nti.contentratings 1.2.0

The Default Categories

This package’s default configuration provides two rating categories. One for user ratings, and one for editorial ratings. They are registered for the marker interfaces IUserRatable and IEditorRatable respectively. The TALES expressions used to determine when they apply are designed to work with objects contained in a Zope 2 CMF application (primarily for backwards compatibility with older versions of contentratings which used direct permission checks). Unless they want to allow all users to set and read the ratings, other applications will need to define categories with custom conditions.

Let’s demonstrate how these categories might be used. We need to create some content and mark it with our marker interface:

>>> from contentratings.tests import SampleContainer
>>> content = SampleContainer()
>>> from contentratings.interfaces import IUserRatable
>>> from zope.interface import alsoProvides
>>> alsoProvides(content, IUserRatable)

Now we can adapt to the rating category using the IUserRating interface:

>>> from contentratings.interfaces import IUserRating
>>> adapted = IUserRating(content)
>>> adapted.title
u'User Rating'
>>> float(adapted.averageRating)
0.0
>>> rating = adapted.rate(7.0)
>>> float(adapted.averageRating)
7.0
>>> adapted.numberOfRatings
1
>>> rating = adapted.rate(8.0, 'me')
>>> float(adapted.averageRating)
7.5
>>> adapted.numberOfRatings
2

For more details on the IUserRating API see tests/userstorage.txt.

Editorial ratings are applied similarly, but have a much simpler implementation:

>>> from contentratings.interfaces import IEditorRatable, IEditorialRating
>>> alsoProvides(content, IEditorRatable)
>>> adapted = IEditorialRating(content)
>>> adapted.title
u'Editor Rating'
>>> adapted.rating is None
True
>>> adapted.rating = 6.0
>>> float(adapted.rating)
6.0
>>> adapted.rating = 8.0
>>> float(adapted.rating)
8.0

See tests/editorialstorage.txt for details on how the IEditorialRating API works.

Let’s remove these markers now so that we can examine custom categories:

>>> from zope.interface import noLongerProvides
>>> noLongerProvides(content, IUserRatable)
>>> noLongerProvides(content, IEditorRatable)
>>> IUserRatable(content) # doctest: +ELLIPSIS
Traceback (most recent call last):
...
TypeError: ('Could not adapt', ...)
>>> IEditorialRating(content) # doctest: +ELLIPSIS
Traceback (most recent call last):
...
TypeError: ('Could not adapt', ...)

Custom Rating Categories

There are two ways to create a new rating category, declaratively using ZCML, or programatically using the category factory directly from python. Let’s look at the ZCML way first. To make it work we need to enable the zcml directive:

>>> from zope.configuration import xmlconfig
>>> import contentratings
>>> context = xmlconfig.file('meta.zcml', contentratings)

Loading the package configuration will do the above automatically.

Now we register our rating category using the contentratings:category directive:

>>> context = xmlconfig.string("""
... <configure
...    xmlns:contentratings="http://namespaces.plone.org/contentratings">
...  <contentratings:category
...      for="zope.container.sample.SampleContainer"
...      title="My Rating Category"
...      />
... </configure>""", context=context)

Here we have made use of all of the category default values. As a result we have registered a category which uses the default ZODB storage and IUserRating API, with no restrictions on who can get and set ratings. We can verify this easily since the categories are simply adapters providing the rating interface provided by the (default) storage:

>>> from contentratings.interfaces import IUserRating
>>> content = SampleContainer()
>>> adapted = IUserRating(content)
>>> IUserRating.providedBy(adapted)
True
>>> adapted.context is content
True
>>> adapted.title
u'My Rating Category'

Note that because we provided no name in the configuration, the adapter was registered as the default (unnamed) adapter. The name of the category is the name under which the adapter is registered and it is stored in the category’s name attribute:

>>> adapted.name
''

To provide multiple categories, just register them with unique names:

>>> context = xmlconfig.string("""
... <configure
...    xmlns:contentratings="http://namespaces.plone.org/contentratings">
...  <contentratings:category
...      for="zope.container.sample.SampleContainer"
...      title="My Other Rating Category"
...      name="other"
...      />
... </configure>""", context=context)
>>> from zope.component import getAdapter
>>> adapted = getAdapter(content, IUserRating, name=u'other')
>>> adapted.title
u'My Other Rating Category'
>>> adapted.name
u'other'

If we wanted to accomplish the same thing programatically, we could instantiate the factory directly and register it as an adpater:

>>> from contentratings.category import RatingsCategoryFactory
>>> category = RatingsCategoryFactory(title=u'Another Title', name=u'another')
>>> from zope.component import provideAdapter
>>> provideAdapter(category, adapts=(SampleContainer,), provides=IUserRating,
...                name=u'another')
>>> adapted = getAdapter(content, IUserRating, name=u'another')
>>> adapted.title
u'Another Title'
>>> adapted.name
u'another'

This involves some redundancy, since the interface provided by the storage has to be explicitly declared, and the category name has to be provided twice. Otherwise they are equivalent.

Note that categories are adapters, and adapters may only be registered under the same name for different interfaces/classes. As usual, for a given name, the adapter registered for the most specific interface will be chosen.

The Rating Views

Each category has an associated view_name which is simply the name of a view registered for the rating interface (e.g. IUserRating) to be used when rendering the category in the UI. These are looked up on the Rating Manager and have access to the IRatingManager API, as well as the protected rating storage API, as provided by the manager (e.g. IUserRating).

Resuable base classes for rating views are provided in browser/basic.py (BasicUserRatingView and BasicEditorialRatingView). These views use a named vocabulary to validate input and use the IRatingManager API to determine who can and cannot rate content. The package configuration provides a few rating views by default:

ratings_view (default):  A rating using 1-5 large (25px) stars
small_stars: A rating using 1-5 small (10px) stars
three_small_stars: A rating using 1-3 small (10px) stars

These are each highly customizable using CSS. They are all registered for IUserRating. Additionally, there is a rating_view is registered for IEditorialRating.

The views are responsible for looking up a rating vocabulary and validating user input, as well as rendering the user interface. The security is enforced by the Rating Manager used by the view, however the view may go directly to the storage from rating manager if it wants to override the expression checking (e.g. showing a user their own ratings, though they cannot see others). Creating new views (e.g. non-starred ratings) is quite simple.

A utility is provided for efficiently determining a reusable session key in a generic manner. This can be used to prevent repeat voting from anonymous users. Applications which implement their own anonymous session tracking mechanisms may override this utility locally if desired.

The Aggregator Views

There are also views which find all the rating categories available on the content object being viewed, rendering them in order. These are intended to be used within a viewlet, portlet, macro, or similar. The aggregator view for user ratings is called user-ratings and the one for editorial ratings is called editorial-ratings.

Author:

• Alec Mitchell <apm13@columbia.edu>

Contributors:

• Maurizio Delmonte

(feel free to add your name above if you have made significant contributions)

Thanks To:

• Geoff Davis author of ATRatings from which icons and ideas were stolen.

• Philipp von Weitershausen author of Web Component Development with Zope 3 which provides a nice example of an annotation based rating product, which was the starting point for this implementation.

• Kai Diefenbach author of iqpp.plone.rating from which other icons and UI ideas were stolen.

• Some icons are from Mark James’ Silk icon set 1.3

• The star rating is based on CSS Star Rating Redux via iqpp.plone.rating