Pylint-django worked just fine up to version 2.3.x but since 2.4.0 it reports an error on every python-django file:
Django was not configured. For more information runpylint --load-plugins=pylint_django --help-msg=django-not-configuredpylint(django-not-configured)
This happens on VSCode and I believe I have it correctly configured:
{
"python.linting.pylintArgs": [
"--load-plugins",
"pylint_django",
"--load-plugins",
"pylint_django.checkers.migrations",
"--disable=C0114, C0115, W0222",
"--disable=imported-auth-user",
"--disable=invalid-name",
"--disable=line-too-long"
]
}
This worked perfectly fine, as I said, up to v.2.3.
I raised an issue on their repository but regrettably it was dismissed with little to no visible effort to address it.
For the time being I'm staying with v.2.3.0, which lints with no issues with the above configuration, but would like to know if this is a bug or otherwise.
Did any get into this issue or is it there anything else I'm missing?
Note:
The error message can be hid by adding this value in VSCode's settings.json:
{
"python.linting.pylintArgs": [
[...]
"--disable=django-not-configured",
]
}
But I'm aware this is sweeping the dust under the carpet.
Add the --django-settings-module argument the VS Code settings:
{
"python.linting.pylintArgs": [
[...]
"--disable=django-not-configured",
"--django-settings-module=<mainapp>.settings",
]
}
Change <mainapp> to be your main app directory. For example, if your settings.py was in sweetstuff/settings.py then the argument value would be sweetstuff.settings. This is the same format as you would import the settings module from inside a Python module or the Django shell.
This problem came up for me in some Bitbucket Pipelines. I solved it there by creating the DJANGO_SETTINGS_MODULE as a repository variable which is then made available as an environment variable when Pylint is run.
If you want to use a .pylintrc file add the following to to your .pylintrc file
Note: replace myproject with the name of your project (the root directory of your Django project.
[MASTER]
load-plugins=pylint_django
django-settings-module=myproject.settings
if you don't have one you can create one here's an example
[MASTER]
load-plugins=pylint_django
django-settings-module=myproject.settings
[FORMAT]
max-line-length=120
[MESSAGES CONTROL]
disable=missing-docstring,invalid-name
[DESIGN]
max-parents=13
[TYPECHECK]
generated-members=REQUEST,acl_users,aq_parent,"[a-zA-Z]+_set{1,2}",save,delete
If you're using a seperate settings file for local vs productin then use the relevant setting file ex:
django-settings-module=myproject.settings.local
this hit me too and I think I have a solution. if you run in terminal, as suggested by the error message, with the parameter --help-msg=django-not-configured you'll end up with the following message:
Finding foreign-key relationships from strings in pylint-django requires
configuring Django. This can be done via the DJANGO_SETTINGS_MODULE
environment variable or the pylint option django-settings-module, eg: `pylint
--load-plugins=pylint_django --django-settings-module=myproject.settings` .
This can also be set as an option in a .pylintrc configuration file. Some
basic default settings were used, however this will lead to less accurate
linting. Consider passing in an explicit Django configuration file to match
your project to improve accuracy. This message belongs to the django foreign
keys referenced by strings checker.
so to solve this you can either pass the needed settings module via command line argument, in a .pylintrc file or set (export) your environment variable.
hope it helped.