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 # reset to one month for 'dump' output type
77 if ($Options{'o'} eq 'dump' and $Options{'p'}) {
78 warn ("$MySelf: W: You cannot combine time periods (-p) with '-o dump', changing output type to '-o pretty'.\n");
79 $Options{'o'} = 'pretty';
83 my $DBHandle = InitDB(\%Conf,1);
86 # get list of newsgroups (-n)
87 my ($QueryGroupList,$QueryThreshold,@GroupList,@Params);
88 my $Newsgroups = $Options{'n'};
90 # explode list of newsgroups for WHERE clause
91 ($QueryGroupList,@GroupList) = &SQLGroupList($Newsgroups);
93 # set to dummy value (always true)
98 if (defined($Options{'t'})) {
100 # -i: list groups below threshold
101 $QueryThreshold .= ' postings < ?';
103 # default: list groups above threshold
104 $QueryThreshold .= ' postings > ?';
106 # push threshold to Params
107 push @Params,$Options{'t'};
109 # set to dummy value (always true)
113 # construct WHERE clause
114 # $QueryGroupList is "list of newsgroup" (or 1),
115 # $QueryThreshold is threshold definition (or 1),
116 # &SQLHierarchies() takes care of the exclusion of hierarchy levels (.ALL)
117 # according to setting of -s
118 my $WhereClause = sprintf('month BETWEEN ? AND ? AND %s AND %s %s',$QueryGroupList,$QueryThreshold,&SQLHierarchies($Options{'s'}));
120 # get length of longest newsgroup delivered by query for formatting purposes
122 my $MaxLength = &GetMaxLenght($DBHandle,$Conf{'DBTableGrps'},'newsgroup',$WhereClause,$StartMonth,$EndMonth,(@GroupList,@Params));
124 my ($OrderClause,$DBQuery);
125 # -b (best of / top list) defined?
126 if (!defined($Options{'b'}) and !defined($Options{'l'})) {
127 # default: neither -b nor -l
128 # set ordering (ORDER BY) to "newsgroups" or "postings", "ASC" or "DESC"
129 # according to -q and -d
130 $OrderClause = 'newsgroup';
131 $OrderClause = 'postings' if $Options{'q'};
132 $OrderClause .= ' DESC' if $Options{'d'};
133 # prepare query: get number of postings per group from groups table for given months and newsgroups
134 $DBQuery = $DBHandle->prepare(sprintf("SELECT month,newsgroup,postings FROM %s.%s WHERE %s ORDER BY month,%s",$Conf{'DBDatabase'},$Conf{'DBTableGrps'},$WhereClause,$OrderClause));
135 } elsif ($Options{'b'}) {
136 # -b is set (then -l can't be!)
137 # set sorting order (-i): top or flop list?
139 $OrderClause = 'postings';
141 $OrderClause = 'postings DESC';
143 # set -b to 10 if < 1 (Top 10)
144 $Options{'b'} = 10 if $Options{'b'} !~ /^\d*$/ or $Options{'b'} < 1;
145 # push LIMIT to Params
146 push @Params,$Options{'b'};
147 # prepare query: get sum of postings per group from groups table for given months and newsgroups with LIMIT
148 $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));
150 # -l must be set now, as all other cases have been taken care of
151 # which kind of level (-i): more than -l x or less than -l x?
158 # prepare and execute query: get list of newsgroups meeting level condition
159 $DBQuery = $DBHandle->prepare(sprintf("SELECT newsgroup FROM %s.%s WHERE %s GROUP BY newsgroup HAVING MAX(postings) %s ?",$Conf{'DBDatabase'},$Conf{'DBTableGrps'},$WhereClause,$Level));
160 $DBQuery->execute($StartMonth,$EndMonth,@GroupList,$Options{'l'})
161 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);
162 # add newsgroups to a comma-seperated list ready for IN(...) query
164 while (my ($Newsgroup) = $DBQuery->fetchrow_array) {
165 $GroupList .= ',' if (defined($GroupList) and $GroupList ne '');
166 $GroupList .= "'$Newsgroup'";
168 $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));
172 $DBQuery->execute($StartMonth,$EndMonth,@GroupList,@Params)
173 or die sprintf("$MySelf: E: Can't get groups data for %s to %s from %s.%s: %s\n",$StartMonth,$EndMonth,$Conf{'DBDatabase'},$Conf{'DBTableGrps'},$DBI::errstr);
176 # print caption (-c) with time period if -m or -p is set
177 # FIXME - month or period should handled differently
178 printf ("----- Report from %s to %s\n",$StartMonth,$EndMonth) if $Options{'c'} and ($Options{'m'} or $Options{'p'});
179 # print caption (-c) with newsgroup list if -n is set
180 printf ("----- Newsgroups: %s\n",join(',',split(/:/,$Newsgroups))) if $Options{'c'} and $Options{'n'};
181 # print caption (-c) with threshold if -t is set, taking -i in account
182 printf ("----- Threshold: %s %u\n",$Options{'i'} ? '<' : '>',$Options{'t'}) if $Options{'c'} and $Options{'t'};
183 if (!defined($Options{'b'}) and !defined($Options{'l'})) {
184 # default: neither -b nor -l
185 &OutputData($Options{'o'},$DBQuery,$MaxLength);
186 } elsif ($Options{'b'}) {
187 # -b is set (then -l can't be!)
188 # we have to read in the query results ourselves, as they do not have standard layout
189 while (my ($Newsgroup,$Postings) = $DBQuery->fetchrow_array) {
190 # we just assign "top x" or "bottom x" instead of a month for the caption
192 print &FormatOutput($Options{'o'}, ($Options{'i'} ? 'Bottom ' : 'Top ').$Options{'b'}, $Newsgroup, $Postings, $MaxLength);
195 # -l must be set now, as all other cases have been taken care of
196 # we have to read in the query results ourselves, as they do not have standard layout
197 while (my ($Month,$Newsgroup,$Postings) = $DBQuery->fetchrow_array) {
198 # we just switch $Newsgroups and $Month for output generation
200 print &FormatOutput($Options{'o'}, $Newsgroup, $Month, $Postings, 7);
205 $DBHandle->disconnect;
209 ################################ Documentation #################################
213 groupstats - create reports on newsgroup usage
217 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>]
221 See doc/README: Perl 5.8.x itself and the following modules from CPAN:
237 This script create reports on newsgroup usage (number of postings per
238 group per month) taken from result tables created by
241 The time period to act on defaults to last month; you can assign
242 another month via the B<-m> switch or a time period via the B<-p>
243 switch; the latter takes preference.
245 B<groupstats> will process all newsgroups by default; you can limit
246 that to only some newsgroups by supplying a list of those groups via
247 B<-n> (see below). You can include hierarchy levels in the output by
248 adding the B<-s> switch (see below).
250 Furthermore you can set a threshold via B<-t> so that only newsgroups
251 with more postings per month will be included in the report. You can
252 invert that by the B<-i> switch so only newsgroups with less than
253 I<threshold> postings per month will be included.
255 You can sort the output by number of postings per month instead of the
256 default (alphabetical list of newsgroups) by using B<-q>; you can
257 reverse the sorting order (from highest to lowest or in reversed
258 alphabetical order) by using B<-d>.
260 Furthermore, you can create a list of newsgroups that had consistently
261 more (or less) than x postings per month during the whole report
262 period by using B<-l> (together with B<i> as needed).
264 Last but not least you can create a "best of" list of the top x
265 newsgroups via B<-b> (or a "worst of" list by adding B<i>).
267 By default, B<groupstats> will dump a very simple alphabetical list of
268 newsgroups, one per line, followed by the number of postings in that
269 month. This output format of course cannot sensibly be combined with
270 time periods, so you can set the output format by using B<-o> (see
271 below). Captions can be added by setting the B<-c> switch.
275 F<groupstats.pl> will read its configuration from F<newsstats.conf>
276 which should be present in the same directory via Config::Auto.
278 See doc/INSTALL for an overview of possible configuration options.
280 You can override configuration options via the B<-g> switch.
286 =item B<-V> (version)
288 Print out version and copyright information on B<yapfaq> and exit.
292 Print this man page and exit.
294 =item B<-m> I<YYYY-MM> (month)
296 Set processing period to a month in YYYY-MM format. Ignored if B<-p>
299 =item B<-p> I<YYYY-MM:YYYY-MM> (period)
301 Set processing period to a time period between two month, each in
302 YYYY-MM format, separated by a colon. Overrides B<-m>.
304 =item B<-n> I<newsgroup(s)> (newsgroups)
306 Limit processing to a certain set of newsgroups. I<newsgroup(s)> can
307 be a single newsgroup name (de.alt.test), a newsgroup hierarchy
308 (de.alt.*) or a list of either of these, separated by colons, for
311 de.test:de.alt.test:de.newusers.*
313 =item B<-t> I<threshold> (threshold)
315 Only include newsgroups with more than I<threshold> postings per
316 month. Can be inverted by the B<-i> switch so that only newsgroups
317 with less than I<threshold> postings will be included.
319 This setting will be ignored if B<-l> or B<-b> is set.
321 =item B<-l> I<level> (level)
323 Only include newsgroups with more than I<level> postings per
324 month, every month during the whole reporting period. Can be inverted
325 by the B<-i> switch so that only newsgroups with less than I<level>
326 postings every single month will be included. Output will be ordered
327 by newsgroup name, followed by month.
329 This setting will be ignored if B<-b> is set. Overrides B<-t> and
330 can't be used together with B<-q> or B<-d>.
332 =item B<-b> I<n> (best of)
334 Create a list of the I<n> newsgroups with the most postings over the
335 whole reporting period. Can be inverted by the B<-i> switch so that a
336 list of the I<n> newsgroups with the least postings over the whole
337 period is generated. Output will be ordered by sum of postings.
339 Overrides B<-t> and B<-l> and can't be used together with B<-q> or
340 B<-d>. Output format is set to I<pretty> (see below).
344 Used in conjunction with B<-t>, B<-l> or B<-b> to set a lower
345 threshold or level or generate a "bottom list" instead of a top list.
347 =item B<-s> (sum per hierarchy level)
349 Include "virtual" groups for every hierarchy level in output, for
356 See the B<gatherstats> man page for details.
358 =item B<-o> I<output type> (output format)
360 Set output format. Default is I<dump>, consisting of an alphabetical
361 list of newsgroups, each on a new line, followed by the number of
362 postings in that month. This default format can't be used with time
363 periods of more than one month.
365 I<list> format is like I<dump>, but will print the month in front of
368 I<dumpgroup> format can only be use with a group list (see B<-n>) of
369 exactly one newsgroup and is like I<dump>, but will output months,
370 followed by the number of postings.
372 If you don't need easily parsable output, you'll mostly use I<pretty>
373 format, which will print a header for each new month and try to align
374 newsgroup names and posting counts. Usage of B<-b> will force this
377 =item B<-c> (captions)
379 Add captions to output (reporting period, newsgroups list, threshold).
381 =item B<-q> (quantity of postings)
383 Sort by number of postings instead of by newsgroup names.
385 Cannot be used with B<-l> or B<-b>.
387 =item B<-d> (descending)
389 Change sort order to descending.
391 Cannot be used with B<-l> or B<-b>.
393 =item B<-g> I<table> (postings per group table)
395 Override I<DBTableGrps> from F<newsstats.conf>.
405 Show number of postings per group for lasth month in I<dump> format:
409 Show that report for January of 2010 and de.alt.* plus de.test,
410 including display of hierarchy levels:
412 groupstats -m 2010-01 -n de.alt.*:de.test -s
414 Show that report for the year of 2010 in I<pretty> format:
416 groupstats -p 2010-01:2010-12 -o pretty
418 Only show newsgroups with less than 30 postings last month, ordered
419 by number of postings, descending, in I<pretty> format:
421 groupstats -iqdt 30 -o pretty
423 Show top 10 for the first half-year of of 2010 in I<pretty> format:
425 groupstats -p 2010-01:2010-06 -b 10 -o pretty
427 Report all groups that had less than 30 postings every singele month
428 in the year of 2010 (I<pretty> format is forced)
430 groupstats -p 2010-01:2010-12 -il 30
436 =item F<groupstats.pl>
440 =item F<NewsStats.pm>
442 Library functions for the NewsStats package.
444 =item F<newsstats.conf>
446 Runtime configuration file for B<yapfaq>.
452 Please report any bugs or feature requests to the author or use the
453 bug tracker at L<http://bugs.th-h.de/>!
473 This script is part of the B<NewsStats> package.
477 Thomas Hochstein <thh@inter.net>
479 =head1 COPYRIGHT AND LICENSE
481 Copyright (c) 2010 Thomas Hochstein <thh@inter.net>
483 This program is free software; you may redistribute it and/or modify it
484 under the same terms as Perl itself.