Convert to Creole

A Perl script to convert from Bo Leuf's Cluster Wiki markup syntax to Creole — more or less! (^_^)

Here's what the code tries to do:

• strip off any Cluster Wiki headers and footers
• change quintuple-tics into properly-nested bold-italic markup
• change triple-tics into double-asterisks (bold)
• change double-tics into double-slashes (italics)
• remove Cluster Wiki markup that indicates small and big type size
• change #+ ... +# to ^^ ... ^^ (superscript)
• change #- ... -# to ,, ... ,, (subscript)
• change $/ to \\ (forced linebreak)
• change double-double-quotes at line-beginning into single colons (blockquotes)
• change !! at beginning of line into == (subhead titles), and likewise for other header levels
• change double-pipes into single-pipes (tables)
• put double-square-brackets around WikiWords
• change [sic] ... [/sic] into triple-curly-braces
• put triple-curly-braces around blocks of lines that begin with spaces

Here are things that the conversion code does NOT attempt to handle:

• changing [esc]...[/esc] into !... or ~... or ?
• fixing definition lists (which in Cluster Wiki use ": ... : ..." syntax)

#! /usr/bin/perl

# convert_to_creole.prl version 0.1 --- ^z --- 25 Dec 2007, 4-5 Jan 2008, ...
# usage:  perl convert_to_creole.prl indir outdir
#
# take all files in "indir", do a bunch of pattern transformations
# globally in every file, store results in "outdir"
# indir and outdir must already exist

print "Converting to Creole:\n";
$indir = $ARGV[0];
$outdir = $ARGV[1];
opendir(INDIR, "$indir") or die "couldn't open input directory $indir";
opendir(OUTDIR, "$outdir") or die "couldn't open output directory $outdir";
@pages = grep !/^\./, readdir INDIR;
undef $/;  # grab entire file at once

foreach $page (@pages) {
  if ( -e "$indir/$page" ) {
    open(F, "$indir/$page") or die "$page: $!";
    print "  $page ... ";
    $body = <F>;
    close(F);

# delete Cluster Wiki headers and footers
    $body =~ s/^.*\263text\263//;
    $body =~ s/\263.*$//;

# put a newline at the beginning of the file, to simplify logic below
    $body = "\n$body";

# patterns to transform whole body, globally, in proper order
    $body =~ s/\'\'\'\'\'(.*?)\'\'\'\'\'/\*\*\/\/\1\/\/\*\*/sog; # nested bold-italics
    $body =~ s/\'\'\'(.*?)\'\'\'/\*\*\1\*\*/sog;  # '''stuff''' -> **stuff**
    $body =~ s/\'\'(.*?)\'\'/\/\/\1\/\//sog; # ''stuff'' -> //stuff//
    $body =~ s/\[\-(.*?)\-\]/\1/sog; # [- stuff -] -> stuff
    $body =~ s/\[\+(.*?)\+\]/\1/sog; # [+ stuff +] -> stuff
    $body =~ s/\#\+(.*?)\+\#/\^\^\1\^\^/sog; # #+ ... +# -> ^^ ... ^^
    $body =~ s/\#\-(.*?)\-\#/\,\,\1\,\,/sog; # #- ... -# -> ,, ... ,,
    $body =~ s/\$\//\\\\/sog; # $/ -> \\
    $body =~ s/\n\"\"/\n\:/sog; # ""stuff -> :stuff (after newline)
    $body =~ s/\n\!\!\!\!\!\!/\n\=\=\=\=\=\=/sog; # !!!!!!stuff -> ======stuff
    $body =~ s/\n\!\!\!\!\!/\n\=\=\=\=\=/sog; # !!!!!stuff -> =====stuff
    $body =~ s/\n\!\!\!\!/\n\=\=\=\=/sog; # !!!!stuff -> ====stuff
    $body =~ s/\n\!\!\!/\n\=\=\=/sog; # !!!stuff -> ===stuff
    $body =~ s/\n\!\!/\n\=\=/sog; # !!stuff -> ==stuff
    $body =~ s/\n\"\"/\n\:/sog; # ""stuff -> :stuff
    $body =~ s/\|\|/\|/sog; # || -> | (table syntax fix)
    $body =~ s/([A-Z][a-z0-9]+([A-Z][a-z0-9]+)+)/\[\[\1\]\]/sog; # WikiWord -> [[WikiWord]]
    $body =~ s/\[sic\]/\{\{\{/sog;  # [sic] -> {{{
    $body =~ s/\[\/sic\]/\}\}\}/sog; # [/sic] -> }}}
    $body =~ s/((\n [^\n]*)+)/\n\{\{\{\1\n\}\}\}/sog; # put {{{ ... }}} around block

# remove newline added to beginning of file
    $body =~ s/\n(.*)/\1/sog;

    open(F, ">$outdir/$page") or die "$page: $!";
    print F $body;
    close(F);
  } else {
      die "$page didn't exist: $!";
  }
  print "\n";
}

(correlates: Multiword link, SnipPattern01SourceCode, ZhurnalyWiki, ...)