man Bric::Templates () - Producing Templates on the Bricolage System

NAME

Bric::Templates - Producing Templates on the Bricolage System

VERSION

$LastChangedRevision$

DATE

$LastChangedDate: 2004-05-24 18:17:44 -0700 (Mon, 24 May 2004) $

BACKGROUND

Before talking about templates, let's take a moment to discuss all the major players in the publish process and get the terminology down. The three elements involved in publishing a story are:

•
Elements
•
Templates
•
Business Data

A quick word about each of these elements.

Elements

These are data definitions. They describe what kind of data a particular story will have. For instance, an 'Editorial' element might define that an Editorial contains An author name, a title, and one or more paragraphs. Elements cannot affect the formatting or look of a story, just the data fields available.

Templates

For a given element, a template will give a format for the display of the data fields available in that element. This is where the HTML will come from.

Business Data

This is the data itself. Given an element to constrain the type of data, an author then enters business data. This data is then run through a template to produce an HTML page.

TEMPLATE OVERVIEW

More About Elements

The first place one starts when making a new story is with an Element. An element defines the structure of a story. It lists the data fields available to those creating a story. So for example, a 'Column' element might define a column as having:

•
An author field
•
A column topic field
•
1 or more paragraph fields

So given this definition someone creating a new element of type 'column' will have these three types of fields available to them.

Elements can also contain other elements as part of their definition. For example, maybe the 'column' element above will be used for a column reviewing music. A music column might want to have a short blurb about the albums they review. Additionally, say somewhere else on the site, there is another type of story, Top Music Picks, that lists blurbs about albums. So we create another another element called 'music_blurb' with these fields:

•
Artist name
•
Album name
•
Label name
•
Summary

Now in our 'column' element we can add the 'music_blurb' element:

•
An author field
•
A column topic field
•
1 or more paragraph fields
•
1 or more Music Blurbs

(The + is just to mark the recursive relationship of these contained elements.)

So, now someone editing this story has the ability to add an author field, a topic for the column, a few paragraphs, and a blurb about the album they are reviewing. Elements can be nested as deeply as needed. So an element can contain an element, that contains another element, etc.

The most important thing to understand about elements is that they simply define the set of data that the author will be asked for when creating/editing a story. The element has no bearing on the actual data itself, or the formatting of that data!

Template Basics

Templates are the things that actually format data. A template is associated 1 to 1 with an element. Said a different way, a template knows how to format the data for one element. So let's say we create a new template and associate it with the 'column' element we defined above. Maybe we want to format a column like this:

    <html>
        <head><title>$column_topic</title></head>
        <body>
            By $author<br />
            <br />

            <p>$para1</p>
            <p>$para2</p>
            ...
            display_element(music_blurb)
            ...
        </body>
    </html>

The above doesn't represent the actual syntax for inserting values, but is just meant to convey a template's relationship to an element. We'll get into the nuts and bolts later.

Note that the 'music_blurb' element is simply passed to a display function. Remember, templates represent the data for only a single element; there is another template that formats a music_blurb element.

Templates And Categories

While a template formats the data of a single element, there may be many templates that all format the same element. Why? Templates also belong to a single category. Imagine that the site for which we've created the current element and template examples is an arts and media site. There might be different sections of the site: one for music, one for theater, and one for gallery art. This site might therefore define the following categories:

•
Music
•
Theater
•
Gallery

Categories (somewhat like elements) can contain other categories. This is just to provide further distinction:

•
Music
•
Music/DJ
•
Music/Classical
•
Music/Rock
•
Theater
•
Theater/Broadway
•
Theater/Community
•
Gallery
•
Gallery/Painting
•
Gallery/Sculpture

In addition to these categories, the Bricolage system always provides a 'root' category by default. This is assumed to always be the top category. If you don't want something to be specific to a particular category, you can assign it to the root category. Root is usually shown as a '/' in the publishing tool:

•
/
•
/Music
•
/Music/DJ
•
/Music/Clasical
•
/Music/Rock
•
...

So, what does this all mean for templates? By allowing many templates, each associated with a different category, to format the same element, we can have a custom look for our data in each section of our site. Say we want to have a column run in all the major categories of our site. That is, we want a Music column, a Theater column, and a Gallery column. We want to collect the same data for each of these columns, but we want the data to look different when published to each of these columns.

Note that a new template for every single category is not necessary. The 'DJ', 'Classical' and 'Rock' subcategories will all inherit from the template associated with the 'Music' category. If there is no need to change how the data is displayed between any category, you can simply create a template associated with the root category. All the categories will inherit from it.

Templates And Output Channels

There is one final parameter that a template has called an 'Output Channel'. For this discussion output channels can be thought of as hidden categories. Basically an output channel is used mostly during distribution. A single story might be published to the web site and a co-brand site at the same time. Both of these would be considered an output channel, with the web site considered the primary output channel.

Again, this just gives the template producer the opportunity to tweak the presentation of the data for display on one output channel (the web site) versus another (the co-brand site).

Bringing It All Together

So you might be asking, where do all my templates go, how does the system know what categories have templates defined, and how are they used during a publish?

Now, while there should be no need to manipulate templates on the filesystem itself (and it could cause the Bricolage system considerable confusion), I'll detail how the files are laid out. This will hopefully make clear a few things about templates and publishing.

Let's start with names. Elements are given names when they are created. In our examples above, we created a 'column'. When you create a template, and choose an element that it should format, it automatically assumes the name of that element. So, if our element is named 'column', our template is named 'column' as well.

The code defined in a template is Mason code. That is, HTML along with some special Perl tags thrown in. So when saving the code of a template to the file-system, we take the template name and append '.mc' or Mason element and use that as the filename. For our 'column' template, the filename would be 'column.mc'.

The category associated with a template is used as part of the path to the template file. So for a template in the '/Music/Rock' category, the path fragment would be 'music/rock/column.mc'.

Finally the output channel is also used as a part of the path. Rather than take the output channel's name, its ID is used. Let's just say that the ID for the primary output channel is 1. So, our column template would live in 'oc_1/music/rock/column.mc'.

If we assume that our element root is CW$COMP_ROOT, the full path to our template is '$COMP_ROOT/oc_1/music/rock/column.mc'.

Publish Examples

So, let's assume that we have three column templates: one associated with the root category, one associated with the music category, and one associated with the sculpture gallery category. We would have these three files on the filesystem:

    $COMP_ROOT/oc_1/column.mc
    $COMP_ROOT/oc_1/music/column.mc
    $COMP_ROOT/oc_1/gallery/sculpture/column.mc

Now, somebody creates a new column. When creating new business data, the author has the opportunity to associate it with one or more categories. The table below shows the 'column.mc' template used for a publish in a particular category. Assume that all publishes are done to the same output channel with an ID of 1:

    Category Story is Published      Formatting element Used
    ---------------------------      -------------------------
    Music/Rock                       $COMP_ROOT/oc_1/music/column.mc
    Theater/Broadway                 $COMP_ROOT/oc_1/column.mc
    Gallery                          $COMP_ROOT/oc_1/column.mc
    Gallery/Art                      $COMP_ROOT/oc_1/column.mc
    Gallery/Sculpture                $COMP_ROOT/oc_1/gallery/sculpture/column.mc

As you can see a story published in a particular category will lookup 'up' the category tree to find a template to format it, if it can't find it in the starting category.

Note that in the case of the story published in 'Music/Rock' and in 'Gallery/Sculpture' there were templates found before the root template at '$COMP_ROOT/oc_1/column.mc'. It's important to realize that templates do not *chain*. That is, the story published to 'Music/Rock' will *not* get formatted by both '$COMP_ROOT/oc_1/music/column.mc' AND '$COMP_ROOT/oc_1/column.mc'. Once a template is found, the search stops.

One Final Trick

As said above, templates do not chain. However, sometimes you might want formatting code to be inherited rather than overridden. Those of you familiar with mason might see the similarity between how templates are handled and the mason concept of 'dhandler's. While not exactly analogous, it's pretty close.

With a small change to how we create templates, we can get them to behave like Mason 'autohandler's. In Mason an 'autohandler' is an element that is called before any other mason element is called. Moreover, Mason autohandlers can chain. For example, assume the following files exist:

    /a/b/c/page.mc
    /a/b/autohandler
    /autohandler

If the file 'page.mc' is called by mason, the autohandlers will be executed before 'page.mc' is called. The order of execution is:

    /autohandler
    /a/b/autohandler
    /a/b/c/page.mc

Given a starting path, mason looks for the first possible 'autohandler' in that path, and then works its way down executing any other autohandlers until it finally reaches the element that was originally called.

In Bricolage, you can choose not to associate a template with an element. That is, you can have a template that is only associated with a category and an output channel. Any template that is not associated with an element will be named 'autohandler' and will behave just like a Mason autohandler.

Autohandler Example

What good are autohandlers? Let's continue with our running example of the arts and media site. Let's say that, irregardless of what kind of story is being used, you want the same basic HTML wrapper. Whether the story is a 'column' or a 'review' or an 'editorial', you want the same header and footer on every page. The easiest way to do this is by using an autohandler.

If we create a new template in the root category for the primary output channel (for this example, the 'web' output channel with ID 1), then we'll get the file:

    $COMP_ROOT/oc_1/autohandler

The code for this template might look like this:

    <!-- Root Autohandler -->
    <html>
        <head><title>Arts n' Media</title></head>
        <body>
            <!-- Header Nav -->
            <table><tr><td><a href="">link1</a></td>
                       <td><a href="">link2</a></td>
                       <td><a href="">link3</a></td></tr></table>

    % $burner->chain_next;

            <hr />
            Comments? <a href="email:foo@bar.com">foo@bar.com</a>
        </body>
    </html>

Most of this should look pretty normal. The only difference is the bit of Perl code in the middle. There is a call to a method named 'chain_next'. This is where all further content in the chain will go. So, given this autohandler template, if we publish a column to the 'Music/Rock' category the following templates will be used in this order:

    $COMP_ROOT/oc_1/autohandler
    $COMP_ROOT/oc_1/music/rock/column.mc

The content that 'column.mc' outputs after being run will wrapped by the HTML in the autohandler template.

As a last example of chaining templates, let's say that we wanted some special HTML to appear just in the music section, irregardless of what type of story we were publishing. We would then create a new template that was not associated with an element, but was associated with the 'Music' category and the primary output channel:

    $COMP_ROOT/oc_1/music/autohandler

The code for this template might look like this:

    <!-- Music Category Autohandler -->
    <h1>Music</h1>
    <table width="570" border="0">
        <tr><td>

    % $burner->chain_next;

        </td></tr>
    </table>

So, given these autohandler templates, if we again publish a column to the 'Music/Rock' category the following templates will be used in this order:

    $COMP_ROOT/oc_1/autohandler
    $COMP_ROOT/oc_1/music/autohandler
    $COMP_ROOT/oc_1/music/rock/column.mc

After publishing this column, the final page output will look like this (just assume that column.mc outputs the story data in <p> tags):

    <!-- Root Autohandler -->
    <html>
        <head><title>Arts n' Media</title></head>
        <body>
            <!-- Header Nav -->
            <table><tr><td><a href="">link1</a></td>
                       <td><a href="">link2</a></td>
                       <td><a href="">link3</a></td></tr></table>

            <!-- Music Category Autohandler -->
            <h1>Music</h1>
            <table width="570" border="0">
                <tr><td>

                    <!-- Column Template -->
                    <p>
                        This is a test column
                    </p>

                </td></tr>
            </table>

            <hr />
            Comments? <a href="email:foo@bar.com">foo@bar.com</a>
        </body>
    </html>

WRITING TEMPLATES

The above sections should give you an idea of how to setup a template what it means to associate a template with an element, a category, and an output channel. Now let's talk a little about actually writing template code.

Writing template code is no different than writing Mason code. If you've written Mason code, you should have no problem writing templates. If you have not written Mason code before, I suggest you get familiar with it by reading HTML::Mason::Devel. A tutorial on Mason is beyond the scope of this document.

Terminology

Given that template code is just Mason code, the only thing you need to know is how to access the story data. Again, let's take a small break to discuss terminology, this time for story objects.

Stories are based on elements. Before you can create a new story you must pick an element upon which your story will be based. After picking an element, you will then be given a list of available fields for which you can enter data. If we use the 'column' element we defined earlier, our fields will be:

•
An author field
•
A column topic field
•
1 or more paragraph fields
•
1 or more Music Blurbs

Now, if you remember, the author field, column topic field and paragraph fields are all just text fields, while the music blurb is another element contained within the column element. This gives us two varieties of fields that can be in a story.

In the interest of keeping things simple for the story authors, these two types of fields are given the same name. Early beta tests showed that authors didn't want to know the difference between text fields and contained elements. It was thought to be too confusing and complicated. Unfortunately that has made your job, the template producer, more confusing and complicated because the term 'element' is about to become wildly overloaded.

In a story, text fields AND contained elements are both referred to as 'elements'. So 'element' now means

1)
A list of allowed data fields that compose a type of story
2)
A particular text data field as defined by 1) in a story
3)
A particular instance of 1) in a story

To make sure these concepts stay clear in the following discussions, let's use this convention. When I speak of an element as defined by 1), I will refer to it as an Element with a capital 'E'. After all, it's the true form of a story; the definition of a particular type of story. When I speak of elements as defined by 2) and 3) I will use a lowercase 'element'.

Furthermore, since a story has two types of elements, let's establish a way of distinguishing between them. We'll call all text data fields of a story 'data elements' and the contained fields that are based on an Element, 'container elements'. When we don't need to distinguish them, we'll collectively refer to them as 'elements'.

So our new story based on the column Element has the following elements available to it:

•
An author data element
•
A topic data element
•
A paragraph data element
•
A music blurb container element

Now that we know what a story can have and how to refer to them, let's look at the methods to access them.

Available Objects and Methods

First, there are 3 new variables that are exported into the Mason Perl-space for use with templates. These are:

$story
The story object
$element
The current element object
$burner
The burn system object

Story Object Methods

The story object is the object containing the story to be published. It is used to access all the metadata of a story. It can be used to indirectly access the actual data of the story, but there is a better way to do this. We'll discuss how and why shortly. A full list of methods available via the story object can be found by reading the POD in Bric::Biz::Asset::Business::Story. However, I'll highlight the most important/useful ones here:

$story->get_title
Retrieve the title of this story
$story->get_description
Retrieve the description for this story.
$story->get_cover_date
Returns the date this story is tied to (not necessarily when it was published). A cover date might be 'June 1, 2001' for the June issue of, say, Seventeen magazine, but it might actually be published some time before or after that exact date.
$story->get_primary_uri
This is the primary URI for this story. This returns the URI for the primary output channel in the primary category.

Element Object Methods

The current element object contains the real data of a story. Since a story is based upon an Element, the CW$element object is an instance of that Element. So if the Element defined an 'author', 'topic' and 'paragraph' fields, you know you can access those data elements via CW$element. Where CW$element really becomes interesting is for contained elements.

In our example, we had a 'music_blurb' contained element. There is no 'music_blurb' story; it's simply part of the larger 'column' story. However, 'music_blurb', just like any other Element, needs to have a template that knows how to format it. So while burning our 'column' story, we will at some point call out to the 'music_blurb' template to format our 'music_blurb'. When the 'music_blurb' template code is run, CW$story will still contain our 'column' story object, but CW$element will contain an instance of our 'music_blurb' Element, and will allow access to the data fields defined in that Element.

Sound confusing? Let just go over the methods available to CW$element, and we'll get to an example soon enough. This returns the data associated with the data element given by CW$name. For our example in our story, CW$name is one of 'author', 'topic' or 'paragraph'. The CW$num argument is optional. It is meant for data elements that can occur more than once, such as 'paragraph's. If CW$num is not given the text for the first CW$name data element will be returned. Note to programmers; CW$num starts at '1' not '0'. This returns a container element object. This object is of the same type as CW$element. Usually you will just pass this object to a method that will call the correct template for you. The CW$num argument is optional. Like the CW$num argument in 'get_data', it is used to retrieve container elements that occur more than once in a given story. If CW$num is not given, the first container element object will be returned. Note to programmers; CW$num starts at '1' not '0'.

$element->get_elements
Returns an array of both container element objects and data element objects. Container element objects, as noted above, are of the same type as CW$element. Data element objects are an object representation of the story data for a data element. This bullet list identifies the methods you can call on a container element object. A short list of methods available for data element objects will be given after this list.
$element->get_place
Story editors have the opportunity to arrange the elements of their story in a particular order. This method returns a number giving this container element's place among all other elements (both data and container). If CW$element is the element object for the main story (i.e. the 'column' element of a 'column' story) this will return 1.
$element->get_object_order
For container elements that can occur more than once in a story, this will return this element's place among like container elements. For example a 'column' story might contain several 'music_blurb' container elements. If this happens to be the 2nd of 3 'music_blurb' container elements, this method will return 2. If CW$element is the element object for the main story (i.e. the 'column' element of a 'column' story) this will return 1.
$element->get_name
Return the name of this container element.

The 'get_elements' method returns some container element objects and some data element objects. Here is a list of methods you can call on a data element object:

$delem->get_data
Returns the data associated with this data element.
$delem->get_place
Same concept as the container element 'get_place' method.
$delem->get_object_order
Same concept as the container element 'get_object_order' method.
$delem->get_name
Returns the name of this data element.

Finally, to distinguish between data and container element objects you can call this method on any type of element:

$element->is_container
Returns 1 if this is a container element object and 0 if this is a data element object.

Burn System Object Methods

Here are the methods available via the burn system object.

$burner->chain_next
From an autohandler template, this calls the next autohandler, or the specific story template.
$burner->display_element($element)
This takes either a data element object or a container element object. Given a container element object, this method will locate the proper template and output the results of burning that template. Given a data element object, this method will simply output the data of that data element. An sprintf version of CW$burner->display_element($element), that is it returns the HTML as a string instead of directly outputting it to the browser.

A Template Example

For this template example, let's use the 'column' Element included in the Bricolage system as our story. This 'column' is a little different than the one we've been discussing. Here is the layout for this Element:

Element Column
*
Data element 'deck'
*
Container element 'page'
Element Page
*
Data element 'paragraph'
*
Data element 'pull quote'
*
Data element 'next'
*
Data element 'previous'
*
Container element 'inset'
Element Inset
*
Data element 'copy'

There are few things to note here. Notice that this story structure is 3 levels deep rather than our two levels before. Additionally, the 'page' Element was created with a flag marking it as a paginated Element. This means that Mason will automatically start a new output file for every 'page' Element burned, provided it is burned with the display_pages() method.

Here is the code for the 'column' template:

    <!-- The column element -->
    <html>
        <head>
            <title><% $story->get_title %></title>
        </head>
        <body>

    % $burner->display_pages('page');

        </body>
    </html>

This code simply iterates through each 'page' container element in the 'column' story and displays it. Remember that the page element was created with the 'paginated' flag and will start a new output file so we'll save the real HTML for the 'page' element. Note that we skip the 'deck' data element. This element is meant for when this is used in a related story context. Using related stories will be discussed later in another document.

Next let's look at some example code for the 'page' element:

    <!-- The page element -->
    % my $page = $burner->get_page;

    <!-- Only display this title if we are on the first page -->
    % unless ($page) {
            <h1><% $story->get_title %></h1>
    % }

    <!-- Show the content for this element -->
    <%perl>
        foreach my $e ($element->get_elements) {
            if ($e->has_name('paragraph')) {
                $m->out('<p>');
                $burner->display_element($e);
                $m->out('</p>');
            } elsif ($e->has_name('pull_quote')) {
                $m->out('<p><i>');
                $burner->display_element($e);
                $m->out('</i></p>');
            } elsif ($e->has_name('inset')) {
                $burner->display_element($e);
            }
        }

        my $next_txt = $element->get_data('next');
        my $prev_txt = $element->get_data('previous');
        my $next = $burner->next_page_file;
        my $prev = $burner->prev_page_file;
    </%perl>

    <!-- Show previous page link if it exists -->
    % if ($prev and $prev_txt) {
            [Page <% $page %>]&lt;&lt;&lt;
            <a href="<% $prev %>"><% $prev_txt %></a>
    % }

            &nbsp;&nbsp;&nbsp;

    <!-- Show next page link if it exists -->
    % if ($next and $next_txt) {
            <a href="<% $next %>"><% $next_txt %></a>
            &gt;&gt;&gt;[Page <% $page + 2 %>]
    % }

This code outputs the main HTML, outputting a header if this is the first page of the column, followed by the data for this story. After the data is output the 'previous' and 'next' links are displayed. Note that in the foreach loop iterating over all elements the method 'display_element' is used to display both data and container elements. For the data elements we could have used:

    $m->out($e->get_data);

to return the data. Also, it is important to realize that there are no restrictions against having a container and a data element with the same name. While this is unlikely to happen, if it were true for the 'column' element the above code would run fine, but may not output what is expected.

Finally here is the template code for the 'inset' container element:

    <!-- The inset element -->
    <table width="570" border="2">
        <tr><td>
    % $element->get_data('copy');
        </td></tr>
    </table>

That's about it. The above code for the three Elements, once used to create new templates and deployed, should allow you to publish 'column' stories. Make sure you create your templates in the root category to ensure you can publish a story to any category.

APPENDIX

Class Names

    Object Type         Associated Perl Class
    -----------         ---------------------
    Element             Bric::Biz::AssetType
    Template            Bric::Biz::Asset::Formatting
    Story               Bric::Biz::Asset::Business::Story
    element             Bric::Biz::Asset::Business::Parts::Tile
    Data element        Bric::Biz::Asset::Business::Parts::Tile::Data
    Container element   Bric::Biz::Asset::Business::Parts::Tile::Container
    Burn System         Bric::Util::Burner

Glossary

autohandler
A mason concept. As applied to Bricolage, it is a variation on a template that can form a line of 'chained' templates.
Burn System
The element of the Bricolage system that joins story data with existing templates.
Category
A way of organizing a web site into separate sections.
Container element
An element containing structured data based on an Element.
Data element
An element that represents textual/simple data.
dhandler
A Mason concept. Templates behave in a similar manner to dhandlers.
Element
The form of a story or story part.
element
An instance of an Element in a story.
Mason
An apache/perl templating system.
Output Channel
A destination for published data.
Story
The data to be published.
Template
An output format for published data.

AUTHOR

Garth Webb <garth@perijove.com>

SEE ALSO

Bric, Bric::AdvTemplates, Bric::Biz::AssetType, Bric::Biz::Asset::Formatting, Bric::Biz::Asset::Business::Story, Bric::Biz::Asset::Business::Parts::Tile, Bric::Biz::Asset::Business::Parts::Tile::Data, Bric::Biz::Asset::Business::Parts::Tile::Container, Bric::Util::Burner