Django test runner not respecting unittest.TestCase? - python

For one of my functional tests, I decided to use unittest.TestCase instead of a Django test class because it was convenient when cleaning up the test to have direct access to my local development database in the test itself.
Running the test in isolation like so passes as I'd expect:
$ python manage.py test functional_tests.test_functionality
System check identified no issues (0 silenced).
...
----------------------------------------------------------------------
Ran 3 tests in 0.040s
OK
When I try to run all tests at the same time, however, that test specifically errors out, complaining that an object DoesNotExist, as though it were using the Django test database:
$ python manage.py test functional_tests
Creating test database for alias 'default'...
System check identified no issues (0 silenced).
..................E..
======================================================================
ERROR: some_functional_test (functional_tests.test_functionality.FunctionalTest)
----------------------------------------------------------------------
Traceback (most recent call last):
... etc.
app.models.Object.DoesNotExist: Object matching query does not exist.
----------------------------------------------------------------------
Ran 21 tests in 0.226s
FAILED (errors=1)
Destroying test database for alias 'default'...
I assume the error is with my trying use Object.objects.latest('created') when no Objects exist in Django's test database.
Is there some way to prevent Django from wrapping all tests in whatever it is about the test runner that prevents my test from accessing an Object directly?

First a little explanation.
By default when you run ./manage.py test django test-runner make few steps which involves creating test database (with test_ prefix to each database name from app settings), running migration and destroying test database (more details on runner steps could find here)
A good explanation on how django treat test database is described here
In your case when you run unittest.TestCase in isolation no test database is created:
$ python manage.py test functional_tests.test_functionality
System check identified no issues (0 silenced).
...
----------------------------------------------------------------------
Ran 3 tests in 0.040s
OK
( ^no logs about creating test database)
That's because there were no django.test.TestCase called. We can see it from sources (virgin unittest.TestCase doesn't have databases property, when django.TestCase it has)
But when you call whole module (python manage.py test functional_tests) looks like you have some django.test.TestCase tests in suite so that's why new test database is creating:
$ python manage.py test functional_tests
Creating test database for alias 'default'... # <-- THIS ONE
##<skipped for readability>
Destroying test database for alias 'default'...
And as you mentioned tests failed because there were no prepared objects for them.
Solution
At this point I see few options to solve this.
Prepare test data for tests explicitly (by fixtures or manually in tests or setups) so they would be independent from current state of database
Explicitly use desired database.
Run tests with --keepdb option (ie ./manage.py test --keepdb, it will use the existing database and will not destroy it after test runs) and set same test database name in app settings as working database (in such case it will not append test_ prefix for test db)
Since you don't want to use django.TestCase, so don't use them at all? Replace them with unittest.TestCase and it will not create test database

Related

How to test django app without giving an app_name

My project structure looks like this: backend/src/apps/app_name/tests/foo.py and when I'm trying to run the tests by entering a command python manage.py test it runs 0 tests. The only correct way to test is python manage.py test src.apps.app_name but I'm to lazy to do it with every app. 😄
Change
backend/src/apps/app_name/tests/foo.py
to
backend/src/apps/app_name/tests/test_foo.py
Make sure all your test files and functions start with test

IntegrityError: Problem installing fixture

I upload my first Django-project into DigitalOcean. After command python manage.py loaddata initial_data.json, I have received this message:
django.db.utils.IntegrityError: Problem installing fixture
'/webapps/django_shop/shop/initial_data.json': Could not load
contenttypes.ContentType(pk=3): duplicate key value violates unique
constraint "django_content_type_app_label_76bd3d3b_uniq" DETAIL: Key
(app_label, model)=(auth, permission) already exists.
How can I fix it?
I had the same problem and I solved this way
DB with data to export from
python manage.py dumpdata --exclude auth.permission --exclude contenttypes > db.json
New DB to import to
python manage.py flush
// Important! Disable all signals on models pre_save and post_save
python manage.py loaddata db.json
// Do not forget to enable all signals that you disabled
It looks like you've generated fixtures that include Django's default data set, i.e. the built-in entries that are inserted normally as part of the first migrate run for some of Django's plumbing data types.
You should review your fixture process, because content type entries will be created automatically when your (and Django's) apps' migrations are run, so they should not be present in fixtures. It's possible there are other tables that will have this same problem, so now would be a good time to make sure you're not including any other data that would result in this situation.

PyCharm can't find Django test

I'm running into this very same issue, but I can't find a solution for this.
Find the dummy Django project I created for this here.
This is my test configuration:
This is the project structure:
Trying to run the test results in:
AttributeError: 'super' object has no attribute 'run_tests'
Process finished with exit code 137
Empty test suite.
Obviously, running this from the shell gives no errors:
(django)mosquito#mosquito-X550LD ~/python_projects/django_test $ python manage.py test
Creating test database for alias 'default'...
.
----------------------------------------------------------------------
Ran 1 test in 0.000s
OK
Destroying test database for alias 'default'...
Any idea what I'm doing wrong??
Thanks,
Alejandro
In the project configuration, add another environment variable of
DJANGO_SETTINGS_MODULE = django_test.settings
You can do this by clicking on the "..." icon

Django can't destroy and create test databases properly

When I try to run my unittest, this is what I get:
python manage.py test dbank --settings=databank_web.settings.dqs.dev_hooman
Creating test database for alias 'default'...
Creating test database for alias 'global'...
Creating test database for alias 'optin_db'...
Creating test database for alias 'vpd3'...
Creating test database for alias 'user_db'...
Creating test database for alias 'vpd1'...
Creating test database for alias 'vpd2'...
.
----------------------------------------------------------------------
Ran 1 test in 0.327s
OK
Destroying test database for alias 'default'...
Warning: Table 'mysql.proc' doesn't exist
It couldn't destroy the database. It gets better, when I rerun the test:
python manage.py test dbank --settings=databank_web.settings.dqs.dev_hooman
Creating test database for alias 'default'...
Creating test database for alias 'global'...
Got an error creating the test database: (1007, "Can't create database 'test_dqs12_full2'; database exists")
Type 'yes' if you would like to try deleting the test database 'test_dqs12_full2', or 'no' to cancel: yes
Destroying old test database 'global'...
Got an error recreating the test database: Table 'mysql.proc' doesn't exist
Any idea why this is going wrong?
Running latest homebrew + mysql-5.6.21 + Django 1.5.5
I have finally found the reason.
Our application has around 7 databases. When I was about to drop them manually to get a clean state, I accidentally also dropped the mysql database. This database should not be ever deleted, as it contains vital information about the root user.
When I realised this, I tried to mitigate by reinstalling MySQL via brew.
brew uninstall mysql
brew install mysql
This worked and mysql database showed up again when I ran show databases;, however the problem persisted. Thats when I came here for help.
As it seems things are a bit more complicated than that with brew.
This is how I got it working again:
1) brew uninstall mysql
2) sudo rm -r /usr/local/var/mysql/
3) brew install mysql
4) unset TMPDIR
5) mysql_install_db --verbose --user='whoami' --basedir="$(brew --prefix mysql)" --datadir=/usr/local/var/mysql --tmpdir=/tmp
6) mysql.server restart
7) mysql_secure_installation
Now it works. Hope this helps other fellow mac users. Thanks.
Well, there are two errors indicating that the mysql.proc table is missing. You might be able to generate it running mysql_update.

How to see which tests were run during Django's manage.py test command

After tests execution is finished using Django's manage.py test command only number of passed tests is printed to the console.
(virtualenv) G:\Project\>python manage.py test
Creating test database for alias 'default'...
True
..
----------------------------------------------------------------------
Ran 2 tests in 0.017s
OK
Destroying test database for alias 'default'...
Is there any way to see:
which tests were actually executed
from what module
in what order
I haven't found any solution in the doc.
You can pass -v 2 to the test command:
python manage.py test -v 2
After running this command you'll get something like this (I'm using django 2, feel free to ignore migrations/database stuff):
Creating test database for alias 'default' ('file:memorydb_default?mode=memory&cache=shared')...
Operations to perform:
Synchronize unmigrated apps: messages, staticfiles
Apply all migrations: admin, auth, contenttypes, sessions
Synchronizing apps without migrations:
Creating tables...
Running deferred SQL...
Running migrations:
Applying contenttypes.0001_initial... OK
...
Applying sessions.0001_initial... OK
System check identified no issues (0 silenced).
test_equal_hard (polls.tests.TestHard) ... ok <--------+
test_equal_simple (polls.tests.TestSimple) ... ok <--------+
|
|
That's your tests! >----------------------------+
By the way, v stands for verbosity (You can also use --verbosity=2):
python manage.py test --verbosity=2
Here's the excerpt from the python manage.py test --help:
-v {0,1,2,3}, --verbosity {0,1,2,3}
Verbosity level; 0=minimal output, 1=normal output,
2=verbose output, 3=very verbose output
Nigel's answer is great and definitely the lowest barrier to entry option. However, you can get even better feedback with django_nose (and it's not that difficult to setup ;).
The below is from: BDD with Python
First: install some requirements:
pip install nose pinocchio django_nose
Then add the following to settings.py
TEST_RUNNER = 'django_nose.NoseTestSuiteRunner'
NOSE_ARGS = ['--with-spec', '--spec-color']
Then run your tests as per normal:
python manage.py test
Output should look something like this:
Note: The comments under your tests can be used to give even better output than just the name.
e.g.:
def test_something(self):
"""Something should happen"""
...
Will output "Something should happen" when running the test.
For extra points: You can also generate / output your code coverage:
pip install coverage
Add the following to your NOSE_ARGS in settings.py: '--with-coverage', '--cover-html', '--cover-package=.', '--cover-html-dir=reports/cover'
e.g.:
NOSE_ARGS = ['--with-spec', '--spec-color',
'--with-coverage', '--cover-html',
'--cover-package=.', '--cover-html-dir=reports/cover']
Then you'll get a nice code-coverage summary when you run python manage.py test as well as a neat html report in reports/cover

Categories

Resources