Current Development Version:
Most Recent Stable Release:
Info:
Have a CLI Python application?
Want to automate testing of the actual console input & output of your user-facing components?
stdio Manager can help.
While some functionality here is more or less duplicative of
redirect_stdout and redirect_stderr in contextlib
within the standard library,
it provides (i) a much more concise way to mock both stdout and stderr at the same time,
and (ii) a mechanism for mocking stdin, which is not available in contextlib.
First, install:
$ pip install stdio-mgr
Then use!
All of the below examples assume stdio_mgr has already
been imported via:
>>> from stdio_mgr import stdio_mgr
Mock stdout:
>>> with stdio_mgr() as (in_, out_, err_):
... print('foobar')
... out_cap = out_.getvalue().replace(os.linesep, '\n')
>>> out_cap
'foobar\n'
>>> in_.closed and out_.closed and err_.closed
True
By default print
appends a newline
after each argument, which is why out_cap is 'foobar\n'
and not just 'foobar'.
As currently implemented, stdio_mgr closes all three mocked streams
upon exiting the managed context.
Mock stderr:
>>> import warnings
>>> with stdio_mgr() as (in_, out_, err_):
... warnings.warn("foo has no bar")
... err_cap = err_.getvalue().replace(os.linesep, '\n')
>>> err_cap
'...UserWarning: foo has no bar\n...'
Mock stdin:
The simulated user input has to be pre-loaded to the mocked stream.
Be sure to include newlines in the input to correspond to
each mocked Enter keypress!
Otherwise, input will hang, waiting for a newline
that will never come.
If the entirety of the input is known in advance,
it can just be provided as an argument to stdio_mgr.
Otherwise, .append() mocked input to in_
within the managed context as needed:
>>> with stdio_mgr('foobar\n') as (in_, out_, err_):
... print('baz')
... in_cap = input('??? ')
...
... _ = in_.append(in_cap[:3] + '\n')
... in_cap2 = input('??? ')
...
... out_cap = out_.getvalue().replace(os.linesep, '\n')
>>> in_cap
'foobar'
>>> in_cap2
'foo'
>>> out_cap
'baz\n??? foobar\n??? foo\n'
The _ = assignment suppresses printing of the return value
from the in_.append() call--otherwise, it would be interleaved
in out_cap, since this example is shown for an interactive context.
For non-interactive execution, as with unittest, pytest, etc.,
these 'muting' assignments should not be necessary.
Both the '??? ' prompts for input
and the mocked input strings
are echoed to out_, mimicking what a CLI user would see.
A subtlety: While the trailing newline on, e.g., 'foobar\n' is stripped
by input, it is retained in out_.
This is because in_ tees the content read from it to out_
before that content is passed to input.
Want to modify internal print calls
within a function or method?
In addition to mocking, stdio_mgr can also be used to
wrap functions that directly output to stdout/stderr. A stdout example:
>>> def emboxen(func):
... def func_wrapper(s):
... from stdio_mgr import stdio_mgr
...
... with stdio_mgr() as (in_, out_, err_):
... func(s)
... content = out_.getvalue()
...
... max_len = max(map(len, content.splitlines()))
... fmt_str = '| {{: <{0}}} |\n'.format(max_len)
...
... newcontent = '=' * (max_len + 4) + '\n'
... for line in content.splitlines():
... newcontent += fmt_str.format(line)
... newcontent += '=' * (max_len + 4)
...
... print(newcontent)
...
... return func_wrapper
>>> @emboxen
... def testfunc(s):
... print(s)
>>> testfunc("""\
... Foo bar baz quux.
... Lorem ipsum dolor sit amet.""")
===============================
| Foo bar baz quux. |
| Lorem ipsum dolor sit amet. |
===============================
Available on PyPI
(pip install stdio-mgr).
Source on GitHub. Bug reports and feature requests are welcomed at the Issues page there.
Copyright (c) 2018-2019 Brian Skinn
License: The MIT License. See LICENSE.txt for full license terms.