Commit | Line | Data |
---|---|---|
2832c235 TH |
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 | # | |
880c3eb2 | 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($0)); | |
19 | } | |
20 | use strict; | |
21 | ||
ad609792 | 22 | use NewsStats qw(:DEFAULT :TimePeriods ListNewsgroups ReadGroupList); |
2832c235 TH |
23 | |
24 | use DBI; | |
880c3eb2 TH |
25 | use Getopt::Long qw(GetOptions); |
26 | Getopt::Long::config ('bundling'); | |
2832c235 TH |
27 | |
28 | ################################# Definitions ################################## | |
29 | ||
30 | # define types of information that can be gathered | |
31 | # all / groups (/ clients / hosts) | |
880c3eb2 TH |
32 | my %LegalStats; |
33 | @LegalStats{('all','groups')} = (); | |
2832c235 TH |
34 | |
35 | ################################# Main program ################################# | |
36 | ||
37 | ### read commandline options | |
880c3eb2 TH |
38 | my ($OptCheckgroupsFile,$OptClientsDB,$OptDebug,$OptGroupsDB,$OptTLH, |
39 | $OptHostsDB,$OptMonth,$OptRawDB,$OptStatsType,$OptTest); | |
40 | GetOptions ('c|checkgroups=s' => \$OptCheckgroupsFile, | |
41 | 'clientsdb=s' => \$OptClientsDB, | |
42 | 'd|debug!' => \$OptDebug, | |
43 | 'groupsdb=s' => \$OptGroupsDB, | |
44 | 'hierarchy=s' => \$OptTLH, | |
45 | 'hostsdb=s' => \$OptHostsDB, | |
46 | 'm|month=s' => \$OptMonth, | |
47 | 'rawdb=s' => \$OptRawDB, | |
48 | 's|stats=s' => \$OptStatsType, | |
49 | 't|test!' => \$OptTest, | |
50 | 'h|help' => \&ShowPOD, | |
51 | 'V|version' => \&ShowVersion) or exit 1; | |
2832c235 TH |
52 | |
53 | ### read configuration | |
880c3eb2 | 54 | my %Conf = %{ReadConfig($HomePath.'/newsstats.conf')}; |
2832c235 TH |
55 | |
56 | ### override configuration via commandline options | |
57 | my %ConfOverride; | |
880c3eb2 TH |
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; | |
2832c235 TH |
63 | &OverrideConfig(\%Conf,\%ConfOverride); |
64 | ||
65 | ### get type of information to gather, defaulting to 'all' | |
880c3eb2 TH |
66 | $OptStatsType = 'all' if !$OptStatsType; |
67 | &Bleat(2, sprintf("Unknown type '%s'!", $OptStatsType)) | |
68 | if !exists($LegalStats{$OptStatsType}); | |
2832c235 | 69 | |
880c3eb2 TH |
70 | ### get time period from --month |
71 | # get verbal description of time period, drop SQL code | |
72 | my ($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'); | |
2832c235 | 75 | |
17ffbeba TH |
76 | ### reformat $Conf{'TLH'} |
77 | my $TLH; | |
78 | if ($Conf{'TLH'}) { | |
79 | # $Conf{'TLH'} is parsed as an array by Config::Auto; | |
80 | # make a flat list again, separated by : | |
43a0fc77 | 81 | if (ref($Conf{'TLH'}) eq 'ARRAY') { |
17ffbeba TH |
82 | $TLH = join(':',@{$Conf{'TLH'}}); |
83 | } else { | |
84 | $TLH = $Conf{'TLH'}; | |
85 | } | |
86 | # strip whitespace | |
87 | $TLH =~ s/\s//g; | |
7773fb6d TH |
88 | # add trailing dots if none are present yet |
89 | # (using negative look-behind assertions) | |
90 | $TLH =~ s/(?<!\.):/.:/g; | |
91 | $TLH =~ s/(?<!\.)$/./; | |
17ffbeba | 92 | # check for illegal characters |
880c3eb2 | 93 | &Bleat(2,'Config error - illegal characters in TLH definition!') |
314e31aa | 94 | if ($TLH !~ /^[a-zA-Z0-9:+.-]+$/); |
7773fb6d TH |
95 | # escape dots |
96 | $TLH =~ s/\./\\./g; | |
17ffbeba | 97 | if ($TLH =~ /:/) { |
880c3eb2 | 98 | # reformat $TLH from a:b to (a)|(b), |
43a0fc77 | 99 | # e.g. replace ':' by ')|(' |
17ffbeba TH |
100 | $TLH =~ s/:/)|(/g; |
101 | $TLH = '(' . $TLH . ')'; | |
102 | }; | |
103 | }; | |
104 | ||
2832c235 TH |
105 | ### init database |
106 | my $DBHandle = InitDB(\%Conf,1); | |
107 | ||
108 | ### get data for each month | |
880c3eb2 TH |
109 | &Bleat(1,'Test mode. Database is not updated.') if $OptTest; |
110 | foreach my $Month (&ListMonth($Period)) { | |
2832c235 | 111 | |
880c3eb2 | 112 | print "---------- $Month ----------\n" if $OptDebug; |
2832c235 | 113 | |
880c3eb2 | 114 | if ($OptStatsType eq 'all' or $OptStatsType eq 'groups') { |
93c8eae2 TH |
115 | # read list of newsgroups from --checkgroups |
116 | # into a hash | |
117 | my %ValidGroups = %{ReadGroupList(sprintf('%s-%s',$OptCheckgroupsFile,$Month))} | |
118 | if $OptCheckgroupsFile; | |
119 | ||
2832c235 TH |
120 | ### ---------------------------------------------- |
121 | ### get groups data (number of postings per group) | |
122 | # get groups data from raw table for given month | |
880c3eb2 TH |
123 | my $DBQuery = $DBHandle->prepare(sprintf("SELECT newsgroups FROM %s.%s ". |
124 | "WHERE day LIKE ? AND NOT disregard", | |
125 | $Conf{'DBDatabase'}, | |
126 | $Conf{'DBTableRaw'})); | |
127 | $DBQuery->execute($Month.'-%') | |
128 | or &Bleat(2,sprintf("Can't get groups data for %s from %s.%s: ". | |
129 | "$DBI::errstr\n",$Month, | |
130 | $Conf{'DBDatabase'},$Conf{'DBTableRaw'})); | |
2832c235 TH |
131 | |
132 | # count postings per group | |
133 | my %Postings; | |
2832c235 TH |
134 | while (($_) = $DBQuery->fetchrow_array) { |
135 | # get list oft newsgroups and hierarchies from Newsgroups: | |
880c3eb2 TH |
136 | my %Newsgroups = ListNewsgroups($_,$TLH, |
137 | $OptCheckgroupsFile ? \%ValidGroups : ''); | |
2832c235 TH |
138 | # count each newsgroup and hierarchy once |
139 | foreach (sort keys %Newsgroups) { | |
2832c235 TH |
140 | $Postings{$_}++; |
141 | }; | |
142 | }; | |
143 | ||
880c3eb2 | 144 | # add valid but empty groups if --checkgroups is set |
ad609792 TH |
145 | if (%ValidGroups) { |
146 | foreach (sort keys %ValidGroups) { | |
147 | if (!defined($Postings{$_})) { | |
148 | $Postings{$_} = 0 ; | |
149 | warn (sprintf("ADDED: %s as empty group\n",$_)); | |
150 | } | |
151 | }; | |
152 | }; | |
153 | ||
71f0178b | 154 | # delete old data for that month |
880c3eb2 TH |
155 | if (!$OptTest) { |
156 | $DBQuery = $DBHandle->do(sprintf("DELETE FROM %s.%s WHERE month = ?", | |
157 | $Conf{'DBDatabase'},$Conf{'DBTableGrps'}), | |
158 | undef,$Month) | |
159 | or &Bleat(2,sprintf("Can't delete old groups data for %s from %s.%s: ". | |
160 | "$DBI::errstr\n",$Month, | |
161 | $Conf{'DBDatabase'},$Conf{'DBTableGrps'})); | |
71f0178b TH |
162 | }; |
163 | ||
880c3eb2 | 164 | print "----- GroupStats -----\n" if $OptDebug; |
2832c235 | 165 | foreach my $Newsgroup (sort keys %Postings) { |
880c3eb2 TH |
166 | print "$Newsgroup => $Postings{$Newsgroup}\n" if $OptDebug; |
167 | if (!$OptTest) { | |
2832c235 | 168 | # write to database |
880c3eb2 TH |
169 | $DBQuery = $DBHandle->prepare(sprintf("INSERT INTO %s.%s ". |
170 | "(month,newsgroup,postings) ". | |
171 | "VALUES (?, ?, ?)", | |
172 | $Conf{'DBDatabase'}, | |
173 | $Conf{'DBTableGrps'})); | |
174 | $DBQuery->execute($Month, $Newsgroup, $Postings{$Newsgroup}) | |
175 | or &Bleat(2,sprintf("Can't write groups data for %s/%s to %s.%s: ". | |
176 | "$DBI::errstr\n",$Month,$Newsgroup, | |
177 | $Conf{'DBDatabase'},$Conf{'DBTableGrps'})); | |
2832c235 TH |
178 | $DBQuery->finish; |
179 | }; | |
180 | }; | |
181 | } else { | |
182 | # other types of information go here - later on | |
183 | }; | |
184 | }; | |
185 | ||
186 | ### close handles | |
187 | $DBHandle->disconnect; | |
188 | ||
189 | __END__ | |
190 | ||
191 | ################################ Documentation ################################# | |
192 | ||
193 | =head1 NAME | |
194 | ||
195 | gatherstats - process statistical data from a raw source | |
196 | ||
197 | =head1 SYNOPSIS | |
198 | ||
93c8eae2 | 199 | B<gatherstats> [B<-Vhdt>] [B<-m> I<YYYY-MM> | I<YYYY-MM:YYYY-MM>] [B<-s> I<stats] [B<-c> I<filename template>]] [B<--hierarchy> I<TLH>] [B<--rawdb> I<database table>] [B<-groupsdb> I<database table>] [B<--clientsdb> I<database table>] [B<--hostsdb> I<database table>] |
2832c235 TH |
200 | |
201 | =head1 REQUIREMENTS | |
202 | ||
880c3eb2 | 203 | See L<doc/README>. |
2832c235 TH |
204 | |
205 | =head1 DESCRIPTION | |
206 | ||
207 | This script will extract and process statistical information from a | |
208 | database table which is fed from F<feedlog.pl> for a given time period | |
313610f6 | 209 | and write its results to (an)other database table(s). Entries marked |
880c3eb2 TH |
210 | with I<'disregard'> in the database will be ignored; currently, you |
211 | have to set this flag yourself, using your database management tools. | |
212 | You can exclude erroneous entries that way (e.g. automatic reposts | |
213 | (think of cancels flood and resurrectors); spam; ...). | |
2832c235 TH |
214 | |
215 | The time period to act on defaults to last month; you can assign | |
880c3eb2 TH |
216 | another time period or a single month via the B<--month> option (see |
217 | below). | |
2832c235 TH |
218 | |
219 | By default B<gatherstats> will process all types of information; you | |
880c3eb2 TH |
220 | can change that using the B<--stats> option and assigning the type of |
221 | information to process. Currently that doesn't matter yet as only | |
222 | processing of the number of postings per group per month is | |
223 | implemented anyway. | |
2832c235 TH |
224 | |
225 | Possible information types include: | |
226 | ||
227 | =over 3 | |
228 | ||
229 | =item B<groups> (postings per group per month) | |
230 | ||
231 | B<gatherstats> will examine Newsgroups: headers. Crosspostings will be | |
232 | counted for each single group they appear in. Groups not in I<TLH> | |
233 | will be ignored. | |
234 | ||
235 | B<gatherstats> will also add up the number of postings for each | |
236 | hierarchy level, but only count each posting once. A posting to | |
237 | de.alt.test will be counted for de.alt.test, de.alt.ALL and de.ALL, | |
238 | respectively. A crossposting to de.alt.test and de.alt.admin, on the | |
239 | other hand, will be counted for de.alt.test and de.alt.admin each, but | |
240 | only once for de.alt.ALL and de.ALL. | |
241 | ||
880c3eb2 TH |
242 | Data is written to I<DBTableGrps> (see L<doc/INSTALL>); you can |
243 | override that default through the B<--groupsdb> option. | |
2832c235 TH |
244 | |
245 | =back | |
246 | ||
247 | =head2 Configuration | |
248 | ||
880c3eb2 | 249 | B<gatherstats> will read its configuration from F<newsstats.conf> |
2832c235 TH |
250 | which should be present in the same directory via Config::Auto. |
251 | ||
880c3eb2 | 252 | See L<doc/INSTALL> for an overview of possible configuration options. |
2832c235 | 253 | |
880c3eb2 TH |
254 | You can override configuration options via the B<--hierarchy>, |
255 | B<--rawdb>, B<--groupsdb>, B<--clientsdb> and B<--hostsdb> options, | |
256 | respectively. | |
2832c235 TH |
257 | |
258 | =head1 OPTIONS | |
259 | ||
260 | =over 3 | |
261 | ||
880c3eb2 | 262 | =item B<-V>, B<--version> |
2832c235 | 263 | |
880c3eb2 | 264 | Print out version and copyright information and exit. |
2832c235 | 265 | |
880c3eb2 | 266 | =item B<-h>, B<--help> |
2832c235 TH |
267 | |
268 | Print this man page and exit. | |
269 | ||
880c3eb2 | 270 | =item B<-d>, B<--debug> |
2832c235 TH |
271 | |
272 | Output debugging information to STDOUT while processing (number of | |
273 | postings per group). | |
274 | ||
880c3eb2 | 275 | =item B<-t>, B<--test> |
2832c235 | 276 | |
880c3eb2 TH |
277 | Do not write results to database. You should use B<--debug> in |
278 | conjunction with B<--test> ... everything else seems a bit pointless. | |
2832c235 | 279 | |
880c3eb2 | 280 | =item B<-m>, B<--month> I<YYYY-MM[:YYYY-MM]> |
2832c235 | 281 | |
880c3eb2 TH |
282 | Set processing period to a single month in YYYY-MM format or to a time |
283 | period between two month in YYYY-MM:YYYY-MM format (two month, separated | |
284 | by a colon). | |
2832c235 | 285 | |
2832c235 | 286 | |
880c3eb2 | 287 | =item B<-s>, B<--stats> I<type> |
2832c235 TH |
288 | |
289 | Set processing type to one of I<all> and I<groups>. Defaults to all | |
290 | (and is currently rather pointless as only I<groups> has been | |
291 | implemented). | |
292 | ||
93c8eae2 TH |
293 | =item B<-c>, B<--checkgroups> I<filename template> |
294 | ||
295 | Check each group against a list of valid newsgroups read from a file, | |
296 | one group on each line and ignoring everything after the first | |
297 | whitespace (so you can use a file in checkgroups format or (part of) | |
298 | your INN active file). | |
299 | ||
300 | The filename is taken from I<filename template>, amended by each B<-- | |
301 | month> B<gatherstats> is processing, so that | |
302 | ||
303 | gatherstats -m 2010-01:2010-12 -c checkgroups | |
ad609792 | 304 | |
93c8eae2 TH |
305 | will check against F<checkgroups-2010-01> for January 2010, against |
306 | F<checkgroups-2010-02> for February 2010 and so on. | |
ad609792 | 307 | |
93c8eae2 TH |
308 | Newsgroups not found in the checkgroups file will be dropped (and |
309 | logged to STDERR), and newsgroups found there but having no postings | |
ad609792 TH |
310 | will be added with a count of 0 (and logged to STDERR). |
311 | ||
880c3eb2 | 312 | =item B<--hierarchy> I<TLH> (newsgroup hierarchy) |
2832c235 TH |
313 | |
314 | Override I<TLH> from F<newsstats.conf>. | |
315 | ||
880c3eb2 | 316 | =item B<--rawdb> I<table> (raw data table) |
2832c235 TH |
317 | |
318 | Override I<DBTableRaw> from F<newsstats.conf>. | |
319 | ||
880c3eb2 | 320 | =item B<--groupsdb> I<table> (postings per group table) |
2832c235 TH |
321 | |
322 | Override I<DBTableGrps> from F<newsstats.conf>. | |
323 | ||
880c3eb2 | 324 | =item B<--clientsdb> I<table> (client data table) |
2832c235 TH |
325 | |
326 | Override I<DBTableClnts> from F<newsstats.conf>. | |
327 | ||
880c3eb2 | 328 | =item B<--hostsdb> I<table> (host data table) |
2832c235 TH |
329 | |
330 | Override I<DBTableHosts> from F<newsstats.conf>. | |
331 | ||
332 | =back | |
333 | ||
334 | =head1 INSTALLATION | |
335 | ||
880c3eb2 | 336 | See L<doc/INSTALL>. |
2832c235 TH |
337 | |
338 | =head1 EXAMPLES | |
339 | ||
340 | Process all types of information for lasth month: | |
341 | ||
342 | gatherstats | |
343 | ||
344 | Do a dry run, showing results of processing: | |
345 | ||
880c3eb2 | 346 | gatherstats --debug --test |
2832c235 TH |
347 | |
348 | Process all types of information for January of 2010: | |
349 | ||
880c3eb2 | 350 | gatherstats --month 2010-01 |
2832c235 | 351 | |
ad609792 | 352 | Process only number of postings for the year of 2010, |
93c8eae2 | 353 | checking against checkgroups-*: |
2832c235 | 354 | |
93c8eae2 | 355 | gatherstats -m 2010-01:2010-12 -s groups -c checkgroups |
2832c235 TH |
356 | |
357 | =head1 FILES | |
358 | ||
359 | =over 4 | |
360 | ||
361 | =item F<gatherstats.pl> | |
362 | ||
363 | The script itself. | |
364 | ||
365 | =item F<NewsStats.pm> | |
366 | ||
367 | Library functions for the NewsStats package. | |
368 | ||
369 | =item F<newsstats.conf> | |
370 | ||
880c3eb2 | 371 | Runtime configuration file. |
2832c235 TH |
372 | |
373 | =back | |
374 | ||
375 | =head1 BUGS | |
376 | ||
377 | Please report any bugs or feature requests to the author or use the | |
378 | bug tracker at L<http://bugs.th-h.de/>! | |
379 | ||
380 | =head1 SEE ALSO | |
381 | ||
382 | =over 2 | |
383 | ||
384 | =item - | |
385 | ||
880c3eb2 | 386 | L<doc/README> |
2832c235 TH |
387 | |
388 | =item - | |
389 | ||
880c3eb2 | 390 | L<doc/INSTALL> |
2832c235 TH |
391 | |
392 | =back | |
393 | ||
394 | This script is part of the B<NewsStats> package. | |
395 | ||
396 | =head1 AUTHOR | |
397 | ||
398 | Thomas Hochstein <thh@inter.net> | |
399 | ||
400 | =head1 COPYRIGHT AND LICENSE | |
401 | ||
880c3eb2 | 402 | Copyright (c) 2010-2012 Thomas Hochstein <thh@inter.net> |
2832c235 TH |
403 | |
404 | This program is free software; you may redistribute it and/or modify it | |
405 | under the same terms as Perl itself. | |
406 | ||
407 | =cut |