Develop Reference

Unable to run azure command line

I have successfully installed the Azure Command line. Also, python is installed on the server. However, when I run /usr/bin/az command it gives error as follows:
Could not import runpy module
The output for python version is as follows:
Python 2.7.5 (default, Nov 16 2018, 04:38:38)
[GCC 4.4.7 20120313 (Red Hat 4.4.7-23)] on linux2
Type "help", "copyright", "credits" or "license" for more information.

The issue you are facing generally comes into the picture whenever there's a broken python environment on the host. In most cases, the fix is to re-install the package, but for the Azure CLI, you will have to make some other arrangements. I solved the issue by using the following method:
Step 1: Find the location of your AZ using which az most likely it should be sitting in /usr/bin/.
Step 2: Take a copy of your existing az and copy it to let's say az_copy
Step 3: Edit the az file using any of the available editors and replace with the following:
#!/usr/bin/env bash
#/usr/lib64/az/bin/python -Esm azure.cli "$#"
export LD_LIBRARY_PATH=/opt/rh/python27/root/usr/lib64
export PYTHONPATH=/usr/lib64/az/lib/python2.7:/usr/lib64/az/lib/python2.7/site-packages
/opt/rh/python27/root/usr/bin/python2.7 -sm azure.cli "$#"
# -E : ignore PYTHON* environment variables (such as PYTHONPATH)
# -s : don't add user site directory to sys.path; also PYTHONNOUSERSITE
Post this, save and execute az, it should work.
Do let me know if that worked out for you.

Python unicode, UCS4 build or UCS2 build

On my Linux machine, directly execute python command, it shows that my Python is UCS4 build.
Python 2.7.3 (default, Jan 8 2018, 17:43:28)
[GCC 4.8.2] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import sys
>>> if sys.maxunicode > 65535:
... print 'UCS4 build'
... else:
... print 'UCS2 build'
...
UCS4 build
However, when I call python in C++ program using
Py_Initialize();
PyRun_SimpleString("import sys");
PyRun_SimpleString("if sys.maxunicode > 65535:\n print 'UCS4 build'\nelse:\n print 'UCS2 build'");
it prints "UCS2 build".
Other information from the python called by c++ are:
platform:Linux-2.6.32_1-19-0-0-x86_64-with-centos-6.3-Final
('Python', '2.7.5 (default, Apr 13 2016, 14:25:24)
[GCC 4.4.6 20120305 (Red Hat 4.4.6-4)]')
('Python', '******/venv')
I have double checked the python executable path. They are from the same path, but python version and gcc version are different.
Anyone know the reason of this strange sympotom?

Well, I solved this problem. Using ldd command, I found that libpython2.7.so.1.0 pointed to a wrong default path, not the one printed before. Correct the LD_LIBRARY_PATH solved this confusion.....

Subprocess popen interaction with virtualenv

I am using virtualenv to run a script that uses subprocess popen to run another program that requires the system wide python and original environment variables. How do I prevent virtualenv from affecting it?

You can pass in a modified PATH for the subprocess with env=.
from subprocess import Popen
from os import environ
from os.path import join as path_join
myenv = environ.copy()
if 'VIRTUAL_ENV' in environ:
myenv['PATH'] = ':'.join(
[x for x in environ['PATH'].split(':')
if x != path_join(environ['VIRTUAL_ENV'], 'bin')])
Popen(['command', '--with', 'arguments'], env=myenv)

virtualenv creates a copy of python executable and you can activate it to your current shell:
This will change your $PATH so its first entry is the virtualenv’s
bin/ directory. (You have to use source because it changes your shell
environment in-place.) This is all it does; it’s purely a convenience.
If you directly run a script or the python interpreter from the
virtualenv’s bin/ directory (e.g. path/to/ENV/bin/pip or
/path/to/ENV/bin/python-script.py) there’s no need for activation.
So when I've activated python in a virtualenv for my project it's the one that will be used:
gonczor#wiktor-papu:~/Projects/papukurier/papukurier$ source ../venv/bin/activate
(venv) gonczor#wiktor-papu:~/Projects/papukurier/papukurier$ which python
/home/gonczor/Projects/papukurier/venv/bin/python
(venv) gonczor#wiktor-papu:~/Projects/papukurier/papukurier$ python
Python 2.7.14 (default, Sep 23 2017, 22:06:14)
[GCC 7.2.0] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import sys
>>> sys.executable
'/home/gonczor/Projects/papukurier/venv/bin/python'
>>>
But at the same time you can give full path to execute any other python instance on your computer:
(venv) gonczor#wiktor-papu:~/Projects/papukurier/papukurier$ /usr/bin/python
Python 2.7.14 (default, Sep 23 2017, 22:06:14)
[GCC 7.2.0] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import sys
>>> sys.executable
'/usr/bin/python'
>>>

I believe subprocess doesn't care about you venv
subprocess.run('/path/to/system/python program.py', stdout=PIPE, stderr=PIPE)

Trying to understand the pythonpath variable

What is the $PYTHONPATH variable, and what's the significance in setting it?
Also, if I want to know the content of my current pythonpath, how do I find that out?

As already answered PYTHONPATH is the search path used by python to find modules when you ''import' them.
Echoing $PYTHONPATH is only useful if you have set it.
However in python itself you can list out what it thinks it is.
$ python3
Python 3.4.3 (default, Oct 14 2015, 20:28:29)
[GCC 4.8.4] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import sys
>>> sys.path
['', '/usr/lib/python3.4', '/usr/lib/python3.4/plat-x86_64-linux-gnu', '/usr/lib/python3.4/lib-dynload', '/usr/local/lib/python3.4/dist-packages', '/usr/lib/python3/dist-packages']
>>>

PYTHONPATH is the default search path for importing modules. If you use bash, you could type echo $PYTHONPATH to look at it.

Clean way to have project-varying sys.path in Spyder

I often have a collection of .py files, maybe split over a few folders, that all use some little special purpose library I've written that doesn't belong in my general PYTHONPATH. What's the cleanest way for me to get that library into the Python path for those scripts that need it? I'll be the only one ever to run any of these, and I will generally be doing so from Spyder.
I've tried with .pth files. This may not be the best way, but I can't get them to work regardless. For example, if I create a .pth file containing
/var
The running the following code out of the same directory via Spyder, I get output of True, False, True:
import sys
import site
print any(["python" in path for path in sys.path]) # True
print any(["/var" in path for path in sys.path]) # False
site.addsitedir(".")
print any(["/var" in path for path in sys.path]) # True
Here are the details of my environment:
Python 2.7.6 |Anaconda 1.8.0 (x86_64)| (default, Nov 11 2013, 10:49:09)
[GCC 4.0.1 (Apple Inc. build 5493)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
Imported NumPy 1.7.1, SciPy 0.13.0, Matplotlib 1.3.1