pyfakefs implements a fake file system that mocks the Python file system modules. Using pyfakefs, your tests operate on a fake file system in memory without touching the real disk. The software under test requires no modification to work with pyfakefs.
pyfakefs works with Linux, Windows and MacOS.
This file provides general usage instructions for pyfakefs. There is more:
- The Release pyfakefs API Reference contains documentation for each pyfakefs class, method and function for the last version released on PyPi
- The Development pyfakefs API Reference contains the same documentation for the current master branch
- The pyfakefs Wiki provides more detailed information on specific topics
- The Release Notes shows a list of changes in the latest versions
In your own documentation, please link to pyfakefs using the canonical URL http://pyfakefs.org. This URL always points to the most relevant top page for pyfakefs.
There are several approaches to implementing tests using pyfakefs.
The first approach is to allow pyfakefs to automatically find all real file functions and modules, and stub these out with the fake file system functions and modules. This is explained in the pyfakefs wiki page
Automatically find and patch file functions and modules
and demonstrated in files example.py
and example_test.py
.
If you use PyTest, you will be interested in the PyTest plugin in pyfakefs. This automatically patches all file system functions and modules in a manner similar to the automatic find and patch approach described above.
The PyTest plugin provides the fs
fixture for use in your test. For example:
def my_fakefs_test(fs):
# "fs" is the reference to the fake file system
fs.create_file('/var/data/xx1.txt')
assert os.path.exists('/var/data/xx1.txt')
If you are using other means of testing like nose, you can do the
patching using fake_filesystem_unittest.Patcher
- the class doing the the actual work
of replacing the filesystem modules with the fake modules in the first two approaches.
The easiest way is to just use Patcher
as a context manager:
from fake_filesystem_unittest import Patcher
with Patcher() as patcher:
# access the fake_filesystem object via patcher.fs
patcher.fs.create_file('/foo/bar', contents='test')
# the following code works on the fake filesystem
with open('/foo/bar') as f:
contents = f.read()
You can also initialize Patcher
manually:
from fake_filesystem_unittest import Patcher
patcher = Patcher()
patcher.setUp() # called in the initialization code
...
patcher.tearDown() # somewhere in the cleanup code
You can also use mock.patch()
to patch the modules manually. This approach will
only work for the directly imported modules, therefore it is not suited for testing
larger code bases. As the other approaches are more convenient, this one is considered
deprecated.
You have to create a fake filesystem object, and afterwards fake modules based on this file system
for the modules you want to patch.
The following modules and functions can be patched:
os
andos.path
byfake_filessystem.FakeOsModule
io
byfake_filessystem.FakeIoModule
pathlib
byfake_pathlib.FakePathlibModule
- build-in
open()
byfake_filessystem.FakeFileOpen
import pyfakefs.fake_filesystem as fake_fs
# Create a faked file system
fs = fake_fs.FakeFilesystem()
# Do some setup on the faked file system
fs.CreateFile('/foo/bar', contents='test')
# Replace some built-in file system related modules you use with faked ones
# Assuming you are using the mock library to ... mock things
try:
from unittest.mock import patch # In Python 3, mock is built-in
except ImportError:
from mock import patch # Python 2
# Note that this fake module is based on the fake fs you just created
os = fake_fs.FakeOsModule(fs)
with patch('mymodule.os', os):
fd = os.open('/foo/bar', os.O_RDONLY)
contents = os.read(fd, 4)
pyfakefs works with CPython 2.7, 3.3 and above, on Linux, Windows and OSX (MacOS), and with PyPy2 and PyPy3.
pyfakefs works with PyTest version 2.8.6 or above.
pyfakefs will not work with Python libraries that use C libraries to access the
file system. This is because pyfakefs cannot patch the underlying C libraries'
file access functions--the C libraries will always access the real file system.
For example, pyfakefs will not work with lxml
. In this case
lxml
must be replaced with a pure Python alternative such as
xml.etree.ElementTree
.
pyfakefs is available on PyPi.
pyfakefs is currently automatically tested:
- On Linux, with Python 2.7, 3.3 and above using Travis
- On MacOS, with Python 2.7 and 3.6, also using Travis.
The Linux/MacOS build is currently
.
- On Windows, with Python 2.7, 3.4 and above using Appveyor.
The Windows build is currently
.
pyfakefs unit tests are available via two test scripts:
$ python all_tests.py
$ py.test pytest_plugin_test.py
These scripts are called by tox
and Travis-CI. tox
can be used to run tests
locally against supported python versions:
$ tox
pyfakefs.py was initially developed at Google by Mike Bland as a modest fake implementation of core Python modules. It was introduced to all of Google in September 2006. Since then, it has been enhanced to extend its functionality and usefulness. At last count, pyfakefs is used in over 2,000 Python tests at Google.
Google released pyfakefs to the public in 2011 as Google Code project pyfakefs:
- Fork jmcgeheeiv-pyfakefs added direct support for unittest and doctest
- Fork shiffdane-jmcgeheeiv-pyfakefs added further corrections
After the shutdown of Google Code was announced, John McGehee merged all three Google Code projects together here on GitHub where an enthusiastic community actively supports, maintains and extends pyfakefs.