Rewrite split script in perl

bin/build/split.awk

@@ -1,135 +0,0 @@
-#!/usr/bin/gawk -f
-
-# Note: run this via "update.sh" rather than directly.
-#
-# Split the master markdown file into separate files for each page.
-#
-#  - Creates a directory structure reflecting the path hierarchy of the
-#    pages.
-#  - Prepends Gatsby frontmatter to each file.
-#  - Rewrites links for image files so that the origial file can use
-#    one path and the split files another.
-#
-# Rules:
-#  - New pages can start at markdown heading levels 1, 2, or 3.
-#  - To start a new page, append " <!-- /path/to/page -->" to the heading.
-#  - The file containing the page will end up in "./md/pages/path/to/page.md"
-#  - For the page to be marked "hidden", append a "*" to the path.
-#  - Images matching /<img src="md.*"/ will have their paths naively rewritten.
-
-BEGIN{
-    n = 0
-    filename_prefix = "md/pages"
-    filename = filename_prefix "/error.md"
-    h_part = ""
-    h_chapter = ""
-    h_part_no = -1 # Number parts from 0
-    h_chapter_no = 0
-    h_section_no = 0
-}
-
-# Headings with an HTML comment at the end trigger a new page
-/^(# |## |### ).* <!-- .* -->$/ {
-
-    # Start a new page
-    if (n > 0) {
-        close (filename)
-    }
-    n++
-
-    # Generate frontmatter contents
-    name = gensub(/^#+ (.*) <!-- .* -->$/, "\\1", "1")
-    h_path = gensub(/^#+ .* <!-- (.*) -->$/, "\\1", "1")
-    heading = gensub (/^(#+ .*) <!-- .* -->$/, "\\1", "1")
-
-    # Is this page hidden?
-    if (h_path ~ /\*$/) {
-        h_path = substr(h_path, 1, length(h_path) - 1)
-        h_hide = "true"
-    } else {
-        h_hide = "false"
-    }
-
-    # Make filesystem path for writing the file
-    file_path = h_path
-    sub(/\/[^/]+$/, "", file_path)
-    system("mkdir -p " filename_prefix file_path " 2>/dev/null")
-    filename = filename_prefix h_path ".md"
-
-    print filename
-        
-    switch ($0) {
-    case /^# /:
-        h_part = name
-        h_chapter = ""
-        h_section = ""
-        h_part_no++
-        h_chapter_no = 0
-        idx = h_part_no
-        break
-    case /^## /:
-        h_chapter = name
-        h_section = ""
-        h_chapter_no++
-        h_section_no = 0
-        idx = h_part_no "," h_chapter_no
-        break
-    case /^### /:
-        h_section = name
-        h_section_no++
-        idx = h_part_no "," h_chapter_no "," h_section_no
-        break
-    default:
-        print "Internal error"
-        exit (1)
-    }
-    print "---" > filename
-    print "hide: " h_hide > filename
-    print "path: " h_path > filename
-    print "titles: [\"" h_part "\",\"" h_chapter "\",\"" h_section "\"]" > filename
-    print "index: [" idx "]" > filename
-    print "sequence: " n > filename
-    print "---" > filename
-
-    if (h_section != "") {
-        print "\n<div class=\"section-header\">\n" > filename
-        print "# " h_part "\n" > filename
-        print "## " h_chapter > filename
-        print "\n</div>\n" > filename
-        print "### " h_section > filename
-    } else if (h_chapter != "") {
-        print "\n<div class=\"chapter-header\">\n" > filename
-        print "# " h_part > filename
-        print "\n</div>\n" > filename
-        print "## " h_chapter > filename
-    } else {
-        print "# " h_part > filename
-    }
-
-    next
-}
-
-# Rewrite image paths to reflect the directory hierarchy: Markdown format
-/!\[.*\]\(md.+\)/ {
-    prefix = substr(h_path, 2)
-    gsub(/[^/]*/, "..", prefix)
-    sub(/\(md/, "(" prefix, $0)
-    print > filename
-
-    next
-}
-
-# Rewrite image paths to reflect the directory hierarchy: HTML format (deprecated)
-/<img src="md.*"/ {
-    prefix = substr(h_path, 2)
-    gsub(/[^/]*/, "..", prefix)
-    sub(/src="md/, "src=\"" prefix, $0)
-    print > filename
-
-    next
-}
-
-# Pass through everything else as-is
-{
-    print > filename
-}

bin/build/split.pl

@@ -0,0 +1,107 @@
+#!/usr/bin/perl
+
+use strict;
+use warnings;
+use Fcntl qw(SEEK_SET SEEK_CUR SEEK_END);
+use File::Path qw(make_path);
+
+my ($file) = @ARGV;
+die "Usage: $0 FILE\n" if not $file;
+
+open my $fh, '<', $file or die "Can't open $file: $!";
+
+my $outFilePrefix = 'md/pages';
+my $sequence = 0;
+my $thisPart = '';
+my $thisChapter = '';
+my $thisSection = '';
+my $thisPartNo = -1; # Number parts from 0
+my $thisChapterNo = 0;
+my $thisSectionNo = 0;
+my $thisPath;
+my $idx;
+my $ofh;
+
+while( my $line = <$fh>)  {
+
+    if ($line =~ /^(#{1,3} )(.*) <!-- ([^*]*)(\*?) -->$/) {
+
+        my $level = $1;
+        my $title = $2;
+        my $path = $3;
+        my $hide = $4 eq '*' ? 'true' : 'false';
+
+        $sequence++;
+
+        my $outFile = $outFilePrefix . $path . '.md';
+        my $outDirectory = $outFile =~ s|/[^/]+$||gr;
+
+        unless(-e $outDirectory or make_path($outDirectory)) {
+            die "Unable to create $outDirectory\n";
+        }
+
+        close $ofh if defined $ofh;
+        open $ofh, '>', $outFile or die "Can't open $file for writing: $!";
+
+        $thisPath = $path;
+        if ($level eq '# ') {
+            $thisPart = $title;
+            $thisChapter = '';
+            $thisSection = '';
+            $thisPartNo++;
+            $thisChapterNo = 0;
+            $idx = $thisPartNo;
+        } elsif ($level eq '## ') {
+            $thisChapter = $title;
+            $thisSection = '';
+            $thisChapterNo++;
+            $thisSectionNo = 0;
+            $idx = $thisPartNo . ',' . $thisChapterNo;
+        } elsif ($level eq '### ') {
+            $thisSection = $title;
+            $thisSectionNo++;
+            $idx = $thisPartNo . ',' . $thisChapterNo . ',' . $thisSectionNo;
+        } else {
+            die "Internal error: can't determine heading level.";
+        }
+
+        print $ofh
+            "---\n",
+            "hide: $hide\n",
+            "path: $path\n",
+            "titles: [\"$thisPart\",\"$thisChapter\",\"$thisSection\"]\n",
+            "index: [$idx]\n",
+            "sequence: $sequence\n",
+            "---\n";
+
+        if ($thisSection ne '') {
+            print $ofh
+                "\n<div class=\"section-header\">\n\n",
+                "# $thisPart\n\n",
+                "## $thisChapter\n",
+                "\n</div>\n\n",
+                "### $thisSection\n";
+        } elsif ($thisChapter ne '') {
+            print $ofh
+                "\n<div class=\"chapter-header\">\n\n",
+                "# $thisPart\n",
+                "\n</div>\n\n",
+                "## $thisChapter\n";
+        } else {
+            print $ofh "# $thisPart\n";
+        }
+
+    } else {
+
+        die "Error: first line of input must be a new page marker" if not defined $ofh;
+
+        # Rewrite any markdown image paths to reflect the directory hierarchy
+        (my $prefix = substr $thisPath, 2) =~ s|[^/]+|..|g;
+        $line =~ s/\(md/($prefix/;
+
+        print $ofh $line;
+
+    }
+}
+
+close $ofh if defined $ofh;

bin/build/update.sh

@@ -7,7 +7,7 @@ cd "$(dirname "$0")/../../src"
 
 # Build the markdown pages
 rm -rf md/pages/
-../bin/build/split.awk book.md
+../bin/build/split.pl book.md
 
 # Build the one page annotated spec
 rm -f md/annotated.md