2013, Journalism++
GNU General Public License
- How to install Spending Stories
- Presentation of the application
- How to customize Spending Stories
- How to translate Spending Stories
- Update stories with last currencies and cpi
This installation guide was created for and tested on Debian/Ubuntu operating systems.
If you find a bug/error in this guide please submit a pull request.
The Spending Stories installation consists of setting up the following components:
- Set up your python environment
- Install dependencies
- Set up the database (example with MySQL)
- Run server (example with mod_wsgi from apache)
- Set up your superuser
a. Install python packages:
sudo apt-get install build-essential git-core python python-pip python-dev
b. Install virtualenv a tool to isolate your dependencies
sudo pip install virtualenv
c. Download the project
git clone https://github.com/jplusplus/okf-spending-stories.git
d. Create the virtualenv folder for this project
Every python dependencies will be installed in this folder to keep your system's environment clean.
cd okf-spending-stories
virtualenv venv --no-site-packages --distribute --prompt=SpendingStories
e. Activate your new virtualenv
source .env
Tips: you can install autoenv to source this file automatically each time you
cd
this folder.
a. Install python modules required
pip install -r requirements_core.txt
b. Install compilers for Less and CoffeeScript
If you don't already have nodejs installed:
sudo apt-get update
sudo apt-get install python-software-properties g++ make
sudo add-apt-repository ppa:chris-lea/node.js
sudo apt-get update
sudo apt-get install nodejs
and then install them
npm install
a. Configure the file webapp/settings.py
DATABASES = {
'default': {
'ENGINE': 'django.db.backends.sqlite3', # Add 'postgresql_psycopg2', 'mysql', 'sqlite3' or 'oracle'.
'NAME' : 'dev.db', # Or path to database file if using sqlite3.
# The following settings are not used with sqlite3:
'USER' : '',
'PASSWORD': '',
'HOST' : '', # Empty for localhost through domain sockets or '127.0.0.1' for localhost through TCP.
'PORT' : '', # Set to empty string for default.
}
}
Currently, the dataset backend is sqlite3, very easy to use for development.
You can change it by mysql, postgres or oracle.
Documentation
For MySQL, you will need to install mysql-python, like that:
sudo apt-get install libmysqlclient-dev
pip install mysql-python
and create the database
mysql -u root -p
mysql> CREATE DATABASE IF NOT EXISTS `<database_name>` DEFAULT CHARACTER SET `utf8` COLLATE `utf8_unicode_ci`;
mysql> \q
**b. Syncing the database **
To create the tables in the database and put some data like categories and currencies, please run and provide asked informations
python manage.py syncdb && python manage.py migrate
a. For development
python manage.py runserver
b. For production
You have a lot a choice to deploy this application.
Take a look on this documentation : Django - deployment
Keep in mind that assets are already served by the wsgi server with dj-static if you're using our wsgi.py file
Dependency:
sudo apt-get install libapache2-mod-wsgi
Please see the documentation about the deployement with mod_wsgi.
This is an Apache configuration that loads the virtualenv :
WSGIScriptAlias / /<PATH_TO_PROJECT>/webapp/wsgi.py
WSGIPythonPath /<PATH_TO_PROJECT>:/<PATH_TO_PROJECT>/venv/lib/python2.6/site-packages:/<PATH_TO_PROJECT>/libs
<Directory /<PATH_TO_PROJECT>/webapp>
<Files wsgi.py>
Order deny,allow
Allow from all
</Files>
</Directory>
In order to add stories you have to configure your user base. If you already synced your database you must have been prompted to configure a superuser.
If you skipped this step then you have to run the following command:
python manage.py createsuperuser
Spending Stories is built with Django, a Python Web framework.
It's composed of :
- a web application
- an API (Available end-points listed on the page
http://<domain_name>/api
) - some scripts to update data (such as CPI data)
HTML pages are located in webapp/templates
Static files (css, javascript, images etc...) are located in webapp/static
Spending Stories uses Twitter Bootstrap for its style.
Spending Stories' title is defined in a translation variable: HEADER_APP_TITLE
. However, if you want to change the HTML title tag, you have to manually modify it in webapp/templates/base.html
.
Colors are defined in less variables. They can be overridden in webapp/static/less/base.less
.
Notable variables:
body-bg
: defines the page backgroundbrand-primary
: defines color for links, buttons and sticky cards' backgroundstrat-main-color
: defines the visualization background colortext-color
: defines the color for most of the texts
Acceptable values are hexadecimal colors, rgb colors or HTML color names:
@body-bg: rgb(80, 80, 80);
@strat-main-color: #403F3F;
@text-color: black;
The Spending Stories translation is focused on the front-end of the application. However we're also using the django default translation system to translate some string that come from the django application (the server) of this project.
The front-end translation system is composed of the following tools:
- angular-translate, makes Angular application translation easy.
- jplusplus/grunt-angular-translate , a fork of grunt-angular-translate. Automates the updates of locales.
This part covers the following topics:
- how to add/edit some translation
- add a supported language
- what is not fully translated for the moment
You have to choose a translation key. We use the following naming convention:
- the key must be in uppercase and underscore (e.g
A_TRANSLATION_KEY
) - the key describes what it translates and where.
- the first part of the key is the name of the script or the template containing the key.
Ex: a key incontribute.html
will start byCONTRIBUTE_
- the last part is the description of the content, or the word if it can be described otherwise.
Ex: the translation key of the 'Compare' button contained in theheader.html
template will beHEADER_COMPARE_BUTTON
- the first part of the key is the name of the script or the template containing the key.
Note: the task for translation key collection is handled by a grunt task and therefore is really limited: it may not detect your translation keys for many reasons because of its design. To avoid that, here's some advice:
- Prefer filter notation over directive notation in your templates files:
<!-- good: --> <a href="">[[ 'MY_TRADUCTION' | translate ]]</a> <!-- not good: --> <a href="" translate>MY_TRADUCTION</a>
- Avoid dynamic keys in your coffescript files and keep the dollar prefix
# good my_string = $translate('GOOD_KEY_IS_GOOD') # bad -> translate service is missing its dollar prefix, it will not detect this key my_string = translate('GOOD_KEY') # bad -> it's a dynamic key, scripts are not evaluated so it will not detect this key either my_key = 'GOOD_KEY_SEEMS_GOOD' my_string = $translate(my_key)
This will produce the collection of every translation keys contained in html, coffee & python files.
This files contains all the translation keys and their translation values, they're located at:
/webapp/static/locales/<locale code>.json
for static application translations/webapp/locale/<locale code>/LC_MESSAGES/django.po
for the Django application translations
If you want to support a new language in your Spending Stories instance you must change Django's settings.
You must update the webapp/settings.py
file and look for the LANGUAGES
variable. This is where we store
the list of supported languages, by default we only support English.
LANGUAGES = (
('en_GB', _('English')),
# add your new language here
)
User query comprehension works only in French & English for the moment.
The way we understand the amounts and currencies entered by user is based on a fuzzy search.
This means all terms that will be recognized have to be entered in an array.
If you look at the getSearchSetData method in the Comprehension
service you can see all translated term in our array.
This way of doing cannot work with German for instance because of the way most numbers are constructed. Fuzzy search can work but the search data set (the data that will be used to perform the search) has to be built by an algorithm because of the tremendous amount of numbers.
It's important to note that the conversion rates & the consumer price indexes we use to deal with inflation need to be refreshed by hand:
python manage.py update_currencies