Distributing a Python Command Line Program with Multiple Modules

Question

I'm attempting to release a command line program written in Python to the public, but would first like to test it to make sure that the program works as I intend. To do so, I create a simple wheel distribution of the program, create a virtualenv instance in another directory, and install the wheel into said environment. Easy, right?

Not so.

When I attempt to call the program in the testing environment, I receive this error message:

Traceback (most recent call last):
  File "~/testing/bin/leprechaun", line 7, in <module>
    from leprechaun.main import main
  File "~/testing/lib/python3.4/site-packages/leprechaun/main.py", line 8, in <module>
    import generator
ImportError: No module named 'generator'

So, it seems that the generator.py file isn't being imported for some reason. What's strange, however, is that generator.py and main.py (the main entry point to the program) exist in the same directory.

For reference, here's my project's tree hierarchy:

├── leprechaun
│   ├── data
│   │   └── A bunch of text files that aren't important
│   ├── db.py
│   ├── generator.py
│   ├── __init__.py
│   ├── main.py
│   └── rainbow.py
├── LICENSE.txt
├── MANIFEST.in
├── README.rst
└── setup.py

And, for kicks, my setup.py script:

#!/usr/bin/env python3

import re
from setuptools import setup, find_packages

version = re.search("^__version__\s*=\s*\"(.*)\"$",
  open("leprechaun/__init__.py").read(), re.M).group(1)

setup(
  name='leprechaun',
  version=version,
  description="A simple rainbow table generator",
  long_description=open("README.rst", encoding="utf-8").read(),
  author="My Name",
  author_email="My Email",
  url="https://github.com/zcdziura/leprechaun",
  license="MIT",
  packages=find_packages(),
  package_data={
    "leprechaun": ["data/wordlist*"]
  },
  entry_points={
    "console_scripts": [
      "leprechaun = leprechaun.main:main"
    ]
  }
)

I assumed that the entry-point script that is generated would simply call the main function from my package, which would then be able to import modules from the same directory. However, that doesn't seem to be the case.

How would I fix this?

pswaminathan · Accepted Answer · 2014-04-10 22:09:50Z

1

The problem you're facing deals with absolute imports. You might want to take a look at this PEP for more information on what the rationale behind it is.

The short answer is that every import statement you make now looks along sys.path. Whereas in Python 2.x, you could use import generator and it would look in the current directory, Python 3.x requires it to be along the path. So if you change to from leprechaun import generator or from . import generator you should be able to clear this up.

answered Apr 10, 2014 at 22:09

pswaminathan

9,1441 gold badge24 silver badges29 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Z. Charles Dziura Over a year ago

That did it! I didn't realize that Python 3 changed the import behavior. I guess it makes sense why they did it, though. Now I simply have to fix a couple of new bugs that have sprung up. Thank you!

Collectives™ on Stack Overflow

Distributing a Python Command Line Program with Multiple Modules

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related