=encoding utf8

=head1 NAME

Log::Report::Extract::Template - collect translatable strings from template files

=head1 INHERITANCE

 Log::Report::Extract::Template
   is a Log::Report::Extract

=head1 SYNOPSIS

 # First use of this module: extract msgids from various kinds
 # of text-files, usually web templates.
 # See script "xgettext-perl" for standard wrapper script

 my $extr = Log::Report::Extract::Template->new
   ( lexicon => '/usr/share/locale'
   , domain  => 'my-web-site'
   , pattern => 'TT2-loc'
   );
 $extr->process('website/page.html');  # many times
 $extr->showStats;
 $extr->write;

 # Second use: connect to Template::Toolkit
 # See DETAILS chapter below

 [% loc("Greetings {name},", name => client.name) %]
 [% | loc(name => client.name) %]Greetings {name}[% END %]
 [% 'Greetings {name}' | loc(name => client.name) %]

=head1 DESCRIPTION

This module helps maintaining the POT files which list translatable
strings from template files (or other flat text files) by updating the
list of message-ids which are kept in them.

After initiation, the L<process()|Log::Report::Extract::Template/"Processors"> method needs to be called for each file
in the domain  and the existing PO files will get updated accordingly.

If no translations exist yet, one C<$textdomain.po> file will be
created as point to start.  Copy that file into C<$textdomain/$lang.po>

Extends L<"DESCRIPTION" in Log::Report::Extract|Log::Report::Extract/"DESCRIPTION">.
 
=head1 METHODS

Extends L<"METHODS" in Log::Report::Extract|Log::Report::Extract/"METHODS">.
 
=head2 Constructors

Extends L<"Constructors" in Log::Report::Extract|Log::Report::Extract/"Constructors">.
 
=over 4

=item Log::Report::Extract::Template-E<gt>B<new>(%options)

 -Option --Defined in          --Default
  charset  Log::Report::Extract  'utf-8'
  domain                         <required>
  lexicon  Log::Report::Extract  <required>
  pattern                        <undef>

=over 2

=item charset => STRING

=item domain => DOMAIN

There is no syntax for specifying domains in templates (yet), so you
must be explicit about the collection we are making now.

=item lexicon => DIRECTORY

=item pattern => PREDEFINED|CODE

See the DETAILS section below for a detailed explenation.

=back

=back

=head2 Accessors

Extends L<"Accessors" in Log::Report::Extract|Log::Report::Extract/"Accessors">.
 
=over 4

=item $obj-E<gt>B<addPot>($domain, $pot, %options)

Inherited, see L<Log::Report::Extract/"Accessors">

=item $obj-E<gt>B<charset>()

Inherited, see L<Log::Report::Extract/"Accessors">

=item $obj-E<gt>B<domain>()

=item $obj-E<gt>B<domains>()

Inherited, see L<Log::Report::Extract/"Accessors">

=item $obj-E<gt>B<index>()

Inherited, see L<Log::Report::Extract/"Accessors">

=item $obj-E<gt>B<pattern>()

=item $obj-E<gt>B<pots>($domain)

Inherited, see L<Log::Report::Extract/"Accessors">

=back

=head2 Processors

Extends L<"Processors" in Log::Report::Extract|Log::Report::Extract/"Processors">.
 
=over 4

=item $obj-E<gt>B<cleanup>(%options)

Inherited, see L<Log::Report::Extract/"Processors">

=item $obj-E<gt>B<process>($filename, %options)

Update the domains mentioned in the $filename.  All textdomains defined
in the file will get updated automatically, but not written before
all files where processed.

 -Option --Default
  charset  'utf-8'
  pattern  <from new(pattern)>

=over 2

=item charset => STRING

The character encoding used in this template file.

=item pattern => PREDEFINED|CODE

Read the DETAILS section about this.

=back

=item $obj-E<gt>B<showStats>( [$domains] )

Inherited, see L<Log::Report::Extract/"Processors">

=item $obj-E<gt>B<store>( $domain, $filename, $linenr, $context, $msg, [$msg_plural] )

Inherited, see L<Log::Report::Extract/"Processors">

=item $obj-E<gt>B<write>( [$domain] )

Inherited, see L<Log::Report::Extract/"Processors">

=back

=head1 DETAILS

=head2 Scan Patterns

Various template systems use different conventions for denoting strings
to be translated.

=head3 Predefined for Template-Toolkit

There is not a single convention for translations in C<Template-Toolkit>
(see Template), so you need to specify which version TT you use and
which function name you want to use.  In extreme cases, you may even build
separate translation tables by simply providing using functions.

For instance

   pattern => 'TT2-loc'

will scan for

  [% loc("msgid", key => value, ...) %]
  [% loc('msgid', key => value, ...) %]
  [% loc("msgid|plural", count, key => value, ...) %]
 
  [% INCLUDE
       title = loc('something')
   %]
 
  [% | loc(n => name) %]hi {n}[% END %]
  [% 'hi {n}' | loc(n => name) %]

For TT1, the brackets can either be '[%...%]' or '%%...%%'.  The function
name is treated case-sensitive.  Some people prefer 'l()' or 'L()'.

The code needed

  # during initiation of the webserver, once in your script (before fork)
  my $lexicons   = 'some-directory-for-translation-tables';
  my $translator = Log::Report::Translator::POT->new(lexicons => $lexicons);
  my $domain     = textdomain $textdomain;
  $domain->configure(translator => $translator);

  # your standard template driver
  sub handler {
     ...
     my $vars      = { ...all kinds of values... };
     $vars->{loc}  = \&translate;           # <--- this is extra

     my $output    = '';
     my $templater = Template->new(...);
     $templater->process($template_fn, $vars, \$output);
     print $output;
  }

  # anywhere in the same file
  sub translate {
    my $textdomain = ...;   # your choice when running xgettext-perl
    my $lang       = ...;   # how do you figure that out?
    my $msg = Log::Report::Message->fromTemplateToolkit($textdomain, @_);
    $msg->toString($lang);
  }

To generate the pod tables, run in the shell something like

  xgettext-perl -p $lexicons --template TT2-loc \
      --domain $textdomain  $templates_dir

If you want to implement your own extractor --to avoid C<xgettext-perl>--
you need to run something like this:

  my $extr = Log::Report::Extract::Template->new
    ( lexicon => $output
    , charset => 'utf-8'
    , domain  => $domain
    , pattern => 'TT2-loc'
    );
  $extr->process($_) for @filenames;
  $extr->write;

=head2 Use in combination with contexts

This example extends the previous with using context sensitive translations,
as implemented by L<Log::Report::Translator::Context|Log::Report::Translator::Context>.

Let's say that the translation of some of the sentences on the website depend
on the gender of the addressed person.  An example of the use in a TT2
template:

  [% loc("{name<gender} forgot his key", name => person.name) %]

The extraction script F<xgettext-perl> will expand this into two records
in the PO file, respectively with msgctxt attribute 'gender=male' and
'gender=female'.

When your PO-files are not generated by 'xgettext-perl', you do not need
a separate domain configuration file:

  $domain->configure
    ( context_rules => +{gender => ['male','female']}
    , translator    => $translator
    );

When your PO-files are generated by 'xgettext-perl', you need to share
the context-rules between that msgid extractor and your runtime code. That
same file needs to be passed with the 'domain' parameter to the script.

  # add context_rules either explicit or via 'config' filename
  $domain->configure
    ( config     => 'my/own/$domain.conf'
    , translator => $translator
    );

Now, when you generate the pages, you need to set-up the right context.
In this case, we set-up the gender of the person who gets addressed.
(The name 'gender' is good for examples, but quite non-descriptive.
Maybe 'user_gender' is more maintainable)

  $domain->setContext( +{gender => 'male'} );  # or ('gender=male')
  $domain->setContext( "gender=male" );        # same

=head1 SEE ALSO

This module is part of Log-Report-Lexicon distribution version 1.12,
built on April 18, 2025. Website: F<http://perl.overmeer.net/CPAN/>

=head1 LICENSE

Copyrights 2007-2025 by [Mark Overmeer <markov@cpan.org>]. For other contributors see ChangeLog.

This program is free software; you can redistribute it and/or modify it
under the same terms as Perl itself.
See F<http://dev.perl.org/licenses/>