Generate Random Test Data.
These are just a bunch of handy functions designed to make it easier to test your code.
To use testdata in your tests, just import it into your testing module:
import testdata
To install, use Pip:
pip install testdata
Or, with Pip using Github:
pip install --upgrade "git+https://github.com/Jaymon/testdata#egg=testdata"
The easiest way to use testdata is to use testdata.TestCase
while writing your unittest tests:
from testdata import TestCase
import testdata
class CustomTest(TestCase):
def test_foo(self):
# use testdata methods by magic:
print(self.get_int())
# or, be more explicit and use the .data attribute:
print(self.data.get_int())
# finally, you can call anything at the module level also:
print(testdata.get_int())
To use testdata with other testing frameworks, or outside of tests, just import the module and use it:
import testdata
testdata.get_int()
The testdata module has a ton of helpful methods, and you can also easily add your own custom methods by extending testdata.TestData
:
from testdata import TestData, NumberData
class CustomData(TestData):
"""Bespoke methods"""
def foobar(self):
return "foobar"
class CustomNumberData(NumberData):
"""overrides default get_int functionality"""
def get_int(self, *args, **kwargs):
return 1
testdata.foobar() # foobar
testdata.get_int() # 1
The only catch to adding/overriding testdata methods is you have to make sure your custom TestData
child classes are imported before you can call them since testdata makes use of init_subclass to add TestData
classes to the resolver.
This is an overview of some of the functions and classes found in the Testdata module, there are other functions (like get_birthday
) that aren't listed here, for the complete list just look at the source. Any methods on any child class that extends testdata.base.TestData
will be available at testdata.<METHOD-NAME>
.
patch(mod, **patches)
Patches a module, instance, or class with the given patches.
Suppose you had a module like this:
# module foo.bar
def boom():
return 1
class FooPatch(object):
@classmethod
def bam(cls): return boom()
Now you can easily patch it for testing:
def mock_boom():
return 2
foo_bar = testdata.patch('foo.bar', boom=mock_boom)
print foo_bar.FooPatch.bam() # 2
# but you can also just pass in objects or modules
from foo.bar import FooPatch
FooPatch = testdata.patch(FooPatch, boom=mock_boom)
print FooPatch.bam() # 2
from foo import bar
bar = testdata.patch(bar, boom=mock_boom)
print bar.FooPatch.bam() # 2
You can also patch a specific instance
Suppose you had a module like this:
# module foo.bar
class Foo(object):
def boom(self): return 1
Now you can easily patch it for testing:
def mock_boom():
return 2
foo = Foo()
foo_patched = testdata.patch(foo, boom=mock_boom)
print foo_patched.boom() # 2
# be aware though, the original instance was modified, foo_patched == foo
print foo.boom() # 2
Run a command on the command line
r = testdata.run("echo 1")
print(r) # 1
Request a url
r = testdata.fetch("http://example.com")
print(r.code) # 200
print(r.body) # the html body of example.com
Output buffering, handy when you want to make sure logging or print statements are doing what you think they should be doing.
with testdata.capture() as c:
print("foo")
if "foo" in c:
print("foo was captured")
A wrapper around python's builtin threading.Thread
class that bubbles errors up to the main thread because, by default, python's threading classes suppress errors, this makes it annoying when using threads for testing. NOTE - This is buggier than I would like.
def run():
raise ValueError("join_2")
thread = testdata.Thread(target=run)
thread.start()
print(thread.exception)
Sometimes you need to test fetching remote files
import requests
server = testdata.create_fileserver({
"foo.txt": ["foo"],
"bar.txt": ["bar"],
})
with server: # the with handles starting and stopping the server
res = testdata.fetch(server.url("foo.txt"))
print(res.body) # foo
Change your environment with this context manager, if you don't pass in an object as the first value it will default to os.environ
with testdata.enviroment(FOO=1):
print(os.environ["FOO"]) # 1
print(os.environ["FOO"]) # raises KeyError
# you can also modify objects:
d = {}
with testdata.enviroment(d, FOO=1):
print(d["FOO"]) # 1
print(d["FOO"]) # raises KeyError
create_dir(path="", tmpdir="")
create a directory hierarchy
base_dir = "/tmp"
d = testdata.create_dir("/foo/bar", base_dir)
print d # /tmp/foo/bar
create_dirs(dirs, tmpdir="")
Create a bunch of files and folders
testdata.create_dirs({
"foo": {
"bar": {
"che.txt": ["line 1", "line 2"],
}
}
})
create_file(data="", path="", tmpdir="", encoding="")
create a file with contents
base_dir = "/tmp"
f = testdata.create_file(path="/foo/bar.txt", data="The file contents", tmpdir=base_dir)
print f # /tmp/foo/bar.txt
create_files(file_dict, tmpdir="")
Create a whole bunch of files, the file_dict
key is the filename, the value is the contents of the file.
The file_dict
is very similar to the create_modules
param module_dict
file_dict = {
"foo/bar.txt": "the foo file contents",
"baz.txt": "the baz file contents",
}
f = testdata.create_files(file_dict)
get_file(path="", tmpdir="")
This will return a Filepath
instance that you can manipulate but unlike create_file
it won't actually create the file, just give you a path to a file that could be created.
create_module(data="", modpath="", tmpdir="", make_importable=True)
create a module with python contents that can be imported
base_dir = "/tmp"
f = testdata.create_module(modpath="foo.bar", data="class Che(object): pass", tmpdir=base_dir)
print f # /tmp/foo/bar.py
create_modules(module_dict, tmpdir="", make_importable=True)
create a whole bunch of modules at once
f = testdata.create_modules(
{
"foo.bar": "class Che(object): pass",
"foo.bar.baz": "class Boom(object): pass",
"foo.che": "class Bam(object): pass",
}
)
get_ascii(str_size=0)
return a string of ascii characters
>>> testdata.get_ascii()
u'IFUKzVAauqgyRY6OV'
get_md5(val="")
return an md5 hash of val (if passed in) or a random val if val is empty
>>> testdata.get_md5()
'e165765400b30772f1d9b3975ce77320'
get_hash(str_size=32)
return a random hash
>>> testdata.get_hash()
"jYw3HseUl8GLoMc8QejLYFogC2lUYoUs"
get_bool()
return a boolean (either True or False)
>>> testdata.get_bool()
False
>>> testdata.get_bool()
True
get_float(min_size=None, max_size=None)
return a floating point number between min_size
and max_size
.
>>> testdata.get_float()
2.932229899095845e+307
get_int(min_size=1, max_size=sys.maxsize)
return an integer between min_size
and max_size
.
>>> testdata.get_int()
3820706953806377295
get_name(name_count=2, as_str=True)
returns a random name that can be outside the ascii range (eg, name can be unicode)
>>> testdata.get_name()
u'jamel clarke-cabrera'
get_email(name=u'')
returns a random email address in the ascii range.
>>> testdata.get_email()
u'shelley@gmail.com'
get_str(str_size=0, chars=None)
return random characters, which can be unicode.
>>> testdata.get_str()
"q\x0bwZ\u79755\ud077\u027aYm\ud0d8JK\x07\U0010df418tx\x16"
get_url()
return a random url.
>>> testdata.get_url()
u'https://sK6rxrCa626TkQddTyf.com'
get_words(word_count=0, as_str=True)
return a random amount of words, which can be unicode.
>>> testdata.get_words()
"\u043f\u043e\u043d\u044f\u0442\u044c \u043c\u043e\u0436\u043d\u043e felis, habitasse ultrices Nam \u0436\u0435\u043d\u0430"
get_past_datetime([now])
return a datetime guaranteed to be in the past from now
>>> testdata.get_past_datetime()
datetime.datetime(2000, 4, 2, 13, 40, 11, 133351)
get_future_datetime([now])
return a datetime guaranteed to be in the future from now
>>> testdata.get_future_datetime()
datetime.datetime(2017, 8, 3, 15, 54, 58, 670249)
get_between_datetime(start[, stop])
return a datetime guaranteed to be in the future from start
and in the past from stop
>>> start = datetime.datetime.utcnow() - datetime.timedelta(days=100)
>>> testdata.get_between_datetime(start)
datetime.datetime(2017, 8, 3, 15, 54, 58, 670249)
Testing on MacOS:
$ python -m unittest testdata_test
Development needs datatypes on the path. This is kind of a strange thing because datatypes depends on testdata
for testing. Making datatypes
available to testdata
for development should be as easy as:
export PYTHONPATH=$PYTHONPATH:/path/to/dir/containing/datatypes