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:an:o:t:l:b:iscqdf:g:');
31 ### read configuration
32 my %Conf = %{ReadConfig('newsstats.conf')};
34 ### override configuration via commandline options
36 $ConfOverride{'DBTableGrps'} = $Options{'g'} if $Options{'g'};
37 &OverrideConfig(\%Conf,\%ConfOverride);
39 ### check for incompatible command line options
40 # you can't mix '-t', '-b' and '-l'
41 # -b/-l take preference over -t, and -b takes preference over -l
42 # you can't use '-f' with '-b' or '-l'
43 if ($Options{'b'} or $Options{'l'}) {
46 warn ("$MySelf: W: You cannot save the report to monthly files when using top lists (-b) or levels (-l). Filename template '-f $Options{'f'}' was ignored.\n");
51 warn ("$MySelf: W: You cannot combine thresholds (-t) and top lists (-b) or levels (-l). Threshold '-t $Options{'t'}' was ignored.\n");
54 if ($Options{'b'} and $Options{'l'}) {
56 warn ("$MySelf: W: You cannot combine top lists (-b) and levels (-l). Level '-l $Options{'l'}' was ignored.\n");
59 # -q/-d don't work with -b or -l
60 warn ("$MySelf: W: Sorting by number of postings (-q) ignored due to top list mode (-b) / levels (-l).\n") if $Options{'q'};
61 warn ("$MySelf: W: Reverse sorting (-d) ignored due to top list mode (-b) / levels (-l).\n") if $Options{'d'};
65 # default output type to 'pretty'
66 $Options{'o'} = 'pretty' if !$Options{'o'};
67 # fail if more than one newsgroup is combined with 'dumpgroup' type
68 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 $Options{'n'} =~ /:|\*/);
69 # accept 'dumpgroup' only with -n
70 if ($Options{'o'} eq 'dumpgroup' and !defined($Options{'n'})) {
71 $Options{'o'} = 'dump';
72 warn ("$MySelf: W: You must submit exactly one newsgroup ('-n news.group') for '-o dumpgroup'. Output type was set to 'dump'.\n");
74 # set output type to 'pretty' for -l
75 if ($Options{'l'} and $Options{'o'} ne 'pretty') {
76 $Options{'o'} = 'pretty';
77 warn ("$MySelf: W: Output type forced to '-o pretty' due to usage of '-l'.\n");
79 # set output type to 'dump' for -f
80 if ($Options{'f'} and $Options{'o'} ne 'dump') {
81 $Options{'o'} = 'dump';
82 warn ("$MySelf: W: Output type forced to '-o dump' due to usage of '-f'.\n");
86 my $DBHandle = InitDB(\%Conf,1);
89 my ($StartMonth,$EndMonth);
90 # if '-a' is set, set start/end month from database
91 # FIXME - it doesn't make that much sense to get first/last month from database to query it
92 # with a time period that equals no time period ...
96 my $DBQuery = $DBHandle->prepare(sprintf("SELECT MIN(month),MAX(month) FROM %s.%s",$Conf{'DBDatabase'},$Conf{'DBTableGrps'}));
97 $DBQuery->execute or die sprintf("$MySelf: E: Can't get MIN/MAX month from %s.%s: %s\n",$Conf{'DBDatabase'},$Conf{'DBTableGrps'},$DBI::errstr);
98 ($StartMonth,$EndMonth) = $DBQuery->fetchrow_array;
100 ($StartMonth,$EndMonth) = &GetTimePeriod($Options{'m'},$Options{'p'});
102 # if time period is more than one month: force output type to '-o pretty' or '-o dumpgroup'
103 if ($Options{'o'} eq 'dump' and ($Options{'p'} or $Options{'a'})) {
104 if (defined($Options{'n'}) and $Options{'n'} !~ /:|\*/) {
105 # just one newsgroup is defined
106 warn ("$MySelf: W: You cannot combine time periods (-p) with '-o dump', changing output type to '-o dumpgroup'.\n");
107 $Options{'o'} = 'dumpgroup';
108 } elsif (!defined($Options{'f'})) {
109 # more than one newsgroup - and no file output
110 warn ("$MySelf: W: You cannot combine time periods (-p) with '-o dump', changing output type to '-o pretty'.\n");
111 $Options{'o'} = 'pretty';
116 # get list of newsgroups (-n)
117 my ($QueryGroupList,$QueryThreshold,@GroupList,@Params);
118 my $Newsgroups = $Options{'n'};
120 # explode list of newsgroups for WHERE clause
121 ($QueryGroupList,@GroupList) = &SQLGroupList($Newsgroups);
123 # set to dummy value (always true)
128 if (defined($Options{'t'})) {
130 # -i: list groups below threshold
131 $QueryThreshold .= ' postings < ?';
133 # default: list groups above threshold
134 $QueryThreshold .= ' postings > ?';
136 # push threshold to Params
137 push @Params,$Options{'t'};
139 # set to dummy value (always true)
143 # construct WHERE clause
144 # $QueryGroupList is "list of newsgroup" (or 1),
145 # $QueryThreshold is threshold definition (or 1),
146 # &SQLHierarchies() takes care of the exclusion of hierarchy levels (.ALL)
147 # according to setting of -s
148 my $WhereClause = sprintf('month BETWEEN ? AND ? AND %s AND %s %s',$QueryGroupList,$QueryThreshold,&SQLHierarchies($Options{'s'}));
150 # get length of longest newsgroup delivered by query for formatting purposes
152 my $MaxLength = &GetMaxLenght($DBHandle,$Conf{'DBTableGrps'},'newsgroup',$WhereClause,$StartMonth,$EndMonth,(@GroupList,@Params));
154 my ($OrderClause,$DBQuery);
155 # -b (best of / top list) defined?
156 if (!defined($Options{'b'}) and !defined($Options{'l'})) {
157 # default: neither -b nor -l
158 # set ordering (ORDER BY) to "newsgroups" or "postings", "ASC" or "DESC"
159 # according to -q and -d
160 $OrderClause = 'newsgroup';
161 $OrderClause = 'postings' if $Options{'q'};
162 $OrderClause .= ' DESC' if $Options{'d'};
163 # prepare query: get number of postings per group from groups table for given months and newsgroups
164 $DBQuery = $DBHandle->prepare(sprintf("SELECT month,newsgroup,postings FROM %s.%s WHERE %s ORDER BY month,%s",$Conf{'DBDatabase'},$Conf{'DBTableGrps'},$WhereClause,$OrderClause));
165 } elsif ($Options{'b'}) {
166 # -b is set (then -l can't be!)
167 # set sorting order (-i): top or flop list?
169 $OrderClause = 'postings';
171 $OrderClause = 'postings DESC';
173 # set -b to 10 if < 1 (Top 10)
174 $Options{'b'} = 10 if $Options{'b'} !~ /^\d*$/ or $Options{'b'} < 1;
175 # push LIMIT to Params
176 push @Params,$Options{'b'};
177 # prepare query: get sum of postings per group from groups table for given months and newsgroups with LIMIT
178 $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));
180 # -l must be set now, as all other cases have been taken care of
181 # which kind of level (-i): more than -l x or less than -l x?
188 # prepare and execute query: get list of newsgroups meeting level condition
189 $DBQuery = $DBHandle->prepare(sprintf("SELECT newsgroup FROM %s.%s WHERE %s GROUP BY newsgroup HAVING MAX(postings) %s ?",$Conf{'DBDatabase'},$Conf{'DBTableGrps'},$WhereClause,$Level));
190 $DBQuery->execute($StartMonth,$EndMonth,@GroupList,$Options{'l'})
191 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);
192 # add newsgroups to a comma-seperated list ready for IN(...) query
194 while (my ($Newsgroup) = $DBQuery->fetchrow_array) {
195 $GroupList .= ',' if (defined($GroupList) and $GroupList ne '');
196 $GroupList .= "'$Newsgroup'";
198 $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));
202 $DBQuery->execute($StartMonth,$EndMonth,@GroupList,@Params)
203 or die sprintf("$MySelf: E: Can't get groups data for %s to %s from %s.%s: %s\n",$StartMonth,$EndMonth,$Conf{'DBDatabase'},$Conf{'DBTableGrps'},$DBI::errstr);
206 # reset caption (-c) if -f is set
207 undef($Options{'c'}) if $Options{'f'};
208 # print caption (-c) with time period if -m or -p is set
211 printf ("----- Report from %s to %s\n",$StartMonth,$EndMonth);
212 } elsif ($Options{'m'}) {
213 printf ("----- Report for %s\n",$StartMonth);
216 # print caption (-c) with newsgroup list if -n is set
217 printf ("----- Newsgroups: %s\n",join(',',split(/:/,$Newsgroups))) if $Options{'c'} and $Options{'n'};
218 # print caption (-c) with threshold if -t is set, taking -i in account
219 printf ("----- Threshold: %s %u\n",$Options{'i'} ? '<' : '>',$Options{'t'}) if $Options{'c'} and $Options{'t'};
220 if (!defined($Options{'b'}) and !defined($Options{'l'})) {
221 # default: neither -b nor -l
222 &OutputData($Options{'o'},$Options{'f'},$DBQuery,$MaxLength);
223 } elsif ($Options{'b'}) {
224 # -b is set (then -l can't be!)
225 # we have to read in the query results ourselves, as they do not have standard layout
226 while (my ($Newsgroup,$Postings) = $DBQuery->fetchrow_array) {
227 # we just assign "top x" or "bottom x" instead of a month for the caption and force an output type of pretty
228 print &FormatOutput('pretty', ($Options{'i'} ? 'Bottom ' : 'Top ').$Options{'b'}, $Newsgroup, $Postings, $MaxLength);
231 # -l must be set now, as all other cases have been taken care of
232 # print caption (-c) with level, taking -i in account
233 printf ("----- Newsgroups with %s than %u postings over the whole time period\n",$Options{'i'} ? 'less' : 'more',$Options{'l'}) if $Options{'c'};
234 # we have to read in the query results ourselves, as they do not have standard layout
235 while (my ($Month,$Newsgroup,$Postings) = $DBQuery->fetchrow_array) {
236 # we just switch $Newsgroups and $Month for output generation
237 print &FormatOutput($Options{'o'}, $Newsgroup, $Month, $Postings, 7);
242 $DBHandle->disconnect;
246 ################################ Documentation #################################
250 groupstats - create reports on newsgroup usage
254 B<groupstats> [B<-Vhiscqd>] [B<-m> I<YYYY-MM> | B<-p> I<YYYY-MM:YYYY-MM> | B<-a>] [B<-n> I<newsgroup(s)>] [B<-t> I<threshold>] [B<-l> I<level>] [B<-b> I<number>] [B<-o> I<output type>] [B<-f> I<filename template>] [B<-g> I<database table>]
258 See doc/README: Perl 5.8.x itself and the following modules from CPAN:
274 This script create reports on newsgroup usage (number of postings per
275 group per month) taken from result tables created by
278 The time period to act on defaults to last month; you can assign
279 another month via the B<-m> switch or a time period via the B<-p>
280 switch; the latter takes preference.
282 B<groupstats> will process all newsgroups by default; you can limit
283 that to only some newsgroups by supplying a list of those groups via
284 B<-n> (see below). You can include hierarchy levels in the output by
285 adding the B<-s> switch (see below).
287 Furthermore you can set a threshold via B<-t> so that only newsgroups
288 with more postings per month will be included in the report. You can
289 invert that by the B<-i> switch so only newsgroups with less than
290 I<threshold> postings per month will be included.
292 You can sort the output by number of postings per month instead of the
293 default (alphabetical list of newsgroups) by using B<-q>; you can
294 reverse the sorting order (from highest to lowest or in reversed
295 alphabetical order) by using B<-d>.
297 Furthermore, you can create a list of newsgroups that had consistently
298 more (or less) than x postings per month during the whole report
299 period by using B<-l> (together with B<i> as needed).
301 Last but not least you can create a "best of" list of the top x
302 newsgroups via B<-b> (or a "worst of" list by adding B<i>).
304 By default, B<groupstats> will dump an alphabetical list of newsgroups,
305 one per line, followed by the number of postings in that group, for
306 every month. You can change the output format by using B<-o> (see
307 below). Captions can be added by setting the B<-c> switch.
311 B<groupstats> will read its configuration from F<newsstats.conf>
312 which should be present in the same directory via Config::Auto.
314 See doc/INSTALL for an overview of possible configuration options.
316 You can override configuration options via the B<-g> switch.
322 =item B<-V> (version)
324 Print out version and copyright information on B<yapfaq> and exit.
328 Print this man page and exit.
330 =item B<-m> I<YYYY-MM> (month)
332 Set processing period to a month in YYYY-MM format. Ignored if B<-p>
335 =item B<-p> I<YYYY-MM:YYYY-MM> (period)
337 Set processing period to a time period between two month, each in
338 YYYY-MM format, separated by a colon. Overrides B<-m>. Ignored if
343 Set no processing period (process whole database). Overrides B<-m>
346 =item B<-n> I<newsgroup(s)> (newsgroups)
348 Limit processing to a certain set of newsgroups. I<newsgroup(s)> can
349 be a single newsgroup name (de.alt.test), a newsgroup hierarchy
350 (de.alt.*) or a list of either of these, separated by colons, for
353 de.test:de.alt.test:de.newusers.*
355 =item B<-t> I<threshold> (threshold)
357 Only include newsgroups with more than I<threshold> postings per
358 month. Can be inverted by the B<-i> switch so that only newsgroups
359 with less than I<threshold> postings will be included.
361 This setting will be ignored if B<-l> or B<-b> is set.
363 =item B<-l> I<level> (level)
365 Only include newsgroups with more than I<level> postings per
366 month, every month during the whole reporting period. Can be inverted
367 by the B<-i> switch so that only newsgroups with less than I<level>
368 postings every single month will be included. Output will be ordered
369 by newsgroup name, followed by month.
371 This setting will be ignored if B<-b> is set. Overrides B<-t> and
372 can't be used together with B<-q>, B<-d> or B<-f>.
374 =item B<-b> I<n> (best of)
376 Create a list of the I<n> newsgroups with the most postings over the
377 whole reporting period. Can be inverted by the B<-i> switch so that a
378 list of the I<n> newsgroups with the least postings over the whole
379 period is generated. Output will be ordered by sum of postings.
381 Overrides B<-t> and B<-l> and can't be used together with B<-q>, B<-d>
382 or B<-f>. Output format is set to I<pretty> (see below).
386 Used in conjunction with B<-t>, B<-l> or B<-b> to set a lower
387 threshold or level or generate a "bottom list" instead of a top list.
389 =item B<-s> (sum per hierarchy level)
391 Include "virtual" groups for every hierarchy level in output, for
398 See the B<gatherstats> man page for details.
400 =item B<-o> I<output type> (output format)
402 Set output format. Default is I<pretty>, which will print a header for
403 each new month, followed by an alphabetical list of newsgroups, each
404 on a new line, followed by the number of postings in that month.
405 B<groupstats> will try to align newsgroup names and posting counts.
406 Usage of B<-b> will force this format; it cannot be used together with
409 I<dump> format is used to create an easily parsable output consisting
410 of an alphabetical list of newsgroups, each on a new line, followed by
411 the number of postings in that month, without any alignment. This
412 default format can't be used with time periods of more than one month.
413 Usage of B<-f> will force this format.
415 I<list> format is like I<dump>, but will print the month in front of
418 I<dumpgroup> format can only be use with a group list (see B<-n>) of
419 exactly one newsgroup and is like I<dump>, but will output months,
420 followed by the number of postings.
422 =item B<-c> (captions)
424 Add captions to output (reporting period, newsgroups list, threshold
427 This setting will be ignored if B<-f> is set.
429 =item B<-q> (quantity of postings)
431 Sort by number of postings instead of by newsgroup names.
433 Cannot be used with B<-l> or B<-b>.
435 =item B<-d> (descending)
437 Change sort order to descending.
439 Cannot be used with B<-l> or B<-b>.
441 =item B<-f> I<filename template> (output file)
443 Save output to file instead of dumping it to STDOUT. B<groupstats>
444 will create one file for each month, with filenames composed by
445 adding year and month to the I<filename template>, for example
452 This setting will be ignored if B<-l> or B<-b> is set. Output format
453 is set to I<dump> (see above).
455 =item B<-g> I<table> (postings per group table)
457 Override I<DBTableGrps> from F<newsstats.conf>.
467 Show number of postings per group for lasth month in I<dump> format:
471 Show that report for January of 2010 and de.alt.* plus de.test,
472 including display of hierarchy levels:
474 groupstats -m 2010-01 -n de.alt.*:de.test -s
476 Show that report for the year of 2010 in I<pretty> format:
478 groupstats -p 2010-01:2010-12 -o pretty
480 Only show newsgroups with less than 30 postings last month, ordered
481 by number of postings, descending, in I<pretty> format:
483 groupstats -iqdt 30 -o pretty
485 Show top 10 for the first half-year of of 2010 in I<pretty> format:
487 groupstats -p 2010-01:2010-06 -b 10 -o pretty
489 Report all groups that had less than 30 postings every singele month
490 in the year of 2010 (I<pretty> format is forced)
492 groupstats -p 2010-01:2010-12 -il 30
498 =item F<groupstats.pl>
502 =item F<NewsStats.pm>
504 Library functions for the NewsStats package.
506 =item F<newsstats.conf>
508 Runtime configuration file for B<yapfaq>.
514 Please report any bugs or feature requests to the author or use the
515 bug tracker at L<http://bugs.th-h.de/>!
535 This script is part of the B<NewsStats> package.
539 Thomas Hochstein <thh@inter.net>
541 =head1 COPYRIGHT AND LICENSE
543 Copyright (c) 2010 Thomas Hochstein <thh@inter.net>
545 This program is free software; you may redistribute it and/or modify it
546 under the same terms as Perl itself.