django-adminstats 0.7.4

An adminstrative charting tool for statistics in Django

Django Admin Stats allows you to create and display charts of your data using the django admin. It uses c3 to display charts.

A sample chart from the demo being visualized

Features

• Easy-no configuration charting using the “explore” function.

• Supports generating statistics from django models and from trackstats metrics.

• Also allows for custom statistics generation by making your own Registration subclass.

• Nice JavaScript charts with c3, falls back to a plain table without JavaScript.

• Add filters & group data by setting query parameters on the criteria

Limitations

• One dimension/axis of the chart is always the date. There’s no way to specify a chart that isn’t “by date”.

Installation

Installation is straightforward. Install django-adminstats with pip, and then add the following INSTALLED_APPS in your config. If you already have something that replaces django.contrib.admin, you’ll need to manually add links to the “Explore” pages if you want to use the Explore tool.

INSTALLED_APPS = [
    ...
    'django_adminstats.apps.AdminConfig', # replaces django.contrib.admin
    'django_adminstats',
    ...
]

Using the Explore Tool

The Explore tool lets any user with access to the admin site view charts for whatever models they have access to. Unlike the following Charts system, these charts are generated on-the-fly, and don’t require any setup.

The explore tool is closely integrated into the functionality that already exists in the django admin. The consequence of this is that if you want to do any grouping or filtering you have to define filters on your admin pages using ModelAdmin.list_filter.

One of the benefits of the Explore tool is that once you’ve applied filters, you can click on the “View Entries” button, to view the selected entries in the regular admin page.

Using Charts

Using charts allows you to generate charts with a high degree of flexibility, including things like averages, sums, substring queries, etcetera. Most of the the regular django querying abilities are available in charts, even if the fields aren’t normally visible in the admin.

This is nice in that it is flexible, but it requires a bit of setup and its not a good idea to give create/edit access to non-superusers due to the ability to query data on almost any field.

Registering Statistics

In order to do anything, you’ll need register some models or trackstats metrics. You can find examples of this in tests/models.py, but the short of it looks like this:

from django_adminstats.registry import register_model

class Currency(models.Model):
    slug = models.CharField(max_length=10)
    name = models.CharField(max_length=100)
    current_usd_rate = models.DecimalField()
    sign = models.CharField(max_length=10, default='$')

class Transaction(models.Model):
    amount = models.DecimalField()
    currency = models.Foreignkey(to=Currency)
    date = models.DateTimeField(auto_now_add=True)

register_model(Transaction)

By default, we look for a field called ‘date’ on the model, and it should be either a DateField or DateTimeField. If you want to use a different, field (for example if you’re using Django’s default user and you want to chart by ‘joined_at’) you need to create a registration subclass.

from django_adminstats.registry import registry, register

class UserRegistration(ModelRegistration):
    date_field = 'joined_at'

register(UserRegistration())

Creating Charts

You can add charts in the admin. In order for the chart to show anything, you need to add criteria. By default, it will just show a count of all the items charted by the date field, if you to change this, you need to add things in the filter query, axis query, and group query fields.

First, the content of these fields is formatted like a URL querystring, for example a filter query of message=Hello%20World&x=y is equivalent to .filter(message='Hello World', x='y'). Note that you only use the key0=value0&key1=value1 form in the filter query, the axis and group queries are just key0&key2.

Second, you can use lookups and relations just like in a normal django query (e.g. field__gt=2 or field__related_field).

Finally, you can also specify functions to use on the field by doing field:function. For example id:count is the default axis query when the field is left blank.

Demo

Just run make demo and log in with user admin and password admin.

CHANGES

0.7.4

• Use the view instead of change permission for viewing charts

0.7.3

• Fix error in description generation

0.7.2

• Make chart control list items more visually distinct

0.7.1

• Fix packaging problem (missing files)

0.7.0

• Add new “Explore” charts, upgrading users may want to adjust INSTALLED_APPS.

0.6.3

• Drop pbr for regular setuptools, add VERSION attribute

• Fix some new problems that flake was catching

• Add screenshot of chart

0.6.2

• Fix problem where criteria with same stats would clobber each other

0.6.1

• fix programming bug with by-month trackstats metrics, better test coverage

0.6

• fix autocomplete options for tracstats criteria

• Fix problems with the autocomplete not (re)initializing properly

0.5

• Add autocomplete for query fields in admin, and rename query fields

0.4

• Add documentation for query fields and registrations

• Add copy chart admin action

0.3

• Add the ability to filter/group related models in trackstats by-object stats registrations