python 'import' package works in shell but not in Jupyter Notebook; issues with setup.py

[robertdefilippi]

I'm trying to import a custom package(s) into my Jupyter Notebook, but I'm having issues using the setup_tools provided by cookiecutter. The package(s) work when I open my termina and run the python3 shell, however I get a 'no module error' when importing the package into my notebook.

I've set up a virtual environment with python3 -m venv and installed the setup.py with pip3 install --editable .

This is my project structure, which is exactly what is provided in cookiecutter

./-- notebook -- my_notebook.ipynb
  └── src -- funct -- func.py
  └── setup.py

I also know I've correctly installed the package:

$ pip3 list
Package    Version Location
---------- ------- -------------------------------------------------------
pip        18.0
setuptools 39.0.1
src        0.1.0   /Users/me/project

In my terminal:

$ python3
>>> import src
>>> from src.funct import func
>>> func.foo()
'Foo'

However, when I run import src in my notebook, I get:

import src
ImportError: No module named src

This is my setup.py

setup(
    name='src',
    packages=find_packages(),
    version='0.1.0',
    description='Example',
    author='me',
    license='MIT',
)

If the package can be imported correctly into my python shell, why can't it be imported into my notebook? Is this a permissions issue, and if so how can I check that?

Here is the SOURCES.txt generated from the setup.py

README.md
src/__init__.py
src.egg-info/PKG-INFO
src.egg-info/SOURCES.txt
src.egg-info/dependency_links.txt
src.egg-info/top_level.txt
src/funct/__init__.py
src/funct/func.py

So now I'm more confused on why it won't import.

Any help would be appreciated.

[pjbull]

I would guess that when you run jupyter it is not part of your environment where the package is installed.

If you run which python3 you should see the environment name in the path to the python3 executable. If you run which jupyter I suspect you will not see your environment name in the path.

To fix this, within the environment run pip install jupyter. Then if you use the jupyter within the environment, it should work. (Check by running which jupyter again).

[robertdefilippi]

# without the virtual env
$ which python3
/usr/local/bin/python3

# with the virtual env
$ which python3
/Users/foo/cookiecutter/bin/python3

$ which python
/usr/bin/python

$ which jupyter
/usr/local/bin/jupyter

I'll try pip install jupyter within the virtual environment, and get back to you.

[pjbull]

Closing, since this isn't a package issue.