This is the historical archive of the now-inactive 'grubstreet' list.
Discussion on OpenGuides development has now moved to OpenGuides-Dev. Discussion on The Open Guide to London now takes place on OpenGuides-London.

[grubstreet] Wiki hoover script

[prev] [thread] [next] [lurker] [Date index for 2003/1/1]

From: Ivor Williams
Subject: [grubstreet] Wiki hoover script
Date: 20:42 on 01 Jan 2003
------=_NextPart_000_0011_01C2B1D6.4F73C4E0
Content-Type: multipart/alternative;
	boundary="----=_NextPart_001_0012_01C2B1D6.4F7B6600"


------=_NextPart_001_0012_01C2B1D6.4F7B6600
Content-Type: text/plain;
	charset="Windows-1252"
Content-Transfer-Encoding: quoted-printable

For those who are interested


------=_NextPart_001_0012_01C2B1D6.4F7B6600
Content-Type: text/html;
	charset="Windows-1252"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=3DContent-Type content=3D"text/html; =
charset=3Dwindows-1252">
<META content=3D"MSHTML 6.00.2800.1106" name=3DGENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=3D#ffffff>
<DIV><FONT size=3D2>For those who are interested</FONT></DIV>
<DIV><FONT size=3D2></FONT>&nbsp;</DIV>
<DIV>&nbsp;</DIV></BODY></HTML>

------=_NextPart_001_0012_01C2B1D6.4F7B6600--

------=_NextPart_000_0011_01C2B1D6.4F73C4E0
Content-Type: text/plain;
	name="wikihoover.pl"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: attachment;
	filename="wikihoover.pl"

#!perl=0A=
=0A=
use strict;=0A=
use warnings;=0A=
=0A=
our $VERSION;=0A=
=0A=
my ($progname) =3D ($0 =3D~ /(\w+)(\.\w*)?$/);=0A=
=0A=
my $revision =3D '1.0';=0A=
$VERSION =3D sprintf("%d.%02d", ($revision =3D~ /(\d+)\.(\d+)/));=0A=
=0A=
use LWP::UserAgent;=0A=
use HTML::TokeParser;=0A=
use HTTP::Request::Common qw(POST);=0A=
=0A=
# Simple arg list for now. May move to Getopt::Long if need to=0A=
=0A=
&usage if @ARGV !=3D 2;=0A=
=0A=
my ($wiki_from,$wiki_to) =3D @ARGV;=0A=
=0A=
# Set up user agent=0A=
=0A=
my $ua =3D new LWP::UserAgent;=0A=
$ua->agent("$progname/$VERSION " . $ua->agent);=0A=
$ua->env_proxy;=0A=
=0A=
# Set up GET request=0A=
=0A=
my $req =3D HTTP::Request->new('GET');=0A=
$req->header('Accept' =3D> 'text/html');=0A=
=0A=
# Get the Wiki index page=0A=
=0A=
$req->url("$wiki_from?action=3Dindex");=0A=
my $res =3D $ua->request($req);=0A=
=0A=
die "Error on index: " . $res->status_line . "\n" =0A=
	unless $res->is_success;=0A=
=0A=
# Get all the links off this page.=0A=
# Ignore RecentChanges and all ?action=3Dfoo type links=0A=
# Deduplicate list via hash %tocopy=0A=
=0A=
my $index =3D $res->content;=0A=
my $indexp =3D HTML::TokeParser->new(\$index) or die "Index: bad HTML";=0A=
my %tocopy;=0A=
while (my $link =3D $indexp->get_tag('a')) {=0A=
	my ($atag,$aatt) =3D @$link;=0A=
	next unless $aatt->{href} =3D~ /\?([^&=3D]+)$/;=0A=
	my $qs =3D $1;=0A=
	$tocopy{$qs}++ unless $qs eq 'RecentChanges';=0A=
}=0A=
=0A=
# Hoover each page=0A=
=0A=
for (keys %tocopy) {=0A=
=0A=
# Bring up page on source Wiki in edit mode=0A=
	$req->url("$wiki_from?action=3Dedit&id=3D$_");=0A=
	my $srcpage =3D $ua->request($req);=0A=
	next unless $srcpage->is_success;=0A=
=0A=
# Parse out the page text=0A=
	my $srcpc =3D $srcpage->content;=0A=
	my $srcp =3D HTML::TokeParser->new(\$srcpc) or (print "Problems with =
$_\n"),next;=0A=
	my $field =3D $srcp->get_tag('textarea');=0A=
	my $text =3D $srcp->get_text('/textarea');=0A=
=0A=
# Bring up edit mode on destination Wiki to get remaining fields=0A=
	$req->url("$wiki_to?action=3Dedit&id=3D$_");=0A=
	my $edpage =3D $ua->request($req);=0A=
	next unless $edpage->is_success;=0A=
	my $edpc =3D $edpage->content;=0A=
=0A=
# Parse out form to get the remaining fields=0A=
	my $edp =3D HTML::TokeParser->new(\$edpc) or (print "Problems with =
$_\n"),next;=0A=
	my @form;=0A=
	while (my $field =3D $edp->get_tag('input','textarea')) {=0A=
		my ($tag,$attr) =3D @$field;=0A=
		if ($tag eq 'textarea') {=0A=
#			my $text =3D $edp->get_text('/textarea');=0A=
			push @form,$attr->{name},$text;=0A=
			next;=0A=
		}=0A=
		if ($attr->{type} eq 'hidden') {=0A=
			push @form,$attr->{name},$attr->{value};=0A=
		}=0A=
	}=0A=
	=0A=
# additional field values=0A=
	push @form,=0A=
		'summary' =3D> 'Bulk Populate',=0A=
		'Save' =3D> 'Save';=0A=
=0A=
# Do the post	=0A=
	my $ureq =3D POST("$wiki_to",\@form);=0A=
	my $upd =3D $ua->request($ureq);=0A=
	print "Copied: $_\n";=0A=
}=0A=
=0A=
sub usage {=0A=
=0A=
print STDERR <<EOF;=0A=
=0A=
Usage: wikihoover.pl fromsite tosite=0A=
=0A=
EOF=0A=
=0A=
exit 1;=0A=
=0A=
}=0A=
=0A=
=3Dhead1 NAME=0A=
=0A=
	wikihoover.pl - copy a Usemod Wiki to another site=0A=
=0A=
=3Dhead1 SYNOPSIS=0A=
=0A=
	wikihoover.pl fromsite tosite=0A=
=0A=

------=_NextPart_000_0011_01C2B1D6.4F73C4E0--


-- 
grubstreet mailing list
http://london.openguides.org/old-list-archives/