I have a Django 1.6.6 project that is not loading fixtures from relative paths using Python 2.7.8 on Windows 8.1. Works just fine in Linux.
Example Test Case:
# cart/tests.py
class CartTestBase(TestCase):
fixtures = ['products/fixtures/product_categories.json',
'products/fixtures/brands.json', 'products/fixtures/products.json']
Which fits into a directory structure of:
ecomm_app/
cart/
__init__.py
models.py
tests.py
. . .
products/
fixtures/
products/fixtures/product_categories.json
products/fixtures/brands.json
products/fixtures/products.json
__init__.py
models.py
. . .
The specific error message is:
UserWarning: No fixture named 'products/fixtures/products' found.
warnings.warn("No fixture named '%s' found." % fixture_name)
The corresponding app modules are listed in INSTALLED_APPS and work fine otherwise. Specifying an absolute path does work, but obviously isn't cross-platform. Specifying a root-relative path does not work, e.g.: /products/fixtures/brands.json
What's going on with these fixtures?
So, the problem was that the path to the fixtures was not being normalized correctly in Windows. I submitted a pull request that was accepted, so this will hopefully go out in the next version of Django after 1.7.1
Related
This is the first time I need to run my app deeper in the directory structure, I have obviously messed up my app structure while setting this up and can't get it right anymore. Could somebody help me out here?
In my start.sh file I define following env vars:
export PYTHONPATH=${PYTHONPATH}:'/home/ubuntu/workspace/skw/3rd_apps/'
export PYTHONPATH=${PYTHONPATH}:'/home/ubuntu/workspace/skw/cskw/'
And my dir structure looks like this:
/home/ubuntu/workspace/skw/start.sh (also manage.py and gulpfile)
/home/ubuntu/workspace/skw/3rd_apps/ (django is here)
/home/ubuntu/workspace/skw/cskw/ (contains __init__.py)
/home/ubuntu/workspace/skw/cskw/apps/ (contains __init__.py)
/home/ubuntu/workspace/skw/cskw/core_app/ (settings.py is here and __init__.py)
I cycle between two errors here:
1. When trying to setup a custom signup form:
ACCOUNT_SIGNUP_FORM_CLASS = 'core_app.forms.SignupForm'
The form is in place but i get this error:
django.core.exceptions.ImproperlyConfigured: Module "core_app.forms" does
not define a "SignupForm" class
And if I fix this I can not import anything from apps, it expects them to be in core_app and does not see the app module at the same level as core_app.
manage.py points to "core_app.settings", so do the variables in settings.py.
Background:
Apparently my question regarding the best ways to build an autoloader in Python was too broad (https://stackoverflow.com/questions/25678974/python-autoloader). Not sure why...
I had hoped to get a few code snippets to make my life easier with regards to choosing between importing modules or building a package. But not so much. So I decided to try to build a package of python functions meant to emulate PHP functions such str_replace.
I've named the package p2p and the directory structure is like this:
lib/p2p
setup.py
CHANGES.txt
MANIFEST.IN
README.txt
/bin (empty)
/docs (empty)
/p2p
__init__.py
/str
__init__.py
str.py
/time
__init__.py
time.py
/json
__init__.py
json.py
/utils
__init__.py
utils.py
The __init__.py are all empty.
In the Console:
Now when I build the package (python setup.py install) I get no errors. When I import p2p I get no errors. However, when I try to use p2p.time.timeID() I get:
AttributeError: 'module' object has no attribute 'timeID'
When I type p2p.time I get:
module 'p2p.time' from '/home/gabe/anaconda/lib/python2.7/site-packages/p2p/time/init.pyc'
I've Googled the attribute error and spent most of yesterday reading up on how to create a python package. What am I doing wrong? (Ultimately this appears like it would have been easier with simple modules and an autoloader. I'm not sure how to build an autoloader that will properly import the modules, due to relative path restrictions in Python, without just dumping my entire library directory structure into the path.)
lib/p2p/setup.py
from distutils.core import setup
setup(
name='p2p',
version='0.1.0',
author='Me',
author_email='me#example.com',
packages=['p2p', 'p2p.test', 'p2p.str', 'p2p.time', 'p2p.json', 'p2p.utils'],
package_dir={'p2p':'p2p', 'str':'p2p/str'},
scripts=[],
url='',
license='LICENSE.txt',
description='PHP to Python functions.',
long_description=open('README.txt').read(),
install_requires=[
"",
],
)
lib/p2p/p2p/time/time.py
import time
# http://www.php2python.com/wiki/function.microtime/
def timeID():
return time.time()
Notes:
Ubuntu 14.04 Dekstop
Anaconda with Python 2.7
Spyder
I'm a complete Python Noob but have worked in PHP and Ruby for a couple of years now
p2p/time/time.py
Well there you go. You want p2p.time.time.timeID instead. Or you could just put it in p2p/time/__init__.py instead.
I know this question has already been asked multiple times but I still can't seem to find a solution that works for me. My Django project folder looks like this:
my_project
__init__.py
manage.py
my_first_app
my_second_app
core
Now the "core" folder looks like this:
__init__.py
some_other_stuff.py
settings
__init__.py
prod.py
dev.py
local.py -> dev.py
local.py is a symbolic link pointing to the right settings file, dev.py on my machine, prod.py in production.
Here's my problem: when I try to use manage.py, I get a weird error ImproperlyConfigured: The SECRET_KEY setting must not be empty. When I pass the path to the settings file local.py as an argument (--settings=core.settings.local) it runs fine. I figured the problem was that Django didn't know where to look for the settings file. How can I tell him (her?) where to look?
I already tried exporting the path to the env (export DJANGO_SETTINGS_MODULE=core.settings.local) and setting the PYTHONPATH to the parent directory, to no avail.
The primary use of __init__.py is to initialize Python packages. The easiest way to demonstrate this is to take a look at the structure of a standard Python module.
package/
__init__.py
file1.py
file2.py
As you can see in the structure above the inclusion of the __init__.py file in a directory indicates to the Python interpreter that the directory should be treated like a Python package
__init__.py can be an empty file but it is often used to perform setup needed for the package(import things, load things into path, etc).
One common thing to do in your __init__.py is to import selected Classes, functions, etc into the package level so they can be convieniently imported from the package.
In our example above we can say that file.py has the Class File. So without anything in our __init__.py you would import with this syntax:
from package.file import File
However you can import File into your __init__.py to make it available at the package level:
# in your __init__.py
from file1 import File
# now import File from package
from package import File
Source
So for conclusion, when you call import in __init__.py in a package mypackage, it's like you use package as a simple python file, that's what my solution do:
from .local import * in __init__.py
I haven't use this before in the settings case but I use it when I wanna to subdivide my models in a Django app, models.py --> models package, ./manage syndb doesn't discover my models declared, I found so this solution that's similar. You can find more details here
Last thing, I'm sure there's others solution to your problem, but this can be the most simple.
Good luck
You are in import hell somewhere. Had this problem too one time. The only way to find out where the root of your problem is, might be to disable all apps, try starting the server, enable the first app, start the server, enable the next etc.
BTW: your project layout should not be used from Django 1.4 onward. https://docs.djangoproject.com/en/dev/releases/1.4/#updated-default-project-layout-and-manage-py
I'd try to use the new layout and hope that it 'just works'.
I think you need to change the name of the file that the manage.py is looking for.
try:
imp.find_module('settings') # Assumed to be in the same directory.
except ImportError:
import sys
sys.stderr.write("Error: Can't find the file 'settings.py' in the directory containing %r. It appears you've customized things.\nYou'll have to run django-admin.py, passing it your settings module.\n" % __file__)
sys.exit(1)
If you had the settings.py file in the same directory simple changing the 'settings' to 'local' would have worked.
But, since you have it in a different directory, I think you need to configure the settings. Refer to this: https://docs.djangoproject.com/en/dev/topics/settings/#using-settings-without-setting-django-settings-module
from django.conf import settings
settings.configure(DEBUG=True, TEMPLATE_DEBUG=True,
TEMPLATE_DIRS=('/home/web-apps/myapp', '/home/web-apps/base'))
Hope that helps.
I have a pycharm project and, presumably, a Django project. Perhaps they are one and the same, perhaps not - I'm unsure as to the distinction.
Anyway, in my settings.py file (which is in the root of project directory, which is what I presume is my pycharm project) I have:
ROOT_URLCONF = 'dumpstown.urls'
Does that mean dumpstown is my project name? Or my pycharm project name? What is the distinction? Because I also have a folder called dumpstownapp and this has all my models.py and view.py files. I would have thought that dumpstownapp was the Django project, but I really don't know!
So, to be concise:
In this folder setup
folderA
folderB
views.py
models.py
<other stuff>
templates folder
settings.py
<other stuff>
which is the "Django project name" ~ and by that I mean, if I have a UserProfile defined in my models.py (shown above) what would be the AUTH_PROFILE_MODULE entry I'd need for it? I'm getting several understandings from the django docs - I'd assume
dumpstownapp.models.UserProfile
But from the docs I'd get
dumpstownapp.UserProfile
Or maybe my app is called dumpstown? and then what do I get?
FolderA is the Django project folder, and folderB is the Django app folder.
I haven't used a UserProfile, but according to the docs ( https://docs.djangoproject.com/en/dev/topics/auth/#storing-additional-information-about-users) it should contain a string containing the Django app name and the model name separated by a dot, so "dumpstownapp.UserProfile" should be correct. If you show the error someone can probably give you better help..
Django documentation used to say that the parent of the project folder (i.e. the parent of folderA) should be on the path, but I believe that has been changed to also include the project folder itself (i.e. folderA) -- since it makes sharing of Django apps much easier. PyCharm seems to assume that is the case, since pressing Alt+F7 to auto-add an import for a newly used module create an import statement that assumes folderA is on the import path (I'm a relative newcomer to PyCharm, and I'm using it on a project that started in the Django 0.96 era, so I might just have things set up wrong..) But folderA is both the Django and the PyCharm project (the .idea file is where PyCharm stores its project data).
In one of my django-app git-submodule projects I needed to find out the name of the Django project that django-app/library was used in. To that end, I tried to get the path of the file that was being executed, find my package in the path and take the folder above it. However, it turned out that on the production server the project was deployed in a folder with a different name (a standard name like www or something along those lines). So this way is not fully reliable.
So I ended up setting a PROJECT variable in the django settings file and using that instead.
I have a Django app that I've set up using Buildout laid out like so:
/workspace
/bin
/src
/myproject
settings.py
/myapp
views.py
...
bootstrap.py
buildout.cfg
setup.py
The issue is that I'd like both myproject.settings and myapp on the python path. I need the myproject.settings on the path so djangorecipe can import it. And I'd like myapp on the path so that I don't have to write import myproject.myapp all the time.
For now I've got both /workspace/src and /workspace/src/myproject in the Python path, but this feels like a hack and practically makes me worried if there might be situations where import some_module might have confusing resolution patterns because I have two directories that are parent-child to each other.
So questions are:
Is there an accepted way to lay this out?
Is it actually bad to have a directory and one of its sub-directories in the path?
There is no problem, on import some_module importer will search in each folder specified at sys.path for some_module/__init__.py and some_module.py. Same for import myproject.some_module, it will search for myproject module, then it will try to find in it some_module with same algorithm.
I'm using the same project structure.
If your buildout.cfg includes develop = . and whatever egg your setup.py defines is included as a dependency for your buildout/parts then whatever code path your setup.py defines will be automatically added to sys.path. Just make sure your setup.py includes src as a code directory. One way to do this is with:
setup(name=...
...
packages=find_packages('src'),
package_dir = {'':'src'},
...