Author image Michael Roberts
and 1 contributors

NAME

Data::Org::Template::Cookbook - Simple recipes for my flavor of templates

DESCRIPTION

I have no memory whatsoever for syntactic detail, which is a significant handicap for a programmer. As a result, I rely on cookbooks, notes, and code snippets to do the most trivial of things - I can only whip out code quickly when it's entirely non-trivial, like tossing closures around.

This set of examples is therefore really more for my own use than yours, but I hope you find it useful as well.

BASIC TEMPLATING

Hello, world

The most fundamental programming task, as we all know, is to greet the entire human race. Here's how you'd do that using a DOTemplate - and of course by changing the input data you can greet anybody you like.

    use Data::Org::Template;

    my $t = Data::Org::Template->new ("Hello, [[name]]!");
    print $t->text({name => 'world'});

So far, this is no different from any other template engine - we have a field delimited by some kind of special brackets, and the field names a data item in a hashref. This is just how you set it all up.

Multiline templates

What I normally do for multiline templates (which is the majority of templates) is this:

    use Data::Org::Template;

    my $t = Data::Org::Template->new (<<'EOF');
    Hello to:
    [[name]]
    EOF
    
    print $t->text({name => 'world'});

I don't like those [[brackets]]

You can also easily use a different type of bracket.

    use Data::Org::Template;

    my $t = Data::Org::Template->new ("Hello, {name}!", '{}');
    print $t->text({name => 'world'});

DATA GETTING

So far we've just passed a single hashref to the template when we express it, but we can get fancier than that.

Register a data getter

You can register a data getter ahead of time so text retrieval will always produce an updated value.

    use Data::Org::Template;

    my $data = {name => 'world'};

    my $t = Data::Org::Template->new ("Hello, {name}!", '{}');
    $t->data_getter ($data);
    
    print $t->text; # --> Hello, world!
    
    $data->{name} = 'Bob';
    print $t->text; # --> Hello, Bob!

Getting data from multiple sources

All the getters we've used so far have had a single source - one hashref holding values. But in reality, a getter can accept a list of sources. The getter will consult each source in turn until it gets a defined value.

    use Data::Org::Template;

    my $t = Data::Org::Template->new ("First [[x]] then [[y]]");
    my $data = {x => 'xval'};
    $t->data_getter($data, {y=>'yval'});
    
    print $t->text; # --> First xval then yval
    
    $data->{yval} = 'not yval';
    print $t->text; # --> First xval then not yval

Notice that the value for y in the first hash now shadows the value in the second hash.

Special cases in multiple-source retrieval

There are two special cases that can come in handy. The first is that if there is a "source" in the list that is a scalar value, then it can be retrieved with the special field name of "." - a period.

    use Data::Org::Template;

    my $t = Data::Org::Template->new ("Value: [[.]]");
    
    print $t->text ('value'); # --> Value: value
    

The second comes into play when a data source list is passed to a template that already has a registered data getter. The getter is only consulted for values if the special source "*" has been included in the list passed to expression. That explanation isn't all that clear, but the code probably makes it obvious:

    use Data::Org::Template;

    my $t = Data::Org::Template->new ("First [[x]] then [[y]]");
    my $data = {x => 'xval'};
    $t->data_getter($data);
    
    print $t->text ({y => 'yval'});      # --> First  then yval
    print $t->text ({y => 'yval'}, '*'); # --> First xval then yval
    print $t->text ({y => 'yval', x => 'not xval'}, '*'); # --> First not xval then yval
    print $t->text ('*', {y => 'yval', x => 'not xval'}); # --> First xval then yval
    

The sources are still consulted in the order they're provided, so in the final print statement the template's own getter returns a value for x before the value passed in for expression.

Magic values

Any value returned by a data getter can also be a coderef (a closure) - if so, it is executed to get the final value. (If the return value is still a coderef, it will be executed again. And again. Try not to shoot yourself in the foot with a self-returning coderef, obviously.) This is a kind of neat way to embed a template in another template - just make the embedded template's expression a magic value in the data for the second template.

    use Data::Org::Template;

    my $data = {name => 'world'};
    $t = Data::Org::Template->new("Hello, [[name]]!");
    $t->data_getter ($data);

    my $t2 = Data::Org::Template->new ("Current greeting: '[[greeting]]'");
    $t2->data_getter ({greeting => sub { $t->text }});

    print $t2->text();  # --> Current greeting: 'Hello, world!'
    $data->{name} = 'Bob';
    print $t2->text();  # --> Current greeting: 'Hello, Bob!'

Personally, I think that's pretty darn cool.

Indentation preservation

One feature I actually wrote this library to provide is preservation of indentation when multi-lined values are inserted into a template.

    use Data::Org::Template;

    my $text = <<'EOF';
    This is a text
    that consists
    of multiple lines.
    EOF
    
    my $t = Data::Org::Template->new (<<'EOF');
    This is the value: [[text]]
    And this is the line after it.
    EOF
    
    print $t->text({text => $text});
    # This is the value: This is a text
    #                    that consists
    #                    of multiple lines.
    # And this is the line after it.
    

My original impetus for this idea was years ago, when I was dabbling with literate programming and realized that it didn't work for Python because it was such a pain to insert values that needed their indentation preserved properly.

FORMATTING VALUES

HTML-encoding values

There's actually an optional step in between retrieving a value from the getter and incorporating it into the template expression, and that's formatting. A really common formatter is HTML encoding, which substitutes HTML entities for the special values &, <, and >. This is so common it's the only formatter I've already implemented standard in the module, and this is how you use it:

    use Data::Org::Template;

    $t = Data::Org::Template->new("<i>[[string|html]]</i>");
    print $t->text({string => '<x>'}); # --> <i>&lt;x&gt;</i>
    

Anything after a vertical bar is a formatter, and you can run the value through an arbitrary number of formatters. Spaces around the bar are ignored. The formatter specification can also include spaces and some punctuation; the initial alphanumeric string identifies the formatter for lookup, but the whole string is passed to the formatter factory to parameterize the formatter instance. (This part isn't well-tested, so it might break if you actually use it; I just like leaving doors open for later use if the need arises.)

Registering your own formatter

Since the formatter represents arbitrary code that can modify the value as it comes from the data source, there are lots of things you can do with it, but since I haven't anticipated all possible (or even obvious) needs, you can also register your own formatters. The formatter factory is created when the data getter is created, so to do a custom formatter, you'll want something like this:

    use Data::Org::Template;

    my $t = Data::Org::Template->new("Hello, [[name | myfmt]]!");
    my $getter = $t->data_getter ({});
    $getter->formatter->register ('myfmt', sub { sub { return '...' . $_[0] . '...' } } );

    print $t->text ({name => "Bob"}); # --> Hello, ...Bob...!
    

FANCY STUFF

Up to this point, all we've done is get values and put them into vanilla templates. We can also use section directives to direct entire sections of the template; three are defined (but of course you can register your own - if you want to read the code; eventually, if the need arises, I'll work out how it works, fix any bugs I find, and document it here, but today is not that day). The three available are [[.if]], [[.with]], and [[.list]]. All section directives are terminated with [[..]] and they can have subsection directives of the form [[..else]].

Conditional templates with .if

The .if directive is probably the most obvious to use. It checks a value, and includes its content if the value is true. If it contains an [[..else]] subsection, that will be displayed if the value is false. Easy, right?

    use Data::Org::Template;

    my $data = {flag => 1};
    my $t = Data::Org::Template->new("Current value: [[.if flag]]yes[[..else]]no[[..]]");

    print $t->text; # --> Current value: yes
    $data->{flag} = 0;
    print $t->text; # --> Current value: no
    

If directives appear on their own line in a multi-line template, then the line end is considered part of the directive when replacing.

    use Data::Org::Template;

    my $data = {flag => 1};
    my $t = Data::Org::Template->new(<<'EOF');
    Current value:
      [[.if flag]]
      yes
      [[..else]]
      no
      [[..]]
    EOF

    print $t->text ({flag => 0});
    # Current value:
    #   no

(This is because I spent too much time agonizing over early template engines that would leave all kinds of blank lines everywhere in the output.)

Embedding hashrefs with .with

If one of your values is a hashref, you can define a subtemplate that accesses its values (and so on recursively) using the .with directive. If there's a chance that there is no such hashref, an [[..else]] subsection can be included.

    use Data::Org::Template;

    my $data = {x => {name => 'world'}};
    my $t = Data::Org::Template->new(<<'EOF');
    [[.with x]]
    Hello, [[name]]!
    [[..]]
    EOF

    print $t->text ({flag => 0}); # --> Hello, world!

List templates with .list

And if one of your values is an arrayref, you can repeat a section of the template for each item in the list. There are a lot of moving parts in a list template; I tried to cover all the use cases I've run into myself. The main section of the list is expressed for each of its items and assumes that the item is a hashref. (If the item is just a scalar, of course, it can be accessed with the special name ..)

If there is an [[..alt]] subsection, it is expressed in the calling data context between rows.

If there is an [[..else]] subsection, it is expressed if the arrayref contains no items.

If the list is in an arrayref, not an iterator (see below for iterators), the default value getter also provides special values .count and .total for the current item's number and the total number of items in the list. This doesn't apply to iterators, because we don't know the total number of rows an iterator will return and you can build your own counting variable for your input iterator if you need one.

    use Data::Org::Template;

    $t = Data::Org::Template->new (<<EOF);
    People:
    [[.list people]]
    - [[name]]
    [[..else]]
    (none listed)
    [[..]]
    EOF

    print $t->text({'people' => [{name => 'Bob'}, {name => 'Sam'}]});
    # People:
    # - Bob
    # - Sam

    print $t->text();
    # People:
    # (none listed)

COOKING WITH ITERATORS

This template engine was specifically written to work with my Iterator::Records module. An itrecs is a factory for Iterator::Simple iterators that return a series of arrayrefs, each arrayref being a field. The names of the fields are also defined in advance. In the upcoming version, a type descriptor will also be definable for each column, but that won't really affect template expression (I suppose we could have a formatter, something like "|by_type" or something).

The really convenient thing about itrecs is that they are a swappable way to represent query results - from SQLite, from the filesystem, from walks of arbitrary data trees, whatever - and this template engine gives us a simple way of turning that into human-readable text.

Turning an SQLite query into an HTML table

This is a ubiquitous use case.

    use Iterator::Records;
    use Data::Org::Template;
   
    my $db = Iterator::Records::db->open ("mydb.sqlt");
    # Assume a table like this:
    # create table people (
    #   first text,
    #   last  text
    # );
    
    $t = Data::Org::Template->new (<<EOF);
    <table>
    <tr><th>First name</th><th>Last name</th></tr>
    [[.list .]]
    <tr><td>[[first|html]]</td><td>[[last|html]]</td></tr>
    [[..else]]
    <tr><td colspan="2">No records found.</td></tr>
    [[..]]
    </table>
    EOF

    $t->text($db->select ("select first, last from people")->iter);
    

Simple. Easy to remember. You could write a generic table displayer that would read the iterator's field list and use the actual field names as column headers; eventually I'm going to write that as part of a general SQL-to-HTML toolset.

YOUR TURN

That's all the recipes I'm going to write at the moment. I've already spent more time on writing this cookbook than I really intended. A couple of the examples aren't tested all that well, so if I use them in something and it turns out they were wrong, I'll update. If you use this module, and have some convenient use cases you'd like to see here, by all means get in touch.