5 # This script will get statistical data on newgroup usage
8 # It is part of the NewsStats package.
10 # Copyright (c) 2010 Thomas Hochstein <thh@inter.net>
12 # It can be redistributed and/or modified under the same terms under
13 # which Perl itself is published.
16 our $VERSION = "0.01";
18 push(@INC, dirname($0));
22 use NewsStats qw(:DEFAULT :TimePeriods :Output :SQLHelper);
26 ################################# Main program #################################
28 ### read commandline options
29 my %Options = &ReadOptions('m:p:n:o:t:l:b:iscqdg:');
31 ### read configuration
32 my %Conf = %{ReadConfig('newsstats.conf')};
34 ### override configuration via commandline options
36 $ConfOverride{'DBTableGrps'} = $Options{'g'} if $Options{'g'};
37 &OverrideConfig(\%Conf,\%ConfOverride);
39 ### check for incompatible command line options
40 # you can't mix '-t', '-b' and '-l'
41 # -b/-l take preference over -t, and -b takes preference over -l
42 if ($Options{'b'} or $Options{'l'}) {
45 warn ("$MySelf: W: You cannot combine thresholds (-t) and top lists (-b) or levels (-l). Threshold '-t $Options{'t'}' was ignored.\n");
48 if ($Options{'b'} and $Options{'l'}) {
50 warn ("$MySelf: W: You cannot combine top lists (-b) and levels (-l). Level '-l $Options{'l'}' was ignored.\n");
53 # -q/-d don't work with -b or -l
54 warn ("$MySelf: W: Sorting by number of postings (-q) ignored due to top list mode (-b) / levels (-l).\n") if $Options{'q'};
55 warn ("$MySelf: W: Reverse sorting (-d) ignored due to top list mode (-b) / levels (-l).\n") if $Options{'d'};
59 # default output type to 'dump'
60 $Options{'o'} = 'dump' if !$Options{'o'};
61 # fail if more than one newsgroup is combined with 'dumpgroup' type
62 die ("$MySelf: E: You cannot combine newsgroup lists (-n) with more than one group with '-o dumpgroup'!\n") if ($Options{'o'} eq 'dumpgroup' and defined($Options{'n'}) and $Options{'n'} =~ /:|\*/);
63 # accept 'dumpgroup' only with -n
64 if ($Options{'o'} eq 'dumpgroup' and !defined($Options{'n'})) {
65 $Options{'o'} = 'dump';
66 warn ("$MySelf: W: You must submit exactly one newsgroup ('-n news.group') for '-o dumpgroup'. Output type was set to 'dump'.\n");
68 # set output type to 'pretty' for -l
70 $Options{'o'} = 'pretty';
71 warn ("$MySelf: W: Output type forced to '-o pretty' due to usage of '-l'.\n");
75 my ($StartMonth,$EndMonth) = &GetTimePeriod($Options{'m'},$Options{'p'});
76 # if time period is more than one month: set output type to '-o pretty' or '-o dumpgroup'
77 if ($Options{'o'} eq 'dump' and $Options{'p'}) {
78 if (defined($Options{'n'}) and $Options{'n'} !~ /:|\*/) {
79 warn ("$MySelf: W: You cannot combine time periods (-p) with '-o dump', changing output type to '-o dumpgroup'.\n");
80 $Options{'o'} = 'dumpgroup';
82 warn ("$MySelf: W: You cannot combine time periods (-p) with '-o dump', changing output type to '-o pretty'.\n");
83 $Options{'o'} = 'pretty';
88 my $DBHandle = InitDB(\%Conf,1);
91 # get list of newsgroups (-n)
92 my ($QueryGroupList,$QueryThreshold,@GroupList,@Params);
93 my $Newsgroups = $Options{'n'};
95 # explode list of newsgroups for WHERE clause
96 ($QueryGroupList,@GroupList) = &SQLGroupList($Newsgroups);
98 # set to dummy value (always true)
103 if (defined($Options{'t'})) {
105 # -i: list groups below threshold
106 $QueryThreshold .= ' postings < ?';
108 # default: list groups above threshold
109 $QueryThreshold .= ' postings > ?';
111 # push threshold to Params
112 push @Params,$Options{'t'};
114 # set to dummy value (always true)
118 # construct WHERE clause
119 # $QueryGroupList is "list of newsgroup" (or 1),
120 # $QueryThreshold is threshold definition (or 1),
121 # &SQLHierarchies() takes care of the exclusion of hierarchy levels (.ALL)
122 # according to setting of -s
123 my $WhereClause = sprintf('month BETWEEN ? AND ? AND %s AND %s %s',$QueryGroupList,$QueryThreshold,&SQLHierarchies($Options{'s'}));
125 # get length of longest newsgroup delivered by query for formatting purposes
127 my $MaxLength = &GetMaxLenght($DBHandle,$Conf{'DBTableGrps'},'newsgroup',$WhereClause,$StartMonth,$EndMonth,(@GroupList,@Params));
129 my ($OrderClause,$DBQuery);
130 # -b (best of / top list) defined?
131 if (!defined($Options{'b'}) and !defined($Options{'l'})) {
132 # default: neither -b nor -l
133 # set ordering (ORDER BY) to "newsgroups" or "postings", "ASC" or "DESC"
134 # according to -q and -d
135 $OrderClause = 'newsgroup';
136 $OrderClause = 'postings' if $Options{'q'};
137 $OrderClause .= ' DESC' if $Options{'d'};
138 # prepare query: get number of postings per group from groups table for given months and newsgroups
139 $DBQuery = $DBHandle->prepare(sprintf("SELECT month,newsgroup,postings FROM %s.%s WHERE %s ORDER BY month,%s",$Conf{'DBDatabase'},$Conf{'DBTableGrps'},$WhereClause,$OrderClause));
140 } elsif ($Options{'b'}) {
141 # -b is set (then -l can't be!)
142 # set sorting order (-i): top or flop list?
144 $OrderClause = 'postings';
146 $OrderClause = 'postings DESC';
148 # set -b to 10 if < 1 (Top 10)
149 $Options{'b'} = 10 if $Options{'b'} !~ /^\d*$/ or $Options{'b'} < 1;
150 # push LIMIT to Params
151 push @Params,$Options{'b'};
152 # prepare query: get sum of postings per group from groups table for given months and newsgroups with LIMIT
153 $DBQuery = $DBHandle->prepare(sprintf("SELECT newsgroup,SUM(postings) AS postings FROM %s.%s WHERE %s GROUP BY newsgroup ORDER BY %s,newsgroup LIMIT ?",$Conf{'DBDatabase'},$Conf{'DBTableGrps'},$WhereClause,$OrderClause));
155 # -l must be set now, as all other cases have been taken care of
156 # which kind of level (-i): more than -l x or less than -l x?
163 # prepare and execute query: get list of newsgroups meeting level condition
164 $DBQuery = $DBHandle->prepare(sprintf("SELECT newsgroup FROM %s.%s WHERE %s GROUP BY newsgroup HAVING MAX(postings) %s ?",$Conf{'DBDatabase'},$Conf{'DBTableGrps'},$WhereClause,$Level));
165 $DBQuery->execute($StartMonth,$EndMonth,@GroupList,$Options{'l'})
166 or die sprintf("$MySelf: E: Can't get groups data for %s to %s from %s.%s: %s\n",$StartMonth,$EndMonth,$Conf{'DBDatabase'},$Conf{'DBTableGrps'},$DBI::errstr);
167 # add newsgroups to a comma-seperated list ready for IN(...) query
169 while (my ($Newsgroup) = $DBQuery->fetchrow_array) {
170 $GroupList .= ',' if (defined($GroupList) and $GroupList ne '');
171 $GroupList .= "'$Newsgroup'";
173 $DBQuery = $DBHandle->prepare(sprintf("SELECT month,newsgroup,postings FROM %s.%s WHERE newsgroup IN (%s) AND %s ORDER BY newsgroup,month",$Conf{'DBDatabase'},$Conf{'DBTableGrps'},$GroupList,$WhereClause));
177 $DBQuery->execute($StartMonth,$EndMonth,@GroupList,@Params)
178 or die sprintf("$MySelf: E: Can't get groups data for %s to %s from %s.%s: %s\n",$StartMonth,$EndMonth,$Conf{'DBDatabase'},$Conf{'DBTableGrps'},$DBI::errstr);
181 # print caption (-c) with time period if -m or -p is set
184 printf ("----- Report from %s to %s\n",$StartMonth,$EndMonth);
185 } elsif ($Options{'m'}) {
186 printf ("----- Report for %s\n",$StartMonth);
189 # print caption (-c) with newsgroup list if -n is set
190 printf ("----- Newsgroups: %s\n",join(',',split(/:/,$Newsgroups))) if $Options{'c'} and $Options{'n'};
191 # print caption (-c) with threshold if -t is set, taking -i in account
192 printf ("----- Threshold: %s %u\n",$Options{'i'} ? '<' : '>',$Options{'t'}) if $Options{'c'} and $Options{'t'};
193 if (!defined($Options{'b'}) and !defined($Options{'l'})) {
194 # default: neither -b nor -l
195 &OutputData($Options{'o'},$DBQuery,$MaxLength);
196 } elsif ($Options{'b'}) {
197 # -b is set (then -l can't be!)
198 # we have to read in the query results ourselves, as they do not have standard layout
199 while (my ($Newsgroup,$Postings) = $DBQuery->fetchrow_array) {
200 # we just assign "top x" or "bottom x" instead of a month for the caption and force an output type of pretty
201 print &FormatOutput('pretty', ($Options{'i'} ? 'Bottom ' : 'Top ').$Options{'b'}, $Newsgroup, $Postings, $MaxLength);
204 # -l must be set now, as all other cases have been taken care of
205 # print caption (-c) with level, taking -i in account
206 printf ("----- Newsgroups with %s than %u postings over the whole time period\n",$Options{'i'} ? 'less' : 'more',$Options{'l'}) if $Options{'c'};
207 # we have to read in the query results ourselves, as they do not have standard layout
208 while (my ($Month,$Newsgroup,$Postings) = $DBQuery->fetchrow_array) {
209 # we just switch $Newsgroups and $Month for output generation
210 print &FormatOutput($Options{'o'}, $Newsgroup, $Month, $Postings, 7);
215 $DBHandle->disconnect;
219 ################################ Documentation #################################
223 groupstats - create reports on newsgroup usage
227 B<groupstats> [B<-Vhiscqd>] [B<-m> I<YYYY-MM>] [B<-p> I<YYYY-MM:YYYY-MM>] [B<-n> I<newsgroup(s)>] [B<-t> I<threshold>] [B<-l> I<level>] [B<-b> I<number>] [B<-o> I<output type>] [B<-g> I<database table>]
231 See doc/README: Perl 5.8.x itself and the following modules from CPAN:
247 This script create reports on newsgroup usage (number of postings per
248 group per month) taken from result tables created by
251 The time period to act on defaults to last month; you can assign
252 another month via the B<-m> switch or a time period via the B<-p>
253 switch; the latter takes preference.
255 B<groupstats> will process all newsgroups by default; you can limit
256 that to only some newsgroups by supplying a list of those groups via
257 B<-n> (see below). You can include hierarchy levels in the output by
258 adding the B<-s> switch (see below).
260 Furthermore you can set a threshold via B<-t> so that only newsgroups
261 with more postings per month will be included in the report. You can
262 invert that by the B<-i> switch so only newsgroups with less than
263 I<threshold> postings per month will be included.
265 You can sort the output by number of postings per month instead of the
266 default (alphabetical list of newsgroups) by using B<-q>; you can
267 reverse the sorting order (from highest to lowest or in reversed
268 alphabetical order) by using B<-d>.
270 Furthermore, you can create a list of newsgroups that had consistently
271 more (or less) than x postings per month during the whole report
272 period by using B<-l> (together with B<i> as needed).
274 Last but not least you can create a "best of" list of the top x
275 newsgroups via B<-b> (or a "worst of" list by adding B<i>).
277 By default, B<groupstats> will dump a very simple alphabetical list of
278 newsgroups, one per line, followed by the number of postings in that
279 month. This output format of course cannot sensibly be combined with
280 time periods, so you can set the output format by using B<-o> (see
281 below). Captions can be added by setting the B<-c> switch.
285 F<groupstats.pl> will read its configuration from F<newsstats.conf>
286 which should be present in the same directory via Config::Auto.
288 See doc/INSTALL for an overview of possible configuration options.
290 You can override configuration options via the B<-g> switch.
296 =item B<-V> (version)
298 Print out version and copyright information on B<yapfaq> and exit.
302 Print this man page and exit.
304 =item B<-m> I<YYYY-MM> (month)
306 Set processing period to a month in YYYY-MM format. Ignored if B<-p>
309 =item B<-p> I<YYYY-MM:YYYY-MM> (period)
311 Set processing period to a time period between two month, each in
312 YYYY-MM format, separated by a colon. Overrides B<-m>.
314 =item B<-n> I<newsgroup(s)> (newsgroups)
316 Limit processing to a certain set of newsgroups. I<newsgroup(s)> can
317 be a single newsgroup name (de.alt.test), a newsgroup hierarchy
318 (de.alt.*) or a list of either of these, separated by colons, for
321 de.test:de.alt.test:de.newusers.*
323 =item B<-t> I<threshold> (threshold)
325 Only include newsgroups with more than I<threshold> postings per
326 month. Can be inverted by the B<-i> switch so that only newsgroups
327 with less than I<threshold> postings will be included.
329 This setting will be ignored if B<-l> or B<-b> is set.
331 =item B<-l> I<level> (level)
333 Only include newsgroups with more than I<level> postings per
334 month, every month during the whole reporting period. Can be inverted
335 by the B<-i> switch so that only newsgroups with less than I<level>
336 postings every single month will be included. Output will be ordered
337 by newsgroup name, followed by month.
339 This setting will be ignored if B<-b> is set. Overrides B<-t> and
340 can't be used together with B<-q> or B<-d>.
342 =item B<-b> I<n> (best of)
344 Create a list of the I<n> newsgroups with the most postings over the
345 whole reporting period. Can be inverted by the B<-i> switch so that a
346 list of the I<n> newsgroups with the least postings over the whole
347 period is generated. Output will be ordered by sum of postings.
349 Overrides B<-t> and B<-l> and can't be used together with B<-q> or
350 B<-d>. Output format is set to I<pretty> (see below).
354 Used in conjunction with B<-t>, B<-l> or B<-b> to set a lower
355 threshold or level or generate a "bottom list" instead of a top list.
357 =item B<-s> (sum per hierarchy level)
359 Include "virtual" groups for every hierarchy level in output, for
366 See the B<gatherstats> man page for details.
368 =item B<-o> I<output type> (output format)
370 Set output format. Default is I<dump>, consisting of an alphabetical
371 list of newsgroups, each on a new line, followed by the number of
372 postings in that month. This default format can't be used with time
373 periods of more than one month.
375 I<list> format is like I<dump>, but will print the month in front of
378 I<dumpgroup> format can only be use with a group list (see B<-n>) of
379 exactly one newsgroup and is like I<dump>, but will output months,
380 followed by the number of postings.
382 If you don't need easily parsable output, you'll mostly use I<pretty>
383 format, which will print a header for each new month and try to align
384 newsgroup names and posting counts. Usage of B<-b> will force this
387 =item B<-c> (captions)
389 Add captions to output (reporting period, newsgroups list, threshold).
391 =item B<-q> (quantity of postings)
393 Sort by number of postings instead of by newsgroup names.
395 Cannot be used with B<-l> or B<-b>.
397 =item B<-d> (descending)
399 Change sort order to descending.
401 Cannot be used with B<-l> or B<-b>.
403 =item B<-g> I<table> (postings per group table)
405 Override I<DBTableGrps> from F<newsstats.conf>.
415 Show number of postings per group for lasth month in I<dump> format:
419 Show that report for January of 2010 and de.alt.* plus de.test,
420 including display of hierarchy levels:
422 groupstats -m 2010-01 -n de.alt.*:de.test -s
424 Show that report for the year of 2010 in I<pretty> format:
426 groupstats -p 2010-01:2010-12 -o pretty
428 Only show newsgroups with less than 30 postings last month, ordered
429 by number of postings, descending, in I<pretty> format:
431 groupstats -iqdt 30 -o pretty
433 Show top 10 for the first half-year of of 2010 in I<pretty> format:
435 groupstats -p 2010-01:2010-06 -b 10 -o pretty
437 Report all groups that had less than 30 postings every singele month
438 in the year of 2010 (I<pretty> format is forced)
440 groupstats -p 2010-01:2010-12 -il 30
446 =item F<groupstats.pl>
450 =item F<NewsStats.pm>
452 Library functions for the NewsStats package.
454 =item F<newsstats.conf>
456 Runtime configuration file for B<yapfaq>.
462 Please report any bugs or feature requests to the author or use the
463 bug tracker at L<http://bugs.th-h.de/>!
483 This script is part of the B<NewsStats> package.
487 Thomas Hochstein <thh@inter.net>
489 =head1 COPYRIGHT AND LICENSE
491 Copyright (c) 2010 Thomas Hochstein <thh@inter.net>
493 This program is free software; you may redistribute it and/or modify it
494 under the same terms as Perl itself.