278
votes

I'm trying to use TDD (test-driven development) with pytest. pytest will not print to the console when I use print.

I am using pytest my_tests.py to run it.

The documentation seems to say that it should work by default: http://pytest.org/latest/capture.html

But:

import myapplication as tum

class TestBlogger:

    @classmethod
    def setup_class(self):
        self.user = "alice"
        self.b = tum.Blogger(self.user)
        print "This should be printed, but it won't be!"

    def test_inherit(self):
        assert issubclass(tum.Blogger, tum.Site)
        links = self.b.get_links(posts)
        print len(links)   # This won't print either.

Nothing gets printed to my standard output console (just the normal progress and how many many tests passed/failed).

And the script that I'm testing contains print:

class Blogger(Site):
    get_links(self, posts):
        print len(posts)   # It won't get printed in the test.

In unittest module, everything gets printed by default, which is exactly what I need. However, I wish to use pytest for other reasons.

Does anyone know how to make the print statements get shown?

7
Maybe stdout is being overwritten. What happens if you use sys.stdout.write("Test")? How about sys.__stdout__.write("Test")? The latter should always write to the system-defined stdout, which should be the console. If the two commands do different things, then stdout is being changed; if they do the same thing, then the problem is something else.TheSoundDefense

7 Answers

313
votes

By default, py.test captures the result of standard out so that it can control how it prints it out. If it didn't do this, it would spew out a lot of text without the context of what test printed that text.

However, if a test fails, it will include a section in the resulting report that shows what was printed to standard out in that particular test.

For example,

def test_good():
    for i in range(1000):
        print(i)

def test_bad():
    print('this should fail!')
    assert False

Results in the following output:

>>> py.test tmp.py
============================= test session starts ==============================
platform darwin -- Python 2.7.6 -- py-1.4.20 -- pytest-2.5.2
plugins: cache, cov, pep8, xdist
collected 2 items

tmp.py .F

=================================== FAILURES ===================================
___________________________________ test_bad ___________________________________

    def test_bad():
        print('this should fail!')
>       assert False
E       assert False

tmp.py:7: AssertionError
------------------------------- Captured stdout --------------------------------
this should fail!
====================== 1 failed, 1 passed in 0.04 seconds ======================

Note the Captured stdout section.

If you would like to see print statements as they are executed, you can pass the -s flag to py.test. However, note that this can sometimes be difficult to parse.

>>> py.test tmp.py -s
============================= test session starts ==============================
platform darwin -- Python 2.7.6 -- py-1.4.20 -- pytest-2.5.2
plugins: cache, cov, pep8, xdist
collected 2 items

tmp.py 0
1
2
3
... and so on ...
997
998
999
.this should fail!
F

=================================== FAILURES ===================================
___________________________________ test_bad ___________________________________

    def test_bad():
        print('this should fail!')
>       assert False
E       assert False

tmp.py:7: AssertionError
====================== 1 failed, 1 passed in 0.02 seconds ======================
102
votes

Short Answer

Use the -s option:

pytest -s

Detailed answer

From the docs:

During test execution any output sent to stdout and stderr is captured. If a test or a setup method fails its according captured output will usually be shown along with the failure traceback.

pytest has the option --capture=method in which method is per-test capturing method, and could be one of the following: fd, sys or no. pytest also has the option -s which is a shortcut for --capture=no, and this is the option that will allow you to see your print statements in the console.

pytest --capture=no     # show print statements in console
pytest -s               # equivalent to previous command

Setting capturing methods or disabling capturing

There are two ways in which pytest can perform capturing:

  1. file descriptor (FD) level capturing (default): All writes going to the operating system file descriptors 1 and 2 will be captured.

  2. sys level capturing: Only writes to Python files sys.stdout and sys.stderr will be captured. No capturing of writes to filedescriptors is performed.

pytest -s            # disable all capturing
pytest --capture=sys # replace sys.stdout/stderr with in-mem files
pytest --capture=fd  # also point filedescriptors 1 and 2 to temp file
96
votes

Using -s option will print output of all functions, which may be too much.

If you need particular output, the doc page you mentioned offers few suggestions:

  1. Insert assert False, "dumb assert to make PyTest print my stuff" at the end of your function, and you will see your output due to failed test.

  2. You have special object passed to you by PyTest, and you can write the output into a file to inspect it later, like

    def test_good1(capsys):
        for i in range(5):
            print i
        out, err = capsys.readouterr()
        open("err.txt", "w").write(err)
        open("out.txt", "w").write(out)
    

    You can open the out and err files in a separate tab and let editor automatically refresh it for you, or do a simple py.test; cat out.txt shell command to run your test.

That is rather hackish way to do stuff, but may be it is the stuff you need: after all, TDD means you mess with stuff and leave it clean and silent when it's ready :-).

18
votes

I needed to print important warning about skipped tests exactly when PyTest muted literally everything.

I didn't want to fail a test to send a signal, so I did a hack as follow:

def test_2_YellAboutBrokenAndMutedTests():
    import atexit
    def report():
        print C_patch.tidy_text("""
In silent mode PyTest breaks low level stream structure I work with, so
I cannot test if my functionality work fine. I skipped corresponding tests.
Run `py.test -s` to make sure everything is tested.""")
    if sys.stdout != sys.__stdout__:
        atexit.register(report)

The atexit module allows me to print stuff after PyTest released the output streams. The output looks as follow:

============================= test session starts ==============================
platform linux2 -- Python 2.7.3, pytest-2.9.2, py-1.4.31, pluggy-0.3.1
rootdir: /media/Storage/henaro/smyth/Alchemist2-git/sources/C_patch, inifile: 
collected 15 items 

test_C_patch.py .....ssss....s.

===================== 10 passed, 5 skipped in 0.15 seconds =====================
In silent mode PyTest breaks low level stream structure I work with, so
I cannot test if my functionality work fine. I skipped corresponding tests.
Run `py.test -s` to make sure everything is tested.
~/.../sources/C_patch$

Message is printed even when PyTest is in silent mode, and is not printed if you run stuff with py.test -s, so everything is tested nicely already.

6
votes

According to the pytest docs, pytest --capture=sys should work. If you want to capture standard out inside a test, refer to the capsys fixture.

3
votes

I originally came in here to find how to make PyTest print in VSCode's console while running/debugging the unit test from there. This can be done with the following launch.json configuration. Given .venv the virtual environment folder.

    "version": "0.2.0",
    "configurations": [
        {
            "name": "PyTest",
            "type": "python",
            "request": "launch",
            "stopOnEntry": false,
            "pythonPath": "${config:python.pythonPath}",
            "module": "pytest",
            "args": [
                "-sv"
            ],
            "cwd": "${workspaceRoot}",
            "env": {},
            "envFile": "${workspaceRoot}/.venv",
            "debugOptions": [
                "WaitOnAbnormalExit",
                "WaitOnNormalExit",
                "RedirectOutput"
            ]
        }
    ]
}
2
votes

This is the cleanest way I know to print a single statement to sys.stdout (without artificially making your test fail or enabling the -s option) - you get to see the specific output you want and nothing more:

  1. Add the in-built parameter capsys to your test function.

  2. In your code, simply insert:

    with capsys.disabled():
        print("this output will not be captured and go straight to sys.stdout")
    

See https://buildmedia.readthedocs.org/media/pdf/pytest/latest/pytest.pdf (2.11 How to capture stdout/stderr output).