Develop Reference

Pipenv could not find platform independent libraries <prefix>

Very weird problem. While python and works perfectly fine, I couldn't use pipenv.
I'm running it on windows, and it doesn't work on windows terminal, powershell, vscode terminal. Python version is 3.10.5, system PATH already has the folder location where python.exe is. I have also tried adding PYTHONPATH and PYTHONHOME variables but still did not work.
I'm still new at this, I tried pip uninstall pipenv, virtualenv, but when I run pipenv it still gives me the same error message and not the "not recognized as a command" error.
This is the error message when I run pipenv --version, pipenv shell, pipenv install etc.:
Could not find platform independent libraries <prefix>
Consider setting $PYTHONHOME to <prefix>[:<exec_prefix>]
Python path configuration:
PYTHONHOME = (not set)
PYTHONPATH = (not set)
program name = 'c:/program files/inkscape/bin/python.exe'
isolated = 0
environment = 1
user site = 1
import site = 1
sys._base_executable = 'C:\\program files\\inkscape\\bin\\python.exe'
sys.base_prefix = 'D:\\a\\_temp\\msys\\msys64\\mingw64'
sys.base_exec_prefix = 'C:\\program files\\inkscape'
sys.executable = 'C:\\program files\\inkscape\\bin\\python.exe'
sys.prefix = 'D:\\a\\_temp\\msys\\msys64\\mingw64'
sys.exec_prefix = 'C:\\program files\\inkscape'
sys.path = [
'D:\\a\\_temp\\msys\\msys64\\mingw64\\lib\\python38.zip',
'D:\\a\\_temp\\msys\\msys64\\mingw64\\lib\\python3.8',
'D:\\a\\_temp\\msys\\msys64\\mingw64\\lib\\python3.8',
'C:\\program files\\inkscape\\lib\\python3.8\\lib-dynload',
]
Fatal Python error: init_fs_encoding: failed to get the Python codec of the filesystem encoding
Python runtime state: core initialized
ModuleNotFoundError: No module named 'encodings'
Current thread 0x00006e4c (most recent call first):
<no Python frame>
I thought inkscape was the problem, so I uninstalled it, but it still doesn't work.
Please help me, thank you

I may not be able to answer your question but I might contribute to the solution:
1.try to find here why you get the ModuleNotFoundError: No module named 'encodings'
what library are you missing ?
2. Check this older thread here it may be of help : ImportError: No module named 'encodings'
3. I used to use Pipenv in my Jupiter terminal and due to colleagues engineering choices I always had problem so decided to move to Poetry check it out is really good :https://python-poetry.org/
I may have not helped you find the answer but hopefully have made a small contribution...all debugging is like criminal investigation sometimes you need to incrementally acquire new evidence to solve crimes so ..expect lots of patience from yourself and don't get disappointed easily!

This was the answer for me:
Long story short, I uninstalled the python from Microsoft Store, restarted my computer, installed using the python 3.10.5 installer from python.org, made sure I check the checkbox to “add python to PATH”, restart my computer. On the terminal, run “which python” and ensure it is on the right path, which should be “c:\users\ [username]\appdata\local\programs\python\python310\python” and not on “c:\program files\windowsapp....” This should be both on the user variable and system variable for the variable “PATH”
I consulted with a programmer friend and he helped me solve this issue.
This is the process for troubleshooting for anyone facing similar issues as I did on Windows:
Find which python, pip, and pipenv the system is using (different terminals like windows terminal, powershell, cmd prompt might have different results, so just use the one you like the most, which was for me was my modified windows terminal). The error code here only tells me the symptoms and not the actual problem. So pursuing the “encoding module not found” was actually jumping deeper into the rabbit hole. Luckily, I realized from experience that encoding was not the issue, but rather the terminal did not recognize where python or pipenv is. (I do realize that when a module or cmd is not recognized, the terminal will usually say “module not recognized” instead of the this long error prompt. So it is still a mistery to me why it did that instead of the simple “module not recognized”.
“Which python” shows the default python installation used, “python -V” shows which python it is associated with and the location, “pip show certifi” is for showing which python pip is using
I found a discrepancy, where most of my modules use appdata, which is where normally pip installs and python.org installers use. But python was using WindowsApp because I had installed it using microsoft store. (On my defense, when you google python installer most guides will tell you to install from microsoft store, which was a rookie mistake)
Uninstall the python you don’t want on “add or remove program” from windows. Then install python on its website and make sure to check the checkbox for adding the newly installed python to the system environment variable. See if it works on the terminal. Normally the computer needs to reboot to make the necessary changes.
Done! If it doesn’t work, make sure to purge every python you have on the computer, install using the installer directly from the python website, and make sure the python path is added to the environment variable

Nest simulator: python says “no module named nest”

After installing the Nest Neural Simulator, I keep getting the following error when trying to run any of the example python files that came in the installation. I've tried re-installing Nest, Python, and using Anaconda, but no go.
Python error:
ImportError: No module named nest
Suggestions?

At https://nest-simulator.org/documentation you now find many different install instructions and how to solve the "ImportError: no module named nest" depends on the way you installed NEST.
System Python
The problem with the nest python module not being found is usually, that NEST is installed for a specific Python version and you can not load it from another. So while many OS still use Python 2.7 you may need to explicitly run
$ python3
>>> import nest
Additionally, if you have multiple Python 3.x versions installed, modules may still be installed for a different version and you have to explicitly start python with python3.6 or python3.8, etc.
Conda package
As #nosratullah-mohammadi already mentioned, if you have a Conda flavour installed, using the pre-built package is a very quick solution. The link in his post is unfortunately broken; this one should work, then go to "Installation" in the side bar.
$ conda create --name nest -c conda-forge python3 nest-simulator
$ conda activate nest
$ python # this should load the Python from the conda env
>>> import nest # this loads nest which is installed explicitly for that Python
From Source
For every install from source, be sure to have Python and other prerequisites installed before building NEST. Then you can create your temporary build directory (can be deleted afterwards) and configure with the flags you need.
cd somewhere
mkdir nest-build
cd nest-build
cmake -DCMAKE_INSTALL_PREFIX:PATH=/install/path -Dwith-python=3 .../sources/of/nest-simulator
Replace somewhere, /install/path and .../sources/of/nest-simulator with the paths correct for your setup. (A popular choice when compiling from source in conjunction with Conda environments, for example, is to use -CMAKE_INSTALL_PREFIX=$CONDA_PREFIX, which installs NEST directly into the active environment. Conda is however in no way necessary for NEST.)
Add more -D... flags as you prefer. Possible flags you see with cmake -LA .../sources/of/nest-simulator, as pointed out here. You are probably interested in many of the with-xyz at the end. Check the aforementioned documentation for deatils.
Check that the paths and libraries reported in the Configuration Summary make sense (you may need to scroll up a bit to see). It could for example look something like this:
--------------------------------------------------------------------------------
NEST Configuration Summary
--------------------------------------------------------------------------------
[...]
Python bindings : Yes (Python 3.6.8: /home/yourname/miniconda3/envs/nest/bin/python3)
Includes : /home/yourname/miniconda3/envs/nest/include/python3.6m
Libraries : /home/yourname/miniconda3/envs/nest/lib/libpython3.6m.so
Cython bindings : Yes (Cython 0.27.3: /home/yourname/miniconda3/envs/nest/bin/cython)
[...]
--------------------------------------------------------------------------------
[...]
PyNEST will be installed to:
/home/yourname/miniconda3/envs/nest/lib/python3.6/site-packages
--------------------------------------------------------------------------------
In this example CMake configured everything for Python3.6 from my conda environment.
If you are satisfied with your settings and all the found Python versions match, run the usual
$ make # optionally with -j$(nproc)
$ make install
$ make installcheck
In case that works out fine you are done and can delete the build directory to free the space. Congratulations!
Also if things get too mixed up and it doesn't seem to do what you expect, it is sometimes useful to delete the build directory and start off clean.

there is a new method of installation added to other methods, wich is installing nest with conda package and it's in its beta version. but it works and it's really simple.
you can find the installation from here!
simply after install mini conda package run your terminal and type this :
conda create --name ENVNAME -c conda-forge nest-simulator python
then type :
conda activate ENVNAME
and you're good to go!

NEST now provide the solution to that problem and similar ones, by providing a script which automatically sets the relevant system variables:
If your operating system does not find the nest executable or Python
does not find the nest module, your path variables may not be set
correctly. This may also be the case if Python cannot load the nest
module due to missing or incompatible libraries. In this case, please
run
source </path/to/nest_install_dir>/bin/nest_vars.sh
to set the necessary environment variables. You may want to include
this line in your .bashrc file, so that the environment variables are
set automatically.
https://nest-simulator.readthedocs.io/en/latest/installation/linux_install.html

Turns out I needed to move the directory where I installed nest (Users/name/opt/nest) into a nest folder in the following directory in anaconda. Specifically, I moved the contents of the folder (from the nest installation):
/Users/name/opt/nest/lib/python2.7/site-packages/nest
Into this one:
/anaconda/lib/python2.7/site-packages/nest
Disclaimer: I could very well run into problems for not having copied all the contents of the Nest installation, but this little hack is helping me run example files now.

Python installation, failing to find bz2 module

OK, I have an old Debian VM. Package managers are useless. No, I'm not going to update the OS.
I have the bzip2 lib and development headers installed correctly on my system (those actually came from a package).
I start with absolutely NO Python on the system. I removed everything manually. I downloaded the Python 2.7.5 source, and configured with ./configure --prefix=/usr. It configures fine. I run make, and it compiles fine. I try ./python -c "import bz2; print bz2.__doc__", it works, and says:
The python bz2 module provides a comprehensive interface for
the bz2 compression library. It implements a complete file
interface, one shot (de)compression functions, and types for
sequential (de)compression.
I then run make test and the whole test suite progresses fine, and notably the "test_bz2" test passes.
I then run make install, which installs my new Python binary into /usr/bin/ like I wanted.
I try /usr/bin/python -c "import bz2; print bz2.__doc__", and it fails with:
Traceback (most recent call last):
File "", line 1, in
ImportError: No module named bz2
I've tried a bunch of different things, including building Python as --enable-shared and not, no luck. I've tried at least 10 times (each time totally cleaning out everything, running make distclean, etc.). No luck.
I tried: PYTHONPATH="/usr/lib/python2.7"; export PYTHONPATH. Still no luck.
HOWEVER, if I delete the symlink that make install creates for /usr/bin/python, and instead do: ln -s /path/to/my/python/compile/python python, NOW it magically works.
So, what the heck? Why is this Python binary I'm getting created only able to find stuff when the binary exists in the compile directory, and not when it's put into normal production install location? What am I missing?
I am root during the entire process, from configure to make to make install to trying to test the Python import call.
I have started from scratch again (this time compiling with --enable-shared btw), and verified that not only in the compile directory is there build/lib.linux-x86_64-2.7/bz2.so, but once I run make install, that file is put into /usr/lib/python2.7/lib-dynload/bz2.so.
I've tried to do some reading on lib-dynload, but haven't been able to determine if there's something else a Python program (like default configuration for the CLI or whatever) would need to be able to tell it to pull module imports from lib-dynload, or if there's some other place or option to tell the make install where it should be putting it instead of dynload.
Still I have no explanation why the /path/to/compilation/python binary can find and load bz2.so fine, but the /usr/bin/python binary can't find (or load) /usr/lib/python2.7/lib-dynload/bz2.so.
I thought maybe it was something to do with the fact that the installation doesn't create like a /usr/lib/python symlink to point at /usr/lib/python2.7 directory. But I created the symlink and still no go.
I am still lost here.

It would appear that a sort of non-answer answer was arrived at accidentally via a long string of Twitter conversation(s).
I've filed another Stack Overflow question here to ask WHY what we found was the solution to this problem: https://stackoverflow.com/questions/17662091/python-installation-prefix-not-being-persisted-in-config
For posterity sake, right now the solution is that I have to set the PYTHONHOME environment variable to /usr, and everything starts working. The puzzling part is that the documentation says PYTHONHOME should default to {prefix}, which I was clearly setting as default during configure to /usr. So why should I have to manually set it?
Running python-config --prefix reveals that the {prefix} default is in fact /usr/bin, NOT /usr like I specified, which leads to me needing to override the default back to the default, bizarrely.

How do I fix there is no such module error in python 2.6?

I've been using python 2.7 and after installation of python 2.7. All of the scripts have run successfully, but suddenly today, when I run python, it is recognized with python 2.6, so for one of the python packages I get the following error:
/Library/Frameworks/Python.framework/Versions/2.6/Resources/Python.app/Contents/MacOS/Python:
No module named htseq-count
i am not sure why it happens. The path environment variable for python is set to:
PATH="/Library/Frameworks/Python.framework/Versions/2.7/Resources/bin:${PATH}" export PATH
How do I fix this error?

If your code relies on a particular version of Python, you should specify python2.7 or python2.6 instead of just python.
If this "suddenly" happened it's possible that installing some other software modified your $PATH so that the old version of Python now has priority (it's possible for multiple versions to be on your $PATH; it uses the first one it finds).
If you move these two lines to the bottom of the file they're in (maybe ~/.profile) it may correct this.
PATH="/Library/Frameworks/Python.framework/Versions/2.7/Resources/bin:${PATH}"
export PATH
You can type which python in Terminal you can see what binary it finds when looking for python on your $PATH.
If you don't mind using Python 2.6, you can probably install the missing package by typing this in your terminal:
easy_install-2.6 htseq

I think you are applying different python version to be used.
type
which python
under your control.
see which python you are using, you should config the one using python 2.7
python --version

How do I find out what Python libraries are installed on my Mac?

I'm just starting out with Python, and have found out that I can import various libraries. How do I find out what libraries exist on my Mac that I can import? How do I find out what functions they include?
I seem to remember using some web server type thing to browse through local help files, but I may have imagined that!

From the Python REPL (the command-line interpreter / Read-Eval-Print-Loop), type help("modules") to see a list of all your available libs.
Then to see functions within a module, do help("posix"), for example. If you haven't imported the library yet, you have to put quotes around the library's name.

For the web server, you can run the pydoc module that is included in the python distribution as a script:
python /path/to/pydoc.py -p 1234
where 1234 is the port you want the server to run at. You can then visit http://localhost:1234/ and browse the documentation.

Every standard python distribution has these libraries, which cover most of what you will need in a project.
In case you need to find out if a library exists at runtime, you do it like this
try:
import ObscureModule
except ImportError:
print "you need to install ObscureModule"
sys.exit(1) # or something like that

You can install another library: yolk.
yolk is a python package manager and will show you everything you have added via pypi. But it will also show you site-packages added through whatever local package manager you run.

just run the Python interpeter and type the command
import "lib_name"
if it gives an error, you don't have the lib installed...else you are good to go

On Leopard, depending on the python package you're using and the version number, the modules can be found in /Library/Python:
/Library/Python/2.5/site-packages
or in /Library/Frameworks
/Library/Frameworks/Python.framework/Versions/Current/lib/python2.6/site-packages
(it could also be 3.0 or whatever version)...
I guess it is quite the same with Tiger

Considering that in every operating system most of python's packages are installed using 'pip' (see pip documentation) you can also use the command 'pip freeze' on a terminal to print a list of all the packages you have installed through it.
Other tools like 'homebrew' for macOS (used when for some reason you can't install a package using pip) have similar commands, in this specific case 'brew list'.