Deptest is a testing framework to handle situation when your need to control the execution order of the test units. Seriously, deptest does not follow the rules of unit testing, in other words, using this tool means you are thinking againest the philosophy of unit testing: “to isolate each part of the program and show that the individual parts are correct”.
But so what? Programming needs diversity, so does testing methodology. If the situation really exists, we should do something with it, that's why deptest is created, it could be considered as a different approach to organize your tests. Try it if you are stuck with unit testing, maybe it'll be helpful :)
pip install deptest
The core part of using deptest is to use depend_on
decorator on your test functions. depend_on
describes that a test function should be run if and
only if its dependency function is OK
. If dependency is FAILED
, then the
test function will not be executed and the status will be set to UNMET
.
Case 1, simple dependency
from deptest import depend_on
@depend_on('test_b')
def test_a():
print 'a, depend on a'
def test_b():
print 'b'
This will ensure test_a
run after test_b
even though test_a
is defined before test_b
.
Case 2, passing return value
from deptest import depend_on
@depend_on('test_b', with_return=True)
def test_a(name):
print 'a, depend on', name
def test_b():
print 'b'
return 'b'
With with_return
argument set to True
, the return value of test_b
will be passed into test_a
. By default return values of dependencies
won't be passed.
Case 3, complicated dependencies
from deptest import depend_on
@depend_on('test_c', with_return=True)
@depend_on('test_b', with_return=True)
def test_a(name1, name2):
print 'a, depend on', name1, name2
return 'a'
@depend_on('test_d')
def test_b():
print 'b'
return 'b'
@depend_on('test_d')
def test_c():
print 'c'
return 'c'
def test_d():
print 'd'
return 'd'
The dependent graph of the four functions will be:
d
| \
b c
| /
a
Thus the execute sequence will be d, b, c, a
or d, c, b, a
, the results are fairly the same.
$ deptest -s test/simple_test.py
d
→ simple_test.test_d... OK
b
→ simple_test.test_b... OK
c
→ simple_test.test_c... OK
a, depend on b c
→ simple_test.test_a... OK
______________________________________________________________________
Ran 4 tests, OK 4, FAILED 0, UNMET 0
You can see some practical examples in examples/
folder,
It's worth mentioning that http_api_test.py
simulates an HTTP API testing case, which is mostly the reason why I develop this tool.
Note: to run
http_api_test.py
, you need HTTPretty installed.
Deptest provides a cli command also called deptest
, it supports some common
arguments of nosetests
, like -s
and --nocapture
, see detail usage by deptest -h
:
usage: deptest [-h] [-s] [--nologcapture] [--dry] [--debug] [PATH [PATH ...]]
positional arguments:
PATH files or dirs to scan
optional arguments:
-h, --help show this help message and exit
-s, --nocapture Don't capture stdout (any stdout output will be printed
immediately)
--nologcapture Don't capture logging
--dry Dry run, only show matched files
--debug Set logging level to debug for deptest logger
See it in action, run deptest examples
:
With --nologcapture
argument: