Add IRI filter to beautify (i.e. canonicalize and internationalize) links (including uglifying workaround to modest URL parsing).

2 files changed, 42 insertions, 1 deletions

diff --git a/Makefile b/Makefile
index 07bd8de..a8f9131 100644
--- a/Makefile
+++ b/Makefile
@@ -10,7 +10,7 @@ re_urlencode = s|\bhttp[^\s\"]+?\K(--+)|uri_escape($$1, "-")|eg
 
 templates = template.tex header.tex before.tex after.tex
 
-filters = ./pandoc-filter-sections ./pandoc-filter-bib
+filters = ./pandoc-filter-sections ./pandoc-filter-bib ./pandoc-filter-iri
 
 export PANDOC_CITEPROC_FILE = $(stem).bib
 #PANDOC_CITEPROC_FILE = $(stem).bib
diff --git a/pandoc-filter-iri b/pandoc-filter-iri
new file mode 100755
index 0000000..94be2e4
--- /dev/null
+++ b/pandoc-filter-iri
@@ -0,0 +1,41 @@
+#!/usr/bin/perl
+
+use warnings;
+use strict;
+
+use Pandoc::Filter;
+use Pandoc::Elements;
+
+use URI;
+use URI::Find;
+
+pandoc_filter(
+	\&link_normalize,
+);
+
+sub link_normalize {
+	my $self = shift;
+	return [ Link(
+		$self->content,
+		[ pp_uri($self->target->[0]) => '' ]
+	)]
+		if ( $self->name eq 'Link' );
+	return [ Str pp_string($self->content) ]
+		if ( $self->name eq 'Str' );
+	return;
+}
+
+sub pp_uri {
+	return URI->new(shift)->canonical->as_string;
+}
+
+sub pp_iri {
+	return URI->new(shift)->canonical->as_iri;
+}
+
+sub pp_string {
+	my $string = shift;
+	my $finder = URI::Find->new( \&pp_iri );
+	$finder->find(\$string);
+	return $string;
+}