How to structure python project for use with coverage.py?

Question

I have a package with the following layout:

tiny-py-interpreter/
|-- setup.py
|-- .coveragerc
|-- tinypy/
    | -- foo/
    | -- bar/
    | -- tests/
    | -- tinypyapp.py
    | -- run_tests.py

Here are some lines from the setup.py:

setup(
    ...
    packages=find_packages(),
    entry_points = {
        'console_scripts' : [ 'tinypy = tinypy.tinypyapp:main']
    },
    test_suite = 'tinypy.run_tests.get_suite',
)

After installing the package a console script called tinypy is installed:

pip3 install .

Then I run coverage:

coverage run setup.py test
coverage combine
coverage report

All the tests I have are implemented in a such way that each test launches a subprocess of the tinypy, so I use parallel = True in .coveragerc to capture the result of the coverage run.

Essentially, I have the same layout as coverage itself, where coverage and cmdline.py are the same things as tinypy and tinypyapp.py in my case.

The problem: when console script tinypy (tinypyapp.py) is executed, it uses an installed version of the tinypy package, which is located in /usr/local/lib/python3.5/site-packages/tinypy/. Coverage ignores sources in ./tinypy (as they're not used). If source parameter is omitted, that it's possible to see coverage for the code from site-packages/tinypy.

The question: how to structure the project properly? When tinypyapp.py is installed as a script, it is installed on the same level as tinypy folder (one level higher, not inside). I can't place tinypyapp.py outside the tinypy folder, as then setup.py can't find it. Because of this I can't use tinypyapp.py and have to use name of the script (which is tinypy).

Ned Batchelder · Accepted Answer · 2015-12-22 17:14:34Z

1

I think the easiest thing would be to not install the code you are working on into site-packages, but instead to use a development install:

pip install -e .

answered Dec 22, 2015 at 17:14

Ned Batchelder

378k77 gold badges583 silver badges675 bronze badges

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

Comments

Max Malysh · Accepted Answer · 2015-12-21 20:28:32Z

0

My current workaround is to create test_entry_point.py in the root folder:

tiny-py-interpreter/
|-- setup.py
|-- test_entry_point.py
|-- ...

With the following content:

import sys
from tinypy.tinypyapp import main

if __name__ == '__main__':
    main()

And use the following filename when launching a subprocess in test:

tinypy_binary = sys.executable + ' ' + os.getcwd() + '/test_entryp.py'
subprocess.run(tinypy_binary, ...)

answered Dec 21, 2015 at 20:28

Max Malysh

32k21 gold badges117 silver badges125 bronze badges

Collectives™ on Stack Overflow

How to structure python project for use with coverage.py?

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related