#!/usr/bin/env perl # Parse builds/versions-list.html etc. and automatically create versions.xml RSS feed # Format based on RSS 2.0; see http://cyber.law.harvard.edu/rss/rss.html use strict; use warnings; use Date::Manip; use HTTP::Date; use File::Basename; my $debug = 0; my $urlPrefix = "http://genome-test.soe.ucsc.edu/builds"; sub usage { print STDERR < example: buildVersionsRss.pl ~/browser/builds/versions-list.html > /usr/local/apache/htdocs/builds/versions.xml END exit(1); } my $versionsFile = shift(@ARGV) or usage(); my $path = dirname($versionsFile); my $versionFile = basename($versionsFile); if($path) { chdir($path) or die "Cannot chdir into $path: $!"; } open(VERSIONS, $versionsFile) or die "cannot open file: $versionsFile: $!"; print < UCSC Genome Browser Versions $urlPrefix/versions.html en-us @{[time2str(time())]} END while() { if(/versions-all\/v(\d+)-(final|preview)\.html/) { my ($version, $prefix) = ($1, $2); print STDERR "$1\t$2\n" if($debug); my $file = "versions-all/v$version-$prefix.html"; if(open(HTML, $file)) { my $html = join("\n", ); my ($title, $date); # Parse out

name - version

if($html =~ /

\s(\S+)\s<\/H1>/) { # title only $title = $1; # Try to extract date from CVS log, but ignore those that were entered via bulk entry on 2006-07-26 my $cvsLog = `cvs log $file`; if($cvsLog =~ /date: (.*?);/ && !($cvsLog =~ /Adding these to CVS for safekeeping/)) { printf STDERR "found title: $title in cvsLog\n" if($debug); $date = str2time($1); my @tmp = gmtime($date); my @months = qw(Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec); $title = "$title - " . $tmp[3] . " " . $months[$tmp[4]] . ". " . (1900 + $tmp[5]); } } else { my $originalDate; # if ($html =~ /

\s*(\S+)\s+-\s+(.+)<\/H1>/) { if ($html =~ /

\s(\S+)\s+-(.+?)\s<\/H1>/) { # Single line; e.g.

v140-final - 08-15-2006

in v140-final.html $title = $1; $originalDate = $2; $title = "$title - $originalDate"; } elsif($html =~ /

\s(\S+)\s+-\s+(.?)\s\n/) { # newer multi-line format (e.g. v180-final.html) $title = $1; $originalDate = $2; $title = "$title - $originalDate"; } if($originalDate) { # try to deal with all the date formats donna uses. my $str; if($date = ParseDate($originalDate)) { # e.g. "12 June 2007" $str = $date; $str =~ s/(\d{8})(.)/$1T$2/; } else { if($originalDate =~ /(\d+)(\D+)\.\s+(\d{4})/) { # e.g. "16 Jan. 2007" $str = "$2 $1 $3"; } elsif($originalDate =~ /(\d+)-(\d+)-(\d{4})/) { # Handle "08-15-2006" $str = "$3-$1-$2"; } else { $str = $originalDate; } } if(!($date = str2time($str))) { die "Couldn't parse $str into HTTP::Date style date"; } } } if($title && $date) { print STDERR "$title\n" if($debug); # RSS 2.0 requires RFC 822 format (which is similar to RFC 1123, which # supported by HTTP::Date). $date = time2str($date); print < $title $urlPrefix/$file $date END } elsif (($version <= 137) || ($version == 138 && $prefix eq 'preview')) { # file syntax is not consistent before 140-final and cvs logs are either missing or due to a mass checkin print STDERR "Ignoring older version: $version-$prefix\n" if($debug); } else { if($debug) { print STDERR "Couldn't extract date from $file\n"; } else { die "Couldn't extract date from $file"; } } } else { if($debug) { print STDERR "can't open file: $file\n"; } else { die "can't open file: $file: $!"; } } } } print < END

name - version

\s*(\S+)\s+-\s+(.+)<\/H1>/) { if ($html =~ /

\s*(\S+)\s+-(.+?)\s*<\/H1>/) { # Single line; e.g.

v140-final - 08-15-2006

\s(\S+)\s+-(.+?)\s<\/H1>/) { # Single line; e.g.