| 1 | #! /usr/bin/perl -W\r |
| 2 | #\r |
| 3 | # feedlog.pl\r |
| 4 | #\r |
| 5 | # This script will log headers and other data to a database\r |
| 6 | # for further analysis by parsing a feed from INN.\r |
| 7 | # \r |
| 8 | # It is part of the NewsStats package.\r |
| 9 | #\r |
| 10 | # Copyright (c) 2010 Thomas Hochstein <thh@inter.net>\r |
| 11 | #\r |
| 12 | # It can be redistributed and/or modified under the same terms under \r |
| 13 | # which Perl itself is published.\r |
| 14 | \r |
| 15 | BEGIN {\r |
| 16 | our $VERSION = "0.01";\r |
| 17 | use File::Basename;\r |
| 18 | push(@INC, dirname($0));\r |
| 19 | }\r |
| 20 | use strict;\r |
| 21 | \r |
| 22 | use NewsStats;\r |
| 23 | \r |
| 24 | use Sys::Syslog qw(:standard :macros);\r |
| 25 | \r |
| 26 | use Date::Format;\r |
| 27 | use DBI;\r |
| 28 | \r |
| 29 | ################################# Main program #################################\r |
| 30 | \r |
| 31 | ### read commandline options\r |
| 32 | my %Options = &ReadOptions('qd');\r |
| 33 | \r |
| 34 | ### read configuration\r |
| 35 | my %Conf = %{ReadConfig('newsstats.conf')};\r |
| 36 | \r |
| 37 | ### init syslog\r |
| 38 | openlog($MySelf, 'nofatal,pid', LOG_NEWS);\r |
| 39 | syslog(LOG_NOTICE, "$MyVersion starting up.") if !$Options{'q'};\r |
| 40 | \r |
| 41 | ### init database\r |
| 42 | my $DBHandle = InitDB(\%Conf,0);\r |
| 43 | if (!$DBHandle) {\r |
| 44 | syslog(LOG_CRIT, 'Database connection failed: %s', $DBI::errstr);\r |
| 45 | while (1) {}; # go into endless loop to suppress further errors and respawning\r |
| 46 | };\r |
| 47 | my $DBQuery = $DBHandle->prepare(sprintf("INSERT INTO %s.%s (day,date,mid,timestamp,token,size,peer,path,newsgroups,headers) VALUES (?,?,?,?,?,?,?,?,?,?)",$Conf{'DBDatabase'},$Conf{'DBTableRaw'}));\r |
| 48 | \r |
| 49 | ### main loop\r |
| 50 | while (<>) {\r |
| 51 | chomp;\r |
| 52 | # catch empty lines trailing or leading\r |
| 53 | if ($_ eq '') {\r |
| 54 | next;\r |
| 55 | }\r |
| 56 | # first line contains: mid, timestamp, token, size, peer, Path, Newsgroups\r |
| 57 | my ($Mid, $Timestamp, $Token, $Size, $Peer, $Path, $Newsgroups) = split;\r |
| 58 | # remaining lines contain headers\r |
| 59 | my $Headers = "";\r |
| 60 | while (<>) {\r |
| 61 | chomp;\r |
| 62 | # empty line terminates this article\r |
| 63 | if ($_ eq '') {\r |
| 64 | last;\r |
| 65 | }\r |
| 66 | # collect headers\r |
| 67 | $Headers .= $_."\n" ;\r |
| 68 | }\r |
| 69 | \r |
| 70 | # parse timestamp to day (YYYY-MM-DD) and to MySQL timestamp\r |
| 71 | my $Day = time2str("%Y-%m-%d", $Timestamp);\r |
| 72 | my $Date = time2str("%Y-%m-%d %H:%M:%S", $Timestamp);\r |
| 73 | \r |
| 74 | # write to database\r |
| 75 | if (!$DBQuery->execute($Day, $Date, $Mid, $Timestamp, $Token, $Size, $Peer, $Path, $Newsgroups, $Headers)) {\r |
| 76 | syslog(LOG_ERR, 'Database error: %s', $DBI::errstr);\r |
| 77 | };\r |
| 78 | $DBQuery->finish;\r |
| 79 | \r |
| 80 | warn sprintf("-----\nDay: %s\nDate: %s\nMID: %s\nTS: %s\nToken: %s\nSize: %s\nPeer: %s\nPath: %s\nNewsgroups: %s\nHeaders: %s\n",$Day, $Date, $Mid, $Timestamp, $Token, $Size, $Peer, $Path, $Newsgroups, $Headers) if !$Options{'d'};\r |
| 81 | }\r |
| 82 | \r |
| 83 | ### close handles\r |
| 84 | $DBHandle->disconnect;\r |
| 85 | syslog(LOG_NOTICE, "$MySelf closing down.") if !$Options{'q'};\r |
| 86 | closelog();\r |
| 87 | \r |
| 88 | __END__\r |
| 89 | \r |
| 90 | ################################ Documentation #################################\r |
| 91 | \r |
| 92 | =head1 NAME\r |
| 93 | \r |
| 94 | feedlog - log data from an INN feed to a database\r |
| 95 | \r |
| 96 | =head1 SYNOPSIS\r |
| 97 | \r |
| 98 | B<feedlog> [B<-Vhdq>]\r |
| 99 | \r |
| 100 | =head1 REQUIREMENTS\r |
| 101 | \r |
| 102 | See doc/README: Perl 5.8.x itself and the following modules from CPAN:\r |
| 103 | \r |
| 104 | =over 2\r |
| 105 | \r |
| 106 | =item -\r |
| 107 | \r |
| 108 | Config::Auto\r |
| 109 | \r |
| 110 | =item -\r |
| 111 | \r |
| 112 | Date::Format\r |
| 113 | \r |
| 114 | =item -\r |
| 115 | \r |
| 116 | DBI\r |
| 117 | \r |
| 118 | =back\r |
| 119 | \r |
| 120 | =head1 DESCRIPTION\r |
| 121 | \r |
| 122 | This script will log overview data and complete headers to a database\r |
| 123 | table for further examination by parsing a feed from INN. It will\r |
| 124 | parse that information and write it to a mysql database table in real\r |
| 125 | time.\r |
| 126 | \r |
| 127 | All reporting is done to I<syslog> via I<news> facility. If B<feedlog>\r |
| 128 | fails to initiate a database connection at startup, it will log to\r |
| 129 | I<syslog> with I<CRIT> priority and go in an endless loop, as\r |
| 130 | terminating would only result in a rapid respawn.\r |
| 131 | \r |
| 132 | =head2 Configuration\r |
| 133 | \r |
| 134 | F<feedlog.pl> will read its configuration from F<newsstats.conf> which\r |
| 135 | should be present in the same directory via Config::Auto.\r |
| 136 | \r |
| 137 | See doc/INSTALL for an overview of possible configuration options.\r |
| 138 | \r |
| 139 | =head1 OPTIONS\r |
| 140 | \r |
| 141 | =over 3\r |
| 142 | \r |
| 143 | =item B<-V> (version)\r |
| 144 | \r |
| 145 | Print out version and copyright information on B<yapfaq> and exit.\r |
| 146 | \r |
| 147 | =item B<-h> (help)\r |
| 148 | \r |
| 149 | Print this man page and exit.\r |
| 150 | \r |
| 151 | =item B<-d> (debug)\r |
| 152 | \r |
| 153 | Output debugging information to STDERR while parsing STDIN. You'll\r |
| 154 | find that information most probably in your B<INN> F<errlog> file.\r |
| 155 | \r |
| 156 | =item B<-q> (quiet)\r |
| 157 | \r |
| 158 | Suppress logging to syslog.\r |
| 159 | \r |
| 160 | =back\r |
| 161 | \r |
| 162 | =head1 INSTALLATION\r |
| 163 | \r |
| 164 | See doc/INSTALL.\r |
| 165 | \r |
| 166 | =head1 EXAMPLES\r |
| 167 | \r |
| 168 | Set up a feed like that in your B<INN> F<newsfeeds> file:\r |
| 169 | \r |
| 170 | ## gather statistics for NewsStats\r |
| 171 | newsstats!\r |
| 172 | :!*,de.*\r |
| 173 | :Tc,WmtfbsPNH,Ac:/path/to/feedlog.pl\r |
| 174 | \r |
| 175 | See doc/INSTALL for further information.\r |
| 176 | \r |
| 177 | =head1 FILES\r |
| 178 | \r |
| 179 | =over 4\r |
| 180 | \r |
| 181 | =item F<feedlog.pl>\r |
| 182 | \r |
| 183 | The script itself.\r |
| 184 | \r |
| 185 | =item F<NewsStats.pm>\r |
| 186 | \r |
| 187 | Library functions for the NewsStats package.\r |
| 188 | \r |
| 189 | =item F<newsstats.conf>\r |
| 190 | \r |
| 191 | Runtime configuration file for B<yapfaq>.\r |
| 192 | \r |
| 193 | =back\r |
| 194 | \r |
| 195 | =head1 BUGS\r |
| 196 | \r |
| 197 | Please report any bugs or feature requests to the author or use the\r |
| 198 | bug tracker at L<http://bugs.th-h.de/>!\r |
| 199 | \r |
| 200 | =head1 SEE ALSO\r |
| 201 | \r |
| 202 | =over 2\r |
| 203 | \r |
| 204 | =item -\r |
| 205 | \r |
| 206 | doc/README\r |
| 207 | \r |
| 208 | =item -\r |
| 209 | \r |
| 210 | doc/INSTALL\r |
| 211 | \r |
| 212 | =back\r |
| 213 | \r |
| 214 | This script is part of the B<NewsStats> package.\r |
| 215 | \r |
| 216 | =head1 AUTHOR\r |
| 217 | \r |
| 218 | Thomas Hochstein <thh@inter.net>\r |
| 219 | \r |
| 220 | =head1 COPYRIGHT AND LICENSE\r |
| 221 | \r |
| 222 | Copyright (c) 2010 Thomas Hochstein <thh@inter.net>\r |
| 223 | \r |
| 224 | This program is free software; you may redistribute it and/or modify it\r |
| 225 | under the same terms as Perl itself.\r |
| 226 | \r |
| 227 | =cut\r |