To place widgets in your dashboard you need create a javascript file, where you call each widget that you need to place with the correct options, each widget provide an events that you can call in any javascript file to update the data.
For example if you create a number widget
var dashboard = new Dashboard();
...
dashboard.addWidget('current_valuation_widget', 'Number', {
getData: function () {
$.extend(this.data, {
title: 'Current Valuation',
moreInfo: 'In billions',
updatedAt: 'Last updated at 14:10',
value: '$35',
detail: '64%'
});
}
});
Then you can publish in any moment the event dashboard.publish('example_widget/getData') to get new data and update the widget.
Note that in this example the getData method will be executed each 1000 milliseconds because is the default value of interval option in a Number widget.
This widget can display an specific day an hour.
JSON object that represent the date and time in format
{
time: 'hh:mm:ss',
date: 'Month Day DD sYYYY'
}
This widget can display a value with an associate graph as background.
JSON object in this format
{
data: [
{x: /x0/, y: /y0/},
{x: /x1/, y: /y1/}
...
],
value: /string/
title: /string/,
moreInfo: /string/
}
To render the graph this widget use Rickshaw library, for now the config options are quite limited, if you need be more specific you can overwrite the rivetsjs binder (rv-dashing-graph) or write a custom widget use this as guide.
To configure the X and Y axis must be passed as extra aparameters in the data xFormat and yFormat methods, also you can use the methods beforeRender and afterRender to execute arbitrary javascript before or after of render, for example:
var xFormat = function(n) {
return '(' + n + ')';
};
$.get('/my/api/url/', function(data) {
data.data.xFormat = xFormat;
data.data.afterRender = function() {
alert('graph shown');
};
$.extend(self.data, data);
});
This class helps to return valid data to be use by the widget, you can see the definition in GitHub
Here’s an example of a graph widget where in value is displayed the total of Errands and in data is returned an array with the last two hour of activity
from dashing.widgets import GraphWidget
class HourlyErrandsWidget(GraphWidget):
title = 'Hourly Errands'
moreInfo = ''
def get_value(self):
return SearchQuerySet().filter(django_ct='errands.errand').count()
def get_data(self):
latest_hours = datetime.now() - timedelta(hours=2)
latest_errands = SearchQuerySet().filter(
django_ct='errands.errand',
created__gt=latest_hours).values('created')
intervals = []
for errand in latest_errands:
delta = datetime.now() - errand['created']
for m in range(10, 120, 10):
if delta < timedelta(minutes=m):
intervals.append(13 - m/10)
break
rlist = Counter([x for x in intervals])
return [{'x': x, 'y': y} for x, y in rlist.most_common()]
This widget can display a list of elements with an associate value.
JSON object in this format
{
data: [
{
label: /string/,
name: /string/
},
{
label: /string/,
name: /string/
},
...
],
title: /string/,
moreInfo: /string/,
updatedAt: /string/
}
This class helps to return valid data to be use by the widget, you can see the definition in GitHub
Here’s an example of a graph widget where in data returns an array with the messengers who have more requests
from dashing.widgets import ListWidget
class ActiveMessengersWidget(ListWidget):
title = 'Active Messengers'
moreInfo = 'Those who have more requests'
def get_updated_at(self):
modified = SearchQuerySet().filter(
django_ct='errand').order_by('-modified')[0].modified
return u'Last updated {}'.format(modified)
def get_data(self):
messengers = SearchQuerySet().filter(
django_ct='messengers', active=True)
rlist = Counter([x for x in messengers])
return [{'label':x, 'value':y} for x, y in rlist.most_common(20)]
This widget can display a value with another interesting information.
JSON object in this format
{
value: /string/,
title: /string/,
detail: /string/,
moreInfo: /string/,
updatedAt: /string/
}
This class helps to return valid data to be use by the widget, you can see the definition in GitHub
Here’s an example of a graph widget where in value is displayed the total of payments and in the detail and moreInfo shows other information of interest
from dashing.widgets import NumberWidget
class PaymentsWidget(NumberWidget):
title = 'Payments Customers'
def get_value(self):
return Payment.objects.all().count()
def get_detail(self):
payments = Payment.objects.all()
total = len([x for x in payments if x.status == Payment.STATUS.waiting])
return '{} to approve'.format(total)
def get_more_info(self):
payments = Payment.objects.all()
total = len([x for x in payments if x.status == Payment.STATUS.rejected])
return '{} rejected'.format(total)
To make a custom widget you must create three static files to define configuration parameters and appearance, in addition, you can create a python class to communicate with the Django project.
To name your widgets should follow a naming convention, where the name must by unique for findable through the settings.
Its location should be <static_directory>/widgets/<widget_name>/<widget_name>.html this file describes its UI in plain HTML using the Rivets.js conventions to bind data to the script file.
For example {% static %}widgets/list/list.html looks like this:
<div>
<h1>{ data.title }</h1>
<ul>
<li rv-each-el="data.data">
<span class="label">{ el.label }</span>
<span class="value">{ el.value }</span>
</li>
</ul>
<p class="more-info">{ data.moreInfo }</p>
<p class="updated-at">{ data.updatedAt }</p>
</div>
The classes are only for the stylesheet.
Your location should be <static_directory>/widgets/<widget_name>.css in this file defines the styles of widget.
Your location should be <static_directory>/widgets/<widget_name>.js in this file will be defined the configuration options and default values for the new widget, the idea is to create an object using the new keyword, then we define properties and methods using this keyword.
We must provide an __init__ method where binding the data with the template and add to the dashboard, this function is quite similar in all widgets, then it is provided by Dashing.utils.widgetInit to facilitate implementation and improve reading of widgets, also must provide a data element which will be binded to the template, and a getData function will surely be the to be overwritten to obtain relevant data as required,
For example {% static %}widgets/list/list.js looks like this:
/* global Dashboard */
Dashing.widgets.List = function (dashboard) {
var self = this,
widget;
this.__init__ = Dashing.utils.widgetInit(dashboard, 'list');
this.row = 2;
this.col = 1;
this.data = {};
this.getWidget = function () {
return widget;
};
this.getData = function () {};
this.interval = 10000;
};
if we want to initialize widget with data we can write:
...
this.col = 1;
this.data = {
title: 'Default Title',
moreInfo: 'No data to display'
};
this.getWidget = function () {
...
Surely in many cases may be necessary give the option to get some Dajngo project data into the widget, for this dashing has a Widget class that can be inherited to deliver properly serialized data, also subsequently can be serve the data using the dashing router.
For example ListWidget in dashing/widgets.py looks like this:
class ListWidget(Widget):
title = ''
more_info = ''
updated_at = ''
data = []
def get_title(self):
return self.title
def get_more_info(self):
return self.more_info
def get_updated_at(self):
return self.updated_at
def get_data(self):
return self.data
def get_context(self):
return {
'title': self.get_title(),
'moreInfo': self.get_more_info(),
'updatedAt': self.get_updated_at(),
'data': self.get_data(),
}
If you develop your widget with python classes necessarily going to have to distribute it via PyPI
To distribute a widget you have two options, the fastest way is throught Django Dashing Channel but is a bit limited, and through PyPI a bit trickier to pack but you have more options when developing the widget.
Using this distribution method the users will only have to add the widget name on INSTALLED_WIDGETS then to loading the dashboard, this locates the static files from a remote location (specified in the preconfigured repository), if the user creates a copy of the files on your local static directory then these will open locally.
You will have to host your files into a CDN, I recommend creating a github project and use RawGit to serve through MaxCDN, you can take dj-dashing-weather-widget project as a guide.
Finally to publish your widget in Django Dashing Channel you need to make a fork of django-dashing-channel, add your repository to repositories.json and send a pull request. In the repository root will be sought the widget static files (.js .css and .html)
You should create a README file for installations instructions.
If your widget requires python code or just want to provide an easy way to get the widget locally then a PyPI package is the way to go.
As a requirement is necessary follow the widgets naming convention (see static files). To create a PyPI package see the documentation, and should create a README file for installations instructions.
This not excluding the previous way, you could create a minimalist version of your widget and upload to django-dashing-channel and in the project instructions leave on how to install the PyPI version