109 lines
3.4 KiB
Raku
109 lines
3.4 KiB
Raku
#| Interaction with pandoc
|
|
unit module Pandoc;
|
|
|
|
use JSON::Fast;
|
|
|
|
#| Run pandoc with the given arguments, dieing on failure
|
|
sub pandoc(*@args --> Str:D) {
|
|
# Call into pandoc
|
|
my $pandoc = run 'pandoc', '--no-highlight', @args, :out, :err;
|
|
|
|
# Collect the output
|
|
my $output = $pandoc.out.slurp: :close;
|
|
my $stderr = $pandoc.err.slurp: :close;
|
|
die "Pandoc exited with {$pandoc.exitcode}\nout: $output\nerr: $stderr"
|
|
unless $pandoc;
|
|
|
|
$output
|
|
}
|
|
|
|
#| Extract the title from a markdown document
|
|
#|
|
|
#| The title is the only top level header, will throw an error if there are
|
|
#| multiple top level headers or are none
|
|
sub markdown-title(IO::Path:D $file --> Str:D) is export {
|
|
# Collect the output
|
|
my $output = pandoc <-f gfm -t JSON>, $file;
|
|
|
|
# Parse out output from pandoc, we are making an executive decision to trust
|
|
# pandoc here, so we won't do any error handling for pandoc's output
|
|
my %parsed = from-json $output;
|
|
# Extract a list of top level headers from the pandoc output, this should
|
|
# only have one element in it, but as this is user input, its untrusted and
|
|
# we need to do some error handling
|
|
my sub is-header($v) {
|
|
$v ~~ Associative && $v<t> ~~ "Header"
|
|
}
|
|
my @headers = %parsed<blocks>.grep(&is-header).grep(*<c>[0] == 1);
|
|
die "More than one top level header in $file" if @headers.elems > 1;
|
|
die "No top level headers in $file" if @headers.elems == 0;
|
|
|
|
# Extract the header and process it into a string
|
|
my @header = @headers[0]<c>[2].flat;
|
|
my $title = "";
|
|
for @header -> $component {
|
|
next unless $component ~~ Associative;
|
|
given $component<t> {
|
|
when "Str" {
|
|
$title = $title ~ $component<c>;
|
|
}
|
|
when "Space" {
|
|
$title = $title ~ " ";
|
|
}
|
|
default {
|
|
die "Invalid component type: $_";
|
|
}
|
|
}
|
|
}
|
|
|
|
return $title;
|
|
}
|
|
|
|
#| Use pandoc to extract the first paragraph of a markdown document
|
|
sub markdown-first-paragraph(IO::Path:D $file --> Str:D) is export {
|
|
my $output = pandoc <-f gfm -t JSON>, $file;
|
|
my %parsed = from-json $output;
|
|
# Extract a list of paragraphs from the pandoc output
|
|
my sub is-para($v) {
|
|
$v ~~ Associative && $v<t> ~~ 'Para'
|
|
}
|
|
my @paras = %parsed<blocks>.grep(&is-para);
|
|
die "No paragraphs in markdown" if @paras.elems == 0;
|
|
my @para = @paras[0][0]<c>.flat;
|
|
# Proces it into a string
|
|
my $para = "";
|
|
for @para -> $component {
|
|
next unless $component ~~ Associative;
|
|
given $component<t> {
|
|
when "Str" {
|
|
$para ~= $component<c>;
|
|
}
|
|
when "Space" {
|
|
$para ~= ' ';
|
|
}
|
|
when "Code" {
|
|
$para ~= $component<c>[*-1];
|
|
}
|
|
when "SoftBreak" {
|
|
$para ~= "\n";
|
|
}
|
|
when "Link" {
|
|
# TODO: Properly descend into links
|
|
$para ~= $component<c>[1][0]<c>;
|
|
}
|
|
default {
|
|
die "Invalid component type: $_";
|
|
}
|
|
}
|
|
}
|
|
|
|
$para
|
|
}
|
|
|
|
#| Use pandoc to render a markdown document to html
|
|
sub markdown-to-html(IO::Path:D $file --> Str:D) is export {
|
|
# Remove the header, we'll regenerate it later
|
|
my $output = pandoc <-f gfm>, $file;
|
|
$output ~~ s:g/'<h1' .* '</h1>'//;
|
|
$output
|
|
}
|