environment as this would give me the error message when trying to initiate it later. Instead, I tried conda install virtualenv then entered virtualenv env_mysite and problem solved.

Like #RyanWilcox mentioned, you might be inadvertently pointing virtualenv to the wrong Python installation. Virtualenv comes with a -p flag to let you specify which interpreter to use.
In my case,
virtualenv test_env
threw the same error as yours, while
virtualenv -p python test_env
worked perfectly.
If you call virtualenv -h, the documentation for the -p flag will tell you which python it thinks it should be using; if it looks wonky, try passing -p python. For reference, I'm on virtualenv 1.11.6.

In case anyone in the future runs into this problem - this is caused by your default Python distribution being conda. Conda has it's own virtual env set up process but if you have the conda distribution of python and still wish to use virtualenv here's how:
Find the other python distribution on your machine: ls -ls /usr/bin/python*
Take note of the availble python version that is not conda and run the code below (note for python 3 and above you have to upgrade virtualenv first): virtualenv -p python2.7(or your python version) flaskapp

I've run across this problem myself. I wrote down the instructions in a README, which I have pasted below....
I have found there are two things that work:
Make sure you're running the latest virtualenv (1.5.1, of this writting)
If you're using a non system Python as your standard Python (which python to check) Forcefully use the System supplied one.
Instead of virtualenv thing use /usr/bin/python2.6 PATH/TO/VIRTUALENV thing (or whatever which
python returned to you - this is what it did for me when I ran into this issue)

I had the same problem and as I see it now, it was caused by a messy Python installation. I have OS X installed for over a year since I bought a new laptop and I have already installed and reinstalled Python for several times using different sources (official binaries, homebrew, official binaries + hand-made adjustments as described here). Don't ask me why I did that, I'm just a miserable newbie believing everything will fix itself after being re-installed.
So, I had a number of different Pythons installed here and there as well as many hardlinks pointing at them inconsistently. Eventually I got sick of all of them and reinstalled OS X carefully cleaned the system from all the Pythons I found using find utility. Also, I have unlinked all the links pointing to whatever Python from everywhere. Then I've installed a fresh Python using homebrew, installed virtualenv and everything works as a charm now.
So, my recipe is:
sudo find / -iname "python*" > python.log
Then analyze this file, remove and unlink everything related to the version of Python you need, reinstall it (I did it with homebrew, maybe official installation will also work) and enjoy. Make sure you unlink everything python-related from /usr/bin and /usr/local/bin as well as remove all the instances of Frameworks/Python.framework/Versions/<Your.Version> in /Library and /System/Library.
It may be a dirty hack, but it worked for me. I prefer not to keep any system-wide Python libraries except pip and virtualenv and create virtual environments for all of my projects, so I do not care about removing the important libraries. If you don't want to remove everything, still try to understand whether your Pythons are, what links point to them and from where. Then think what may cause the problem and fix it.

I ran into a variation of this "not functioning" error.
I was trying to create an environment in a folder that included the path ".../Programming/Developing..." which is actually "/Users/eric/Documents/Programming:Developing/"
and got this error:
ImportError: No module named site
ERROR: The executable env/bin/python2.7 is not functioning
ERROR: It thinks sys.prefix is u'/Users/eric/Documents/Programming:Developing/heroku' (should be u'/Users/eric/Documents/Programming:Developing/heroku/env')
ERROR: virtualenv is not compatible with this system or executable
I tried the same in a different folder and it worked fine, no errors and env/bin has what I expect (activate, etc.).

I got the same problem and I found that it happens when you do not specify the python executable name properly. So for python 2x, for example:
virtualenv --system-site-packages -p python mysite
But for python 3.6 you need to specify the executable name like python3.6
virtualenv --system-site-packages -p python3.6 mysite

On on OSX 10.6.8 leopard, after having "upgraded" to Lion, then downgrading again (ouch - AVOID!), I went through the Wolf Paulus method a few months ago, completely ignorant of python. Deleted python 2.7 altogether and "replaced" it with 3.something. My FTP program stopped working (Fetch) and who knows what else relies on Python 2.7. So at that point I downloaded the latest version of 2.7 from python.org and it's installer got me up and running - until i tried to use virtualenv.
What seems to have worked for me this time was totally deleting Python 2.7 with this code:
sudo rm -R /System/Library/Frameworks/Python.framework/Versions/2.7
removing all the links with this code:
sudo rm /usr/bin/pydoc
sudo rm /usr/bin/python
sudo rm /usr/bin/pythonw
sudo rm /usr/bin/python-config
I had tried to install python with homebrew, but apparently it will not work unless all of XTools is installed, which I have been avoiding, since the version of XTools compatible with 10.6 is ancient and 4GB and mostly all I need is GCC, the compiler, which you can get here.
So I just installed with the latest download from python.org.
Then had to reinstall easy_install, pip, virtualenv.
Definitely wondering when it will be time for a new laptop, but there's a lot to be said for buying fewer pieces of hardware (slave labor, unethical mining, etc).

The above solutions failed for me, but the following worked:
python3 -m venv --without-pip <ENVIRONMENT_NAME>
. <ENVIRONMENT_NAME>/bin/activate
curl https://bootstrap.pypa.io/get-pip.py | python
deactivate
It's hacky, but yes, the core problem really did just seem to be pip.

I did the following steps to get virtualenv working :
Update virtualenv as follows :
==> sudo pip install --upgrade virtualenv
Initialize python3 virtualenv :
==> virtualenv -p python3 venv

I had this same issue, and I can confirm that the problem was with an outdated virtualenv.py file.
It was not necessary to do a whole install --upgrade.
Replacing the virtualenv.py file with the most recent version sufficed.

I also had this problem, and I tried the following method which worked for me:
conda install virtualenv
virtualenv --system-site-packages /anaconda/envs/tensorflow (here envs keeps all the virtual environments made by user)
source /anaconda/envs/tensorflow/bin/activate
Hope it's helpful.

I had this same issue when trying to install py2.7 on a newer system. The root issue was that virtualenv was part of py3.7 and thus was not compatible:
$ virtualenv -p python2.7 env
Running virtualenv with interpreter /usr/local/bin/python2.7
New python executable in /Users/blah/env/bin/python
ERROR: The executable /Users/blah/env/bin/python is not functioning
ERROR: It thinks sys.prefix is u'/Library/Frameworks/Python.framework/Versions/2.7' (should be u'/Users/blah/env')
ERROR: virtualenv is not compatible with this system or executable
$ which virtualenv
/Library/Frameworks/Python.framework/Versions/3.7/bin/virtualenv
# install proper version of virtualenv
$ pip2.7 install virtualenv
$ /Library/Frameworks/Python.framework/Versions/2.7/bin/virtualenv -p python2.7 env
$ . ./env/bin/activate
(env) $

Open terminal and type /Library/Frameworks/Python.framework/Versions/
then type ls /Library/Frameworks/Python.framework/Versions/2.7/bin/
if you are using Python2(or any other else).
Edit ~/.bash_profile and add the following line:
export PATH=$PATH:/Library/Frameworks/Python.framework/Versions/2.7/bin/
cat ~/.bash_profile
In my case the content of ~/.bash_profile is as follows:
export PATH=$PATH:/Library/Frameworks/Python.framework/Versions/2.7/bin/
Now the virtualenv command should work.

If you continue to have trouble with virtualenv, you might try pythonbrew, instead. It's an alternate solution to the same problem. It works more like Ruby's rvm: It builds and creates an entire instance of Python, under $HOME/.pythonbrew, and then sets up some bash functions that allow you to switch easily between versions. Where virtualenv shadows the system version of Python, using symbolic links as part of its solution, pythonbrew builds entirely self-contained installations of Python.
I used virtualenv for years. It's a decent solution, but I've switched to pythonbrew lately. Having completely self-contained Python instances means that installing a new one takes awhile (since pythonbrew actually compiles Python from scratch), but the self-contained nature of each installation appeals to me. And disk is cheap.