Be more fault-tolerant when reading checkgroups.
[usenet/newsstats.git] / gatherstats.pl
... / ...
CommitLineData
1#! /usr/bin/perl -W
2#
3# gatherstats.pl
4#
5# This script will gather statistical information from a database
6# containing headers and other information from a INN feed.
7#
8# It is part of the NewsStats package.
9#
10# Copyright (c) 2010-2012 Thomas Hochstein <thh@inter.net>
11#
12# It can be redistributed and/or modified under the same terms under
13# which Perl itself is published.
14
15BEGIN {
16 our $VERSION = "0.01";
17 use File::Basename;
18 push(@INC, dirname($0));
19}
20use strict;
21
22use NewsStats qw(:DEFAULT :TimePeriods ListNewsgroups ReadGroupList);
23
24use DBI;
25use Getopt::Long qw(GetOptions);
26Getopt::Long::config ('bundling');
27
28################################# Definitions ##################################
29
30# define types of information that can be gathered
31# all / groups (/ clients / hosts)
32my %LegalStats;
33@LegalStats{('all','groups')} = ();
34
35################################# Main program #################################
36
37### read commandline options
38my ($OptCheckgroupsFile,$OptClientsDB,$OptDebug,$OptGroupsDB,$OptTLH,
39 $OptHostsDB,$OptMonth,$OptRawDB,$OptStatsType,$OptTest);
40GetOptions ('c|checkgroups=s' => \$OptCheckgroupsFile,
41 'clientsdb=s' => \$OptClientsDB,
42 'd|debug!' => \$OptDebug,
43 'groupsdb=s' => \$OptGroupsDB,
44 'hierarchy=s' => \$OptTLH,
45 'hostsdb=s' => \$OptHostsDB,
46 'm|month=s' => \$OptMonth,
47 'rawdb=s' => \$OptRawDB,
48 's|stats=s' => \$OptStatsType,
49 't|test!' => \$OptTest,
50 'h|help' => \&ShowPOD,
51 'V|version' => \&ShowVersion) or exit 1;
52
53### read configuration
54my %Conf = %{ReadConfig($HomePath.'/newsstats.conf')};
55
56### override configuration via commandline options
57my %ConfOverride;
58$ConfOverride{'DBTableRaw'} = $OptRawDB if $OptRawDB;
59$ConfOverride{'DBTableGrps'} = $OptGroupsDB if $OptGroupsDB;
60$ConfOverride{'DBTableClnts'} = $OptClientsDB if $OptClientsDB;
61$ConfOverride{'DBTableHosts'} = $OptHostsDB if $OptHostsDB;
62$ConfOverride{'TLH'} = $OptTLH if $OptTLH;
63&OverrideConfig(\%Conf,\%ConfOverride);
64
65### get type of information to gather, defaulting to 'all'
66$OptStatsType = 'all' if !$OptStatsType;
67&Bleat(2, sprintf("Unknown type '%s'!", $OptStatsType))
68 if !exists($LegalStats{$OptStatsType});
69
70### get time period from --month
71# get verbal description of time period, drop SQL code
72my ($Period) = &GetTimePeriod($OptMonth);
73&Bleat(2,"--month option has an invalid format - please use 'YYYY-MM' or ".
74 "'YYYY-MM:YYYY-MM'!") if (!$Period or $Period eq 'all time');
75
76### reformat $Conf{'TLH'}
77my $TLH;
78if ($Conf{'TLH'}) {
79 # $Conf{'TLH'} is parsed as an array by Config::Auto;
80 # make a flat list again, separated by :
81 if (ref($Conf{'TLH'}) eq 'ARRAY') {
82 $TLH = join(':',@{$Conf{'TLH'}});
83 } else {
84 $TLH = $Conf{'TLH'};
85 }
86 # strip whitespace
87 $TLH =~ s/\s//g;
88 # add trailing dots if none are present yet
89 # (using negative look-behind assertions)
90 $TLH =~ s/(?<!\.):/.:/g;
91 $TLH =~ s/(?<!\.)$/./;
92 # check for illegal characters
93 &Bleat(2,'Config error - illegal characters in TLH definition!')
94 if ($TLH !~ /^[a-zA-Z0-9:+.-]+$/);
95 # escape dots
96 $TLH =~ s/\./\\./g;
97 if ($TLH =~ /:/) {
98 # reformat $TLH from a:b to (a)|(b),
99 # e.g. replace ':' by ')|('
100 $TLH =~ s/:/)|(/g;
101 $TLH = '(' . $TLH . ')';
102 };
103};
104
105# read list of newsgroups from --checkgroups
106# into a hash
107my %ValidGroups = %{ReadGroupList($OptCheckgroupsFile)} if $OptCheckgroupsFile;
108
109### init database
110my $DBHandle = InitDB(\%Conf,1);
111
112### get data for each month
113&Bleat(1,'Test mode. Database is not updated.') if $OptTest;
114foreach my $Month (&ListMonth($Period)) {
115
116 print "---------- $Month ----------\n" if $OptDebug;
117
118 if ($OptStatsType eq 'all' or $OptStatsType eq 'groups') {
119 ### ----------------------------------------------
120 ### get groups data (number of postings per group)
121 # get groups data from raw table for given month
122 my $DBQuery = $DBHandle->prepare(sprintf("SELECT newsgroups FROM %s.%s ".
123 "WHERE day LIKE ? AND NOT disregard",
124 $Conf{'DBDatabase'},
125 $Conf{'DBTableRaw'}));
126 $DBQuery->execute($Month.'-%')
127 or &Bleat(2,sprintf("Can't get groups data for %s from %s.%s: ".
128 "$DBI::errstr\n",$Month,
129 $Conf{'DBDatabase'},$Conf{'DBTableRaw'}));
130
131 # count postings per group
132 my %Postings;
133 while (($_) = $DBQuery->fetchrow_array) {
134 # get list oft newsgroups and hierarchies from Newsgroups:
135 my %Newsgroups = ListNewsgroups($_,$TLH,
136 $OptCheckgroupsFile ? \%ValidGroups : '');
137 # count each newsgroup and hierarchy once
138 foreach (sort keys %Newsgroups) {
139 $Postings{$_}++;
140 };
141 };
142
143 # add valid but empty groups if --checkgroups is set
144 if (%ValidGroups) {
145 foreach (sort keys %ValidGroups) {
146 if (!defined($Postings{$_})) {
147 $Postings{$_} = 0 ;
148 warn (sprintf("ADDED: %s as empty group\n",$_));
149 }
150 };
151 };
152
153 # delete old data for that month
154 if (!$OptTest) {
155 $DBQuery = $DBHandle->do(sprintf("DELETE FROM %s.%s WHERE month = ?",
156 $Conf{'DBDatabase'},$Conf{'DBTableGrps'}),
157 undef,$Month)
158 or &Bleat(2,sprintf("Can't delete old groups data for %s from %s.%s: ".
159 "$DBI::errstr\n",$Month,
160 $Conf{'DBDatabase'},$Conf{'DBTableGrps'}));
161 };
162
163 print "----- GroupStats -----\n" if $OptDebug;
164 foreach my $Newsgroup (sort keys %Postings) {
165 print "$Newsgroup => $Postings{$Newsgroup}\n" if $OptDebug;
166 if (!$OptTest) {
167 # write to database
168 $DBQuery = $DBHandle->prepare(sprintf("INSERT INTO %s.%s ".
169 "(month,newsgroup,postings) ".
170 "VALUES (?, ?, ?)",
171 $Conf{'DBDatabase'},
172 $Conf{'DBTableGrps'}));
173 $DBQuery->execute($Month, $Newsgroup, $Postings{$Newsgroup})
174 or &Bleat(2,sprintf("Can't write groups data for %s/%s to %s.%s: ".
175 "$DBI::errstr\n",$Month,$Newsgroup,
176 $Conf{'DBDatabase'},$Conf{'DBTableGrps'}));
177 $DBQuery->finish;
178 };
179 };
180 } else {
181 # other types of information go here - later on
182 };
183};
184
185### close handles
186$DBHandle->disconnect;
187
188__END__
189
190################################ Documentation #################################
191
192=head1 NAME
193
194gatherstats - process statistical data from a raw source
195
196=head1 SYNOPSIS
197
198B<gatherstats> [B<-Vhdt>] [B<-m> I<YYYY-MM> | I<YYYY-MM:YYYY-MM>] [B<-s> I<stats] [B<-c> I<checkgroups file>]] [B<--hierarchy> I<TLH>] [B<--rawdb> I<database table>] [B<-groupsdb> I<database table>] [B<--clientsdb> I<database table>] [B<--hostsdb> I<database table>]
199
200=head1 REQUIREMENTS
201
202See L<doc/README>.
203
204=head1 DESCRIPTION
205
206This script will extract and process statistical information from a
207database table which is fed from F<feedlog.pl> for a given time period
208and write its results to (an)other database table(s). Entries marked
209with I<'disregard'> in the database will be ignored; currently, you
210have to set this flag yourself, using your database management tools.
211You can exclude erroneous entries that way (e.g. automatic reposts
212(think of cancels flood and resurrectors); spam; ...).
213
214The time period to act on defaults to last month; you can assign
215another time period or a single month via the B<--month> option (see
216below).
217
218By default B<gatherstats> will process all types of information; you
219can change that using the B<--stats> option and assigning the type of
220information to process. Currently that doesn't matter yet as only
221processing of the number of postings per group per month is
222implemented anyway.
223
224Possible information types include:
225
226=over 3
227
228=item B<groups> (postings per group per month)
229
230B<gatherstats> will examine Newsgroups: headers. Crosspostings will be
231counted for each single group they appear in. Groups not in I<TLH>
232will be ignored.
233
234B<gatherstats> will also add up the number of postings for each
235hierarchy level, but only count each posting once. A posting to
236de.alt.test will be counted for de.alt.test, de.alt.ALL and de.ALL,
237respectively. A crossposting to de.alt.test and de.alt.admin, on the
238other hand, will be counted for de.alt.test and de.alt.admin each, but
239only once for de.alt.ALL and de.ALL.
240
241Data is written to I<DBTableGrps> (see L<doc/INSTALL>); you can
242override that default through the B<--groupsdb> option.
243
244=back
245
246=head2 Configuration
247
248B<gatherstats> will read its configuration from F<newsstats.conf>
249which should be present in the same directory via Config::Auto.
250
251See L<doc/INSTALL> for an overview of possible configuration options.
252
253You can override configuration options via the B<--hierarchy>,
254B<--rawdb>, B<--groupsdb>, B<--clientsdb> and B<--hostsdb> options,
255respectively.
256
257=head1 OPTIONS
258
259=over 3
260
261=item B<-V>, B<--version>
262
263Print out version and copyright information and exit.
264
265=item B<-h>, B<--help>
266
267Print this man page and exit.
268
269=item B<-d>, B<--debug>
270
271Output debugging information to STDOUT while processing (number of
272postings per group).
273
274=item B<-t>, B<--test>
275
276Do not write results to database. You should use B<--debug> in
277conjunction with B<--test> ... everything else seems a bit pointless.
278
279=item B<-m>, B<--month> I<YYYY-MM[:YYYY-MM]>
280
281Set processing period to a single month in YYYY-MM format or to a time
282period between two month in YYYY-MM:YYYY-MM format (two month, separated
283by a colon).
284
285
286=item B<-s>, B<--stats> I<type>
287
288Set processing type to one of I<all> and I<groups>. Defaults to all
289(and is currently rather pointless as only I<groups> has been
290implemented).
291
292=item B<-c>, B<--checkgroups> I<filename>
293
294Check each group against a list of valid newsgroups read from
295I<filename>, one group on each line and ignoring everything after the
296first whitespace (so you can use a file in checkgroups format or (part
297of) your INN active file).
298
299Newsgroups not found in I<filename> will be dropped (and logged to
300STDERR), and newsgroups found in I<filename> but having no postings
301will be added with a count of 0 (and logged to STDERR).
302
303=item B<--hierarchy> I<TLH> (newsgroup hierarchy)
304
305Override I<TLH> from F<newsstats.conf>.
306
307=item B<--rawdb> I<table> (raw data table)
308
309Override I<DBTableRaw> from F<newsstats.conf>.
310
311=item B<--groupsdb> I<table> (postings per group table)
312
313Override I<DBTableGrps> from F<newsstats.conf>.
314
315=item B<--clientsdb> I<table> (client data table)
316
317Override I<DBTableClnts> from F<newsstats.conf>.
318
319=item B<--hostsdb> I<table> (host data table)
320
321Override I<DBTableHosts> from F<newsstats.conf>.
322
323=back
324
325=head1 INSTALLATION
326
327See L<doc/INSTALL>.
328
329=head1 EXAMPLES
330
331Process all types of information for lasth month:
332
333 gatherstats
334
335Do a dry run, showing results of processing:
336
337 gatherstats --debug --test
338
339Process all types of information for January of 2010:
340
341 gatherstats --month 2010-01
342
343Process only number of postings for the year of 2010,
344checking against checkgroups-2010.txt:
345
346 gatherstats -m 2010-01:2010-12 -s groups -c checkgroups-2010.txt
347
348=head1 FILES
349
350=over 4
351
352=item F<gatherstats.pl>
353
354The script itself.
355
356=item F<NewsStats.pm>
357
358Library functions for the NewsStats package.
359
360=item F<newsstats.conf>
361
362Runtime configuration file.
363
364=back
365
366=head1 BUGS
367
368Please report any bugs or feature requests to the author or use the
369bug tracker at L<http://bugs.th-h.de/>!
370
371=head1 SEE ALSO
372
373=over 2
374
375=item -
376
377L<doc/README>
378
379=item -
380
381L<doc/INSTALL>
382
383=back
384
385This script is part of the B<NewsStats> package.
386
387=head1 AUTHOR
388
389Thomas Hochstein <thh@inter.net>
390
391=head1 COPYRIGHT AND LICENSE
392
393Copyright (c) 2010-2012 Thomas Hochstein <thh@inter.net>
394
395This program is free software; you may redistribute it and/or modify it
396under the same terms as Perl itself.
397
398=cut
This page took 0.01037 seconds and 4 git commands to generate.