#!/usr/bin/perl

use strict;
use DBI;

my ($host, $start_page, $user, $dbase, $pass) = @ARGV;

die "Usage: $0 <host> [start_page] [database] [user] [password]"
  unless ($host);

$start_page ||= 'Main_Page';
$dbase ||= 'wikidb';
$user ||= 'root';

my $db = DBI->connect("DBI:mysql:database=$dbase;host=$host", $user, $pass);

my $page = get_page($db, $start_page);
transform_page($page, translate_page($page));
output_page($page);
system "ci -u -t-'Intitial revision imported from MediaWiki' *.txt";
chmod 0644, glob("*.txt");

sub get_page {
    my ($db, $title, $namespace) = @_;

    my $cmd;

    # First find the page_id of the page
    if ($namespace) {
	$cmd = $db->prepare("SELECT * from mw_page where page_title = ?" .
			       "AND page_namespace = ?");
	$cmd->execute($title, $namespace);
    } else {
	$cmd = $db->prepare("SELECT * from mw_page where page_title = ?");
	$cmd->execute($title);
    }

    return undef unless $cmd->rows;
    my $pg = $cmd->fetchrow_hashref();
    
    $cmd = $db->prepare("SELECT * from mw_revision where rev_id = ?");
    $cmd->execute($pg->{page_latest});

    die "Couldn't find last revision $pg->{page_latest} for page $title"
      unless $cmd->rows;
    my $rev = $cmd->fetchrow_hashref();

    $cmd = $db->prepare("SELECT * from mw_text where " .
			 "old_id = $rev->{rev_text_id}");
    $cmd->execute;

    die "Couldn't find last revision $pg->{page_latest} for page $title"
      unless $cmd->rows;

    my $ret = { title => $title,
		text => $cmd->fetchrow_hashref()->{old_text},
		namespace => $pg->{page_namespace},
		links => [] };

    $cmd = $db->prepare("SELECT * from mw_pagelinks where " .
			 "pl_from = $pg->{page_id}");
    $cmd->execute;
    while (my $link = $cmd->fetchrow_hashref()) {
	my $sub_page = get_page($db, $link->{pl_title}, $link->{pl_namespace});
	push @{$ret->{links}}, { title => $link->{pl_title},
				 page => $sub_page  } if $sub_page;
    }

    $ret;
}

sub translate_page {
    my ($page, $file_transforms) = @_;

    $file_transforms ||= [];

    if ($page->{title} =~ /[\._']/) {
	$page->{link_name} = $page->{title};
	$page->{title} =~ s/_/ /g;
	$page->{link_name} =~ s/[\.']//g;
	push @$file_transforms, [ $page->{title}, $page->{link_name} ];
    }

    # These regular expressions were snarfed from:
    #     http://twiki.org/cgi-bin/view/Plugins/MediawikiEditSyntaxRegex

    $_ = $page->{text};

    s/(^|[\n\r])\; *([^:]*):/$1   \$ $2\:/g; # ; definition : term
    s/(^|[\n\r])[\#\*]{9}\# /$1                              1. /g; # level 10 bullet
    s/(^|[\n\r])[\#\*]{8}\# /$1                           1. /g; # level 9 bullet
    s/(^|[\n\r])[\#\*]{7}\# /$1                        1. /g; # level 8 bullet
    s/(^|[\n\r])[\#\*]{6}\# /$1                     1. /g; # level 7 bullet
    s/(^|[\n\r])[\#\*]{5}\# /$1                  1. /g; # level 6 bullet
    s/(^|[\n\r])[\#\*]{4}\# /$1               1. /g; # level 5 bullet
    s/(^|[\n\r])[\#\*]{3}\# /$1            1. /g; # level 4 bullet
    s/(^|[\n\r])[\#\*]{2}\# /$1         1. /g; # level 3 bullet
    s/(^|[\n\r])[\#\*]{1}\# /$1      1. /g; # level 2 bullet
    s/(^|[\n\r])\# /$1   1. /g; # level 1 bullet
    s/(^|[\n\r])[\#\*]{9}\* /$1                              \* /g; # level 10 bullet
    s/(^|[\n\r])[\#\*]{8}\* /$1                           \* /g; # level 9 bullet
    s/(^|[\n\r])[\#\*]{7}\* /$1                        \* /g; # level 8 bullet
    s/(^|[\n\r])[\#\*]{6}\* /$1                     \* /g; # level 7 bullet
    s/(^|[\n\r])[\#\*]{5}\* /$1                  \* /g; # level 6 bullet
    s/(^|[\n\r])[\#\*]{4}\* /$1               \* /g; # level 5 bullet
    s/(^|[\n\r])[\#\*]{3}\* /$1            \* /g; # level 4 bullet
    s/(^|[\n\r])[\#\*]{2}\* /$1         \* /g; # level 3 bullet
    s/(^|[\n\r])[\#\*]{1}\* /$1      \* /g; # level 2 bullet
    s/(^|[\n\r])\* /$1   \* /g; # level 1 bullet
    s/\[\[Image\:([^\]]*)\]\]/\%ATTACHURL\%\/$1/g; # [[Image:file.png]]
    s/\[\[([^\|]*)\|(\1)\]\]/\[\[$1\]\]/g; # internal link [[WikiWord|WikiWord]]
    s/\[\[([^\|]*)\|([^\]]*?)\]\]/\[\[$1\]\[$2\]\]/g; # internal link [[WikiWord|label]]
    s/(?=[^\]]|^)\[(https?\:.*?) (.*?)\](?=[^\]]|$)/\[\[$1\]\[$2\]\]/g; # external link [http:... label]
    s/(^|[\n\r])======(.*?) ?======(?=[\n\r]|$)/$1---\+\+\+\+\+\+$2/g; # H6
    s/(^|[\n\r])=====(.*?) ?=====(?=[\n\r]|$)/$1---\+\+\+\+\+$2/g; # H5
    s/(^|[\n\r])====(.*?) ?====(?=[\n\r]|$)/$1---\+\+\+\+$2/g; # H4
    s/(^|[\n\r])===(.*?) ?===(?=[\n\r]|$)/$1---\+\+\+$2/g; # H3
    s/(^|[\n\r])==(.*?) ?==(?=[\n\r]|$)/$1---\+\+$2/g; # H2
    s/(^|[\n\r])=(.*?) ?=(?=[\n\r]|$)/$1---\+$2/g; # H1
    s/(^|[\s\(])<tt>([^ ].*?[^ ])<\/tt>([\s\)\.\,\:\;\!\?]|$)/$1=$2=$3/g; # monospaced
    s/(^|[\s\(])'''<tt>([^ ].*?[^ ])<\/tt>'''([\s\)\.\,\:\;\!\?]|$)/$1==$2==$3/g; # monospaced bold
    s/(^|[\s\(])'''([^ ].*?[^ ])'''([\s\)\.\,\:\;\!\?]|$)/$1\*$2\*$3/g; # bold
    s/(^|[\s\(])''<b>([^ ].*?[^ ])<\/b>''([\s\)\.\,\:\;\!\?]|$)/$1\_\_$2\_\_$3/g; # italic bold
    s/(^|[\s\(])''([^ ].*?[^ ])''([\s\)\.\,\:\;\!\?]|$)/$1\_$2\_$3/g; # italic

    # Replace the page content
    $page->{text} = $_;

    # ... and recurse
    foreach my $link (@{$page->{links}}) {
	translate_page($link->{page}, $file_transforms);
    }

    $file_transforms;
}

sub transform_page {
    my ($page, $transform_list) = @_;

    foreach my $xform (@$transform_list) {
	my ($from, $to) = @$xform;
	# We need to fixup the input side of the RE to escape the various
	# special RE's.
	my $source = $from;
	$source =~ s/\./\\./g;
	$source =~ s/\(/\\\(/g;
	$source =~ s/\)/\\\)/g;
	$page->{text} =~ s/\[\[$source\]\]/\[\[$to\]\[$from\]\]/g;
    }

    # Place a title at the beginning of the page, like MediaWiki would
    # have automatically done
    $page->{text} = "---+!! $page->{title}\n$page->{text}";

    # ... and recurse
    foreach my $link (@{$page->{links}}) {
	transform_page($link->{page}, $transform_list);
    }
}

sub output_page {
    my $page = shift;

    my $file = $page->{link_name} || $page->{title};
    open(OUT, "> $file.txt");
    print OUT $page->{text};
    close(OUT);

    # ... and recurse
    foreach my $link (@{$page->{links}}) {
	output_page($link->{page});
    }
}