doc/book/en/A02a-create-cube.en.txt
author Graziella Toutoungis <graziella.toutoungis@logilab.fr>
Tue, 31 Mar 2009 17:09:38 +0200
changeset 1194 e224f064a268
parent 296 65564e83853e
child 1249 905d76e38433
permissions -rw-r--r--
optimize this code part

.. -*- coding: utf-8 -*-

Create your cube
----------------

Once your `CubicWeb` development environment is set up, you can create a new
cube::

  cubicweb-ctl newcube blog

This will create in the cubes directory (``/path/to/forest/cubes`` for Mercurial
installation, ``/usr/share/cubicweb/cubes`` for debian packages installation) 
a directory named ``blog`` reflecting the structure described in :ref:`cubesConcepts`.

.. _DefineDataModel:

Define your data model
----------------------

The data model or schema is the core of your `CubicWeb` application.
It defines the type of content your application will handle.

The data model of your cube ``blog`` is defined in the file ``schema.py``:

::

  from cubicweb.schema import format_constraint
  
  class Blog(EntityType):
    title = String(maxsize=50, required=True)
    description = String()

  class BlogEntry(EntityType):
    title = String(required=True, fulltextindexed=True, maxsize=256)
    publish_date = Date(default='TODAY')
    content = String(required=True, fulltextindexed=True)
    entry_of = SubjectRelation('Blog', cardinality='?*') 


A Blog has a title and a description. The title is a string that is
required by the class EntityType and must be less than 50 characters. 
The description is a string that is not constrained.

A BlogEntry has a title, a publish_date and a content. The title is a
string that is required and must be less than 100 characters. The
publish_date is a Date with a default value of TODAY, meaning that
when a BlogEntry is created, its publish_date will be the current day
unless it is modified. The content is a string that will be indexed in
the full-text index and has no constraint.

A BlogEntry also has a relationship ``entry_of`` that links it to a
Blog. The cardinality ``?*`` means that a BlogEntry can be part of
zero or one Blog (``?`` means `zero or one`) and that a Blog can
have any number of BlogEntry (``*`` means `any number including
zero`). For completeness, remember that ``+`` means `one or more`.


Create your instance
--------------------

To use this cube as an application and create a new instance named ``blogdemo``, do::
  
  cubicweb-ctl create blog blogdemo


This command will create a directory ``~/etc/cubicweb.d/blogdemo``
which will contain all the configuration files required to start
you web application.

Welcome to your web application
-------------------------------

Start your application in debug mode with the following command: ::

  cubicweb-ctl start -D blogdemo


You can now access your web application to create blogs and post messages
by visiting the URL http://localhost:8080/.

A login form will appear. By default, the application will not allow anonymous
users to enter the application. To login, you need then use the admin account
you created at the time you initialized the database with ``cubicweb-ctl
create``.

.. image:: images/login-form.png


Once authenticated, you can start playing with your application 
and create entities.

.. image:: images/blog-demo-first-page.png

Please notice that so far, the `CubicWeb` franework managed all aspects of 
the web application based on the schema provided at first.


Add entities
------------

We will now add entities in our web application.

Add a Blog
~~~~~~~~~~

Let us create a few of these entities. Click on the `[+]` at the left of the
link Blog on the home page. Call this new Blog ``Tech-blog`` and type in
``everything about technology`` as the description, then validate the form by
clicking on ``Validate``.

.. image:: images/cbw-create-blog.en.png
   :alt: from to create blog

Click on the logo at top left to get back to the home page, then
follow the Blog link that will list for you all the existing Blog.
You should be seeing a list with a single item ``Tech-blog`` you
just created.

.. image:: images/cbw-list-one-blog.en.png
   :alt: displaying a list of a single blog

Clicking on this item will get you to its detailed description except
that in this case, there is not much to display besides the name and
the phrase ``everything about technology``.

Now get back to the home page by clicking on the top-left logo, then
create a new Blog called ``MyLife`` and get back to the home page
again to follow the Blog link for the second time. The list now
has two items.

.. image:: images/cbw-list-two-blog.en.png
   :alt: displaying a list of two blogs

Add a BlogEntry
~~~~~~~~~~~~~~~

Get back to the home page and click on [+] at the left of the link
BlogEntry. Call this new entry ``Hello World`` and type in some text
before clicking on ``Validate``. You added a new blog entry without
saying to what blog it belongs. There is a box on the left entitled
``actions``, click on the menu item ``modify``. You are back to the form
to edit the blog entry you just created, except that the form now has
another section with a combobox titled ``add relation``. Chose
``entry_of`` in this menu and a second combobox appears where you pick
``MyLife``. 

You could also have, at the time you started to fill the form for a
new entity BlogEntry, hit ``Apply`` instead of ``Validate`` and the 
combobox titled ``add relation`` would have showed up.


.. image:: images/cbw-add-relation-entryof.en.png
   :alt: editing a blog entry to add a relation to a blog

Validate the changes by clicking ``Validate``. The entity BlogEntry
that is displayed now includes a link to the entity Blog named
``MyLife``.

.. image:: images/cbw-detail-one-blogentry.en.png
   :alt: displaying the detailed view of a blogentry

Note that all of this was handled by the framework and that the only input
that was provided so far is the schema. To get a graphical view of the schema,
point your browser to the URL http://localhost:8080/schema

.. image:: images/cbw-schema.en.png
   :alt: graphical view of the schema (aka data-model)


.. _DefineViews:

Define your entities views
--------------------------

Each entity defined in a model inherits defaults views allowing
different rendering of the data. You can redefine each of them
according to your needs and preferences. If you feel like it then
you have to know how a view is defined.


The views selection principle
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

A view is defined by a Python class which includes: 
  
  - an identifier (all objects in `CubicWeb` are entered in a registry
    and this identifier will be used as a key)
  
  - a filter to select the resulsets it can be applied to

A view has a set of methods complying
with the `View` class interface (`cubicweb.common.view`).

`CubicWeb` provides a lot of standard views for the type
`EntityView`, for a complete list, you
will have to read the code in directory ``cubicweb/web/views/``

A view is applied on a `result set` which contains a set of
entities we are trying to display. `CubicWeb` uses a selector
mechanism which computes a score used to identify which view
is the best to apply for the `result set` we are trying to 
display. The standard library of selectors is in 
``cubicweb.common.selector`` and a library of methods used to
compute scores is available in ``cubicweb.vregistry.vreq``.

It is possible to define multiple views for the same identifier
and to associate selectors and filters to allow the application
to find the best way to render the data. We will see more details
on this in :ref:`DefinitionVues`.

For example, the view named ``primary`` is the one used to display
a single entity. We will now show you hos to customize this view.


View customization
~~~~~~~~~~~~~~~~~~

If you wish to modify the way a `BlogEntry` is rendered, you will have to 
overwrite the `primary` view defined in the module ``views`` of the cube
``cubes/blog/views.py``.

We can for example add in front of the pulication date a prefix specifying
the date we see is the publication date.

To do so, please apply the following changes:

:: 

  from cubicweb.web.views import baseviews


  class BlogEntryPrimaryView(baseviews.PrimaryView):

    accepts = ('BlogEntry',)

    def render_entity_title(self, entity):
        self.w(u'<h1>%s</h1>' % html_escape(entity.dc_title()))

    def content_format(self, entity):
        return entity.view('reledit', rtype='content_format')

    def cell_call(self, row, col):
        entity = self.entity(row, col)

        # display entity attributes with prefixes
        self.w(u'<h1>%s</h1>' % entity.title)
        self.w(u'<p>published on %s</p>' % entity.publish_date.strftime('%Y-%m-%d'))
        self.w(u'<p>%s</p>' % entity.content)
        
        # display relations
        siderelations = []
        if self.main_related_section:
            self.render_entity_relations(entity, siderelations)

.. note::
  When a view is modified, it is not required to restart the application
  server. Save the Python file and reload the page in your web browser
  to view the changes.

You can now see that the publication date has a prefix.

.. image:: images/cbw-update-primary-view.en.png
   :alt: modified primary view


The above source code defines a new primary view for
``BlogEntry``. 

Since views are applied to resultsets and resulsets can be tables of
data, it is needed to recover the entity from its (row,col)
coordinates. We will get to this in more detail later.

The view has a ``self.w()`` method that is used to output data. In our
example we use it to output HTML tags and values of the entity's attributes.