Commit | Line | Data |
---|---|---|
2832c235 TH |
1 | #! /usr/bin/perl -W |
2 | # | |
3 | # feedlog.pl | |
4 | # | |
5 | # This script will log headers and other data to a database | |
6 | # for further analysis by parsing a feed from INN. | |
7 | # | |
8 | # It is part of the NewsStats package. | |
9 | # | |
f05ec4ae | 10 | # Copyright (c) 2010-2012 Thomas Hochstein <thh@inter.net> |
2832c235 TH |
11 | # |
12 | # It can be redistributed and/or modified under the same terms under | |
13 | # which Perl itself is published. | |
14 | ||
15 | BEGIN { | |
16 | our $VERSION = "0.01"; | |
17 | use File::Basename; | |
18 | push(@INC, dirname($0)); | |
19 | } | |
20 | use strict; | |
21 | ||
22 | use NewsStats; | |
23 | ||
24 | use Sys::Syslog qw(:standard :macros); | |
25 | ||
26 | use Date::Format; | |
27 | use DBI; | |
f05ec4ae TH |
28 | use Getopt::Long qw(GetOptions); |
29 | Getopt::Long::config ('bundling'); | |
2832c235 TH |
30 | |
31 | ################################# Main program ################################# | |
32 | ||
33 | ### read commandline options | |
f05ec4ae TH |
34 | my ($OptDebug,$OptQuiet); |
35 | GetOptions ('d|debug!' => \$OptDebug, | |
36 | 'q|test!' => \$OptQuiet, | |
37 | 'h|help' => \&ShowPOD, | |
38 | 'V|version' => \&ShowVersion) or exit 1; | |
2832c235 TH |
39 | |
40 | ### read configuration | |
f05ec4ae | 41 | my %Conf = %{ReadConfig($HomePath.'/newsstats.conf')}; |
2832c235 TH |
42 | |
43 | ### init syslog | |
f05ec4ae TH |
44 | openlog($0, 'nofatal,pid', LOG_NEWS); |
45 | syslog(LOG_NOTICE, "$MyVersion starting up.") if !$OptQuiet; | |
2832c235 TH |
46 | |
47 | ### init database | |
48 | my $DBHandle = InitDB(\%Conf,0); | |
49 | if (!$DBHandle) { | |
50 | syslog(LOG_CRIT, 'Database connection failed: %s', $DBI::errstr); | |
51 | while (1) {}; # go into endless loop to suppress further errors and respawning | |
52 | }; | |
f05ec4ae TH |
53 | my $DBQuery = $DBHandle->prepare(sprintf("INSERT INTO %s.%s (day,date,mid, |
54 | timestamp,token,size,peer,path, | |
55 | newsgroups,headers) | |
56 | VALUES (?,?,?,?,?,?,?,?,?,?)", | |
57 | $Conf{'DBDatabase'},$Conf{'DBTableRaw'})); | |
2832c235 TH |
58 | |
59 | ### main loop | |
60 | while (<>) { | |
61 | chomp; | |
62 | # catch empty lines trailing or leading | |
63 | if ($_ eq '') { | |
64 | next; | |
65 | } | |
66 | # first line contains: mid, timestamp, token, size, peer, Path, Newsgroups | |
67 | my ($Mid, $Timestamp, $Token, $Size, $Peer, $Path, $Newsgroups) = split; | |
68 | # remaining lines contain headers | |
69 | my $Headers = ""; | |
70 | while (<>) { | |
71 | chomp; | |
72 | # empty line terminates this article | |
73 | if ($_ eq '') { | |
74 | last; | |
75 | } | |
76 | # collect headers | |
77 | $Headers .= $_."\n" ; | |
78 | } | |
79 | ||
80 | # parse timestamp to day (YYYY-MM-DD) and to MySQL timestamp | |
81 | my $Day = time2str("%Y-%m-%d", $Timestamp); | |
82 | my $Date = time2str("%Y-%m-%d %H:%M:%S", $Timestamp); | |
83 | ||
84 | # write to database | |
f05ec4ae TH |
85 | if (!$DBQuery->execute($Day, $Date, $Mid, $Timestamp, $Token, $Size, $Peer, |
86 | $Path, $Newsgroups, $Headers)) { | |
2832c235 TH |
87 | syslog(LOG_ERR, 'Database error: %s', $DBI::errstr); |
88 | }; | |
89 | $DBQuery->finish; | |
90 | ||
f05ec4ae TH |
91 | warn sprintf("-----\nDay: %s\nDate: %s\nMID: %s\nTS: %s\nToken: %s\n". |
92 | "Size: %s\nPeer: %s\nPath: %s\nNewsgroups: %s\nHeaders: %s\n", | |
93 | $Day, $Date, $Mid, $Timestamp, $Token, $Size, $Peer, $Path, | |
94 | $Newsgroups, $Headers) if $OptDebug; | |
2832c235 TH |
95 | } |
96 | ||
97 | ### close handles | |
98 | $DBHandle->disconnect; | |
f05ec4ae | 99 | syslog(LOG_NOTICE, "$0 closing down.") if !$OptQuiet; |
2832c235 TH |
100 | closelog(); |
101 | ||
102 | __END__ | |
103 | ||
104 | ################################ Documentation ################################# | |
105 | ||
106 | =head1 NAME | |
107 | ||
108 | feedlog - log data from an INN feed to a database | |
109 | ||
110 | =head1 SYNOPSIS | |
111 | ||
112 | B<feedlog> [B<-Vhdq>] | |
113 | ||
114 | =head1 REQUIREMENTS | |
115 | ||
f05ec4ae | 116 | See L<doc/README>. |
2832c235 TH |
117 | |
118 | =head1 DESCRIPTION | |
119 | ||
120 | This script will log overview data and complete headers to a database | |
121 | table for further examination by parsing a feed from INN. It will | |
122 | parse that information and write it to a mysql database table in real | |
123 | time. | |
124 | ||
125 | All reporting is done to I<syslog> via I<news> facility. If B<feedlog> | |
126 | fails to initiate a database connection at startup, it will log to | |
127 | I<syslog> with I<CRIT> priority and go in an endless loop, as | |
128 | terminating would only result in a rapid respawn. | |
129 | ||
130 | =head2 Configuration | |
131 | ||
f05ec4ae | 132 | B<feedlog> will read its configuration from F<newsstats.conf> which |
2832c235 TH |
133 | should be present in the same directory via Config::Auto. |
134 | ||
f05ec4ae | 135 | See L<doc/INSTALL> for an overview of possible configuration options. |
2832c235 TH |
136 | |
137 | =head1 OPTIONS | |
138 | ||
139 | =over 3 | |
140 | ||
f05ec4ae | 141 | =item B<-V>, B<--version> |
2832c235 | 142 | |
f05ec4ae | 143 | Print out version and copyright information and exit. |
2832c235 | 144 | |
f05ec4ae | 145 | =item B<-h>, B<--help> |
2832c235 TH |
146 | |
147 | Print this man page and exit. | |
148 | ||
f05ec4ae | 149 | =item B<-d>, B<--debug> |
2832c235 TH |
150 | |
151 | Output debugging information to STDERR while parsing STDIN. You'll | |
152 | find that information most probably in your B<INN> F<errlog> file. | |
153 | ||
f05ec4ae | 154 | =item B<-q>, B<--quiet> |
2832c235 TH |
155 | |
156 | Suppress logging to syslog. | |
157 | ||
158 | =back | |
159 | ||
160 | =head1 INSTALLATION | |
161 | ||
f05ec4ae | 162 | See L<doc/INSTALL.> |
2832c235 TH |
163 | |
164 | =head1 EXAMPLES | |
165 | ||
166 | Set up a feed like that in your B<INN> F<newsfeeds> file: | |
167 | ||
168 | ## gather statistics for NewsStats | |
169 | newsstats! | |
170 | :!*,de.* | |
171 | :Tc,WmtfbsPNH,Ac:/path/to/feedlog.pl | |
172 | ||
f05ec4ae | 173 | See L<doc/INSTALL> for further information. |
2832c235 TH |
174 | |
175 | =head1 FILES | |
176 | ||
177 | =over 4 | |
178 | ||
179 | =item F<feedlog.pl> | |
180 | ||
181 | The script itself. | |
182 | ||
183 | =item F<NewsStats.pm> | |
184 | ||
185 | Library functions for the NewsStats package. | |
186 | ||
187 | =item F<newsstats.conf> | |
188 | ||
f05ec4ae | 189 | Runtime configuration file. |
2832c235 TH |
190 | |
191 | =back | |
192 | ||
193 | =head1 BUGS | |
194 | ||
195 | Please report any bugs or feature requests to the author or use the | |
196 | bug tracker at L<http://bugs.th-h.de/>! | |
197 | ||
198 | =head1 SEE ALSO | |
199 | ||
200 | =over 2 | |
201 | ||
202 | =item - | |
203 | ||
f05ec4ae | 204 | L<doc/README> |
2832c235 TH |
205 | |
206 | =item - | |
207 | ||
f05ec4ae | 208 | L<doc/INSTALL> |
2832c235 TH |
209 | |
210 | =back | |
211 | ||
212 | This script is part of the B<NewsStats> package. | |
213 | ||
214 | =head1 AUTHOR | |
215 | ||
216 | Thomas Hochstein <thh@inter.net> | |
217 | ||
218 | =head1 COPYRIGHT AND LICENSE | |
219 | ||
f05ec4ae | 220 | Copyright (c) 2010-2012 Thomas Hochstein <thh@inter.net> |
2832c235 TH |
221 | |
222 | This program is free software; you may redistribute it and/or modify it | |
223 | under the same terms as Perl itself. | |
224 | ||
225 | =cut |