Djangui has been merged with Wooey! You can get the latest and greatest at the organization Wooey
Djangui is designed to take scripts implemented with a python command line argument parser (such as argparse), and convert them into a web interface.
This project was inspired by how simply and powerfully sandman could expose users to a database. It was also based on my own needs as a data scientist to have a system that could:
1. Autodocument my workflows for data analysis
(simple model saving).
2. Enable fellow lab members with no command line
experience to utilize python scripts.
3. Enable the easy wrapping of any program in simple
python instead of having to use language specific
to existing tools such as Galaxy.pip install django-djanguiThere is a bootstrapper included with djangui, which will create a Django project and setup most of the needed settings automagically for you.
1. djanguify -p ProjectName
2. Follow the instructions at the end of the bootstrapper
to create the admin user and access the admin page
3. Login to the admin page wherever the project is
being hosted (locally this would be localhost:8000/admin)1. Add 'djangui' to INSTALLED_APPS in settings.py (and optionally, djcelery unless you wish to tie into an existing celery instance)
2. Add the following to your urls.py:
`url(r'^', include('djangui.urls')),`
(Note: it does not need to be rooted at your site base,
you can have r'^djangui/'... as your router):
3. Migrate your database:
# Django 1.6 and below:
`./manage.py syncdb`
# Django 1.7 and above
`./manage.py makemigrations`
`./manage.py migrate`
4. Ensure the following are in your TEMPLATE_CONTEXT_PROCSSORS variable:
TEMPLATE_CONTEXT_PROCESSORS = [
...
'django.contrib.auth.context_processors.auth',
'django.core.context_processors.request'
...]Djangui depends on a distributed worker to handle tasks, you can disable this by setting DJANGUI_CELERY to False in your settings, which will allow you to run Djangui through the simple command:
python manage.py runserverHowever, this will cause the server to execute tasks, which will block the site.
The recommended ways to run Djangui are:
The simplest way to run Djangui is to use a Procfile with honcho, which can be installed via pip. Make a file, called Procfile in the root of your project (the same place as manage.py) with the following contents:
web: python manage.py runserver
worker: python manage.py celery worker -c 1 --beat -l info
EOMYour server can then be run by the simple command:
honcho startYou can also run djangui by invoking two commands (you will need a separate process for each):
python manage.py celery worker -c 1 --beat -l info
python manage.py runserverScripts may be added in two ways, through the Django admin interface as well as through the addscript command in manage.py.
Within the django admin interface, scripts may be added to through the 'scripts' model. Here, the user permissions may be set, as well as cosmetic features such as the script's display name, description (if provided, otherwise the script name and description will be automatically populated by the description from argparse if available).
./manage.py addscript
This will add a script or a folder of scripts to Djangui (if a folder is passed instead of a file). By default, scripts will be created in the 'Djangui Scripts' group.
Scripts can be viewed at the root url of Djangui. The ordering of scripts, and groupings of scripts can be altered by changing the 'Script order' or 'Group order' options within the admin.
Scripts and script groups can be relegated to certain groups of users. The 'user groups' option, if set, will restrict script usage to users within selected groups.
Scripts and groups may also be shutoff to all users by unchecked the 'script/group active' option.
DJANGUI_FILE_DIR: String, where the files uploaded by the user will be saved (Default: djangui_files)
DJANGUI_CELERY: Boolean, whether or not celery is enabled. If disabled, tasks will run locally and block execution. (Default: True)
DJANGUI_CELERY_TASKS: String, the name of the celery tasks for Djangui. (Default: 'djangui.tasks')
DJANGUI_ALLOW_ANONYMOUS: Boolean, whether to allow submission of jobs by anonymous users. (Default: True)
By default, Djangui has a basic user account system. It is very basic, and doesn't confirm registrations via email.
DJANGUI_AUTH: Boolean, whether to use the authorization system of Djangui for simple login/registration. (Default: True)
DJANGUI_LOGIN_URL: String, if you have an existing authorization system, the login url: (Default: settings.LOGIN_URL
DJANGUI_REGISTER_URL: String, if you have an existing authorization system, the registration url: (Default: /accounts/register/)
DJANGUI_EPHEMERAL_FILES: Boolean, if your file system changes with each restart. (Default: False)
DJANGUI_SHOW_LOCKED_SCRIPTS: Boolean, whether to show locked scripts as disabled or hide them entirely. (Defalt: True -- show as disabled)
Djangui has been tested on heroku with S3 as a file storage system. Settings for this can be seen in the user_settings.py, which give you a starting point for a non-local server. In short, you need to change your storage settings like such:
STATICFILES_STORAGE = DEFAULT_FILE_STORAGE = 'djangui.djanguistorage.CachedS3BotoStorage' DJANGUI_EPHEMERAL_FILES = True
The easiest way to make your scripts compatible with Djangui is to define your ArgParse class in the global scope. For instance:
import argparse
import sys
parser = argparse.ArgumentParser(description="https://projecteuler.net/problem=1 -- Find the sum of all the multiples of 3 or 5 below a number.")
parser.add_argument('--below', help='The number to find the sum of multiples below.', type=int, default=1000)
def main():
args = parser.parse_args()
...
if __name__ == "__main__":
sys.exit(main())
If you have failing scripts, please open an issue with their contents so I can handle cases as they appear and try to make this as all-encompasing as possible. One known area which fails currently is defining your argparse instance inside the if __name__ == "__main__" block