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 'dump'
66 $Options{'o'} = 'dump' if !$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
76 $Options{'o'} = 'pretty';
77 warn ("$MySelf: W: Output type forced to '-o pretty' due to usage of '-l'.\n");
81 my $DBHandle = InitDB(\%Conf,1);
84 my ($StartMonth,$EndMonth);
85 # if '-a' is set, set start/end month from database
86 # FIXME - it doesn't make that much sense to get first/last month from database to query it
87 # with a time period that equals no time period ...
91 my $DBQuery = $DBHandle->prepare(sprintf("SELECT MIN(month),MAX(month) FROM %s.%s",$Conf{'DBDatabase'},$Conf{'DBTableGrps'}));
92 $DBQuery->execute or die sprintf("$MySelf: E: Can't get MIN/MAX month from %s.%s: %s\n",$Conf{'DBDatabase'},$Conf{'DBTableGrps'},$DBI::errstr);
93 ($StartMonth,$EndMonth) = $DBQuery->fetchrow_array;
95 ($StartMonth,$EndMonth) = &GetTimePeriod($Options{'m'},$Options{'p'});
97 # if time period is more than one month: set output type to '-o pretty' or '-o dumpgroup'
98 if ($Options{'o'} eq 'dump' and ($Options{'p'} or $Options{'a'})) {
99 if (defined($Options{'n'}) and $Options{'n'} !~ /:|\*/) {
100 # just one newsgroup is defined
101 warn ("$MySelf: W: You cannot combine time periods (-p) with '-o dump', changing output type to '-o dumpgroup'.\n");
102 $Options{'o'} = 'dumpgroup';
103 } elsif (!defined($Options{'f'})) {
104 # more than one newsgroup - and no file output
105 warn ("$MySelf: W: You cannot combine time periods (-p) with '-o dump', changing output type to '-o pretty'.\n");
106 $Options{'o'} = 'pretty';
111 # get list of newsgroups (-n)
112 my ($QueryGroupList,$QueryThreshold,@GroupList,@Params);
113 my $Newsgroups = $Options{'n'};
115 # explode list of newsgroups for WHERE clause
116 ($QueryGroupList,@GroupList) = &SQLGroupList($Newsgroups);
118 # set to dummy value (always true)
123 if (defined($Options{'t'})) {
125 # -i: list groups below threshold
126 $QueryThreshold .= ' postings < ?';
128 # default: list groups above threshold
129 $QueryThreshold .= ' postings > ?';
131 # push threshold to Params
132 push @Params,$Options{'t'};
134 # set to dummy value (always true)
138 # construct WHERE clause
139 # $QueryGroupList is "list of newsgroup" (or 1),
140 # $QueryThreshold is threshold definition (or 1),
141 # &SQLHierarchies() takes care of the exclusion of hierarchy levels (.ALL)
142 # according to setting of -s
143 my $WhereClause = sprintf('month BETWEEN ? AND ? AND %s AND %s %s',$QueryGroupList,$QueryThreshold,&SQLHierarchies($Options{'s'}));
145 # get length of longest newsgroup delivered by query for formatting purposes
147 my $MaxLength = &GetMaxLenght($DBHandle,$Conf{'DBTableGrps'},'newsgroup',$WhereClause,$StartMonth,$EndMonth,(@GroupList,@Params));
149 my ($OrderClause,$DBQuery);
150 # -b (best of / top list) defined?
151 if (!defined($Options{'b'}) and !defined($Options{'l'})) {
152 # default: neither -b nor -l
153 # set ordering (ORDER BY) to "newsgroups" or "postings", "ASC" or "DESC"
154 # according to -q and -d
155 $OrderClause = 'newsgroup';
156 $OrderClause = 'postings' if $Options{'q'};
157 $OrderClause .= ' DESC' if $Options{'d'};
158 # prepare query: get number of postings per group from groups table for given months and newsgroups
159 $DBQuery = $DBHandle->prepare(sprintf("SELECT month,newsgroup,postings FROM %s.%s WHERE %s ORDER BY month,%s",$Conf{'DBDatabase'},$Conf{'DBTableGrps'},$WhereClause,$OrderClause));
160 } elsif ($Options{'b'}) {
161 # -b is set (then -l can't be!)
162 # set sorting order (-i): top or flop list?
164 $OrderClause = 'postings';
166 $OrderClause = 'postings DESC';
168 # set -b to 10 if < 1 (Top 10)
169 $Options{'b'} = 10 if $Options{'b'} !~ /^\d*$/ or $Options{'b'} < 1;
170 # push LIMIT to Params
171 push @Params,$Options{'b'};
172 # prepare query: get sum of postings per group from groups table for given months and newsgroups with LIMIT
173 $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));
175 # -l must be set now, as all other cases have been taken care of
176 # which kind of level (-i): more than -l x or less than -l x?
183 # prepare and execute query: get list of newsgroups meeting level condition
184 $DBQuery = $DBHandle->prepare(sprintf("SELECT newsgroup FROM %s.%s WHERE %s GROUP BY newsgroup HAVING MAX(postings) %s ?",$Conf{'DBDatabase'},$Conf{'DBTableGrps'},$WhereClause,$Level));
185 $DBQuery->execute($StartMonth,$EndMonth,@GroupList,$Options{'l'})
186 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);
187 # add newsgroups to a comma-seperated list ready for IN(...) query
189 while (my ($Newsgroup) = $DBQuery->fetchrow_array) {
190 $GroupList .= ',' if (defined($GroupList) and $GroupList ne '');
191 $GroupList .= "'$Newsgroup'";
193 $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));
197 $DBQuery->execute($StartMonth,$EndMonth,@GroupList,@Params)
198 or die sprintf("$MySelf: E: Can't get groups data for %s to %s from %s.%s: %s\n",$StartMonth,$EndMonth,$Conf{'DBDatabase'},$Conf{'DBTableGrps'},$DBI::errstr);
201 # reset caption (-c) if -f is set
202 undef($Options{'c'}) if $Options{'f'};
203 # print caption (-c) with time period if -m or -p is set
206 printf ("----- Report from %s to %s\n",$StartMonth,$EndMonth);
207 } elsif ($Options{'m'}) {
208 printf ("----- Report for %s\n",$StartMonth);
211 # print caption (-c) with newsgroup list if -n is set
212 printf ("----- Newsgroups: %s\n",join(',',split(/:/,$Newsgroups))) if $Options{'c'} and $Options{'n'};
213 # print caption (-c) with threshold if -t is set, taking -i in account
214 printf ("----- Threshold: %s %u\n",$Options{'i'} ? '<' : '>',$Options{'t'}) if $Options{'c'} and $Options{'t'};
215 if (!defined($Options{'b'}) and !defined($Options{'l'})) {
216 # default: neither -b nor -l
217 &OutputData($Options{'o'},$Options{'f'},$DBQuery,$MaxLength);
218 } elsif ($Options{'b'}) {
219 # -b is set (then -l can't be!)
220 # we have to read in the query results ourselves, as they do not have standard layout
221 while (my ($Newsgroup,$Postings) = $DBQuery->fetchrow_array) {
222 # we just assign "top x" or "bottom x" instead of a month for the caption and force an output type of pretty
223 print &FormatOutput('pretty', ($Options{'i'} ? 'Bottom ' : 'Top ').$Options{'b'}, $Newsgroup, $Postings, $MaxLength);
226 # -l must be set now, as all other cases have been taken care of
227 # print caption (-c) with level, taking -i in account
228 printf ("----- Newsgroups with %s than %u postings over the whole time period\n",$Options{'i'} ? 'less' : 'more',$Options{'l'}) if $Options{'c'};
229 # we have to read in the query results ourselves, as they do not have standard layout
230 while (my ($Month,$Newsgroup,$Postings) = $DBQuery->fetchrow_array) {
231 # we just switch $Newsgroups and $Month for output generation
232 print &FormatOutput($Options{'o'}, $Newsgroup, $Month, $Postings, 7);
237 $DBHandle->disconnect;
241 ################################ Documentation #################################
245 groupstats - create reports on newsgroup usage
249 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>]
253 See doc/README: Perl 5.8.x itself and the following modules from CPAN:
269 This script create reports on newsgroup usage (number of postings per
270 group per month) taken from result tables created by
273 The time period to act on defaults to last month; you can assign
274 another month via the B<-m> switch or a time period via the B<-p>
275 switch; the latter takes preference.
277 B<groupstats> will process all newsgroups by default; you can limit
278 that to only some newsgroups by supplying a list of those groups via
279 B<-n> (see below). You can include hierarchy levels in the output by
280 adding the B<-s> switch (see below).
282 Furthermore you can set a threshold via B<-t> so that only newsgroups
283 with more postings per month will be included in the report. You can
284 invert that by the B<-i> switch so only newsgroups with less than
285 I<threshold> postings per month will be included.
287 You can sort the output by number of postings per month instead of the
288 default (alphabetical list of newsgroups) by using B<-q>; you can
289 reverse the sorting order (from highest to lowest or in reversed
290 alphabetical order) by using B<-d>.
292 Furthermore, you can create a list of newsgroups that had consistently
293 more (or less) than x postings per month during the whole report
294 period by using B<-l> (together with B<i> as needed).
296 Last but not least you can create a "best of" list of the top x
297 newsgroups via B<-b> (or a "worst of" list by adding B<i>).
299 By default, B<groupstats> will dump a very simple alphabetical list of
300 newsgroups, one per line, followed by the number of postings in that
301 month. This output format of course cannot sensibly be combined with
302 time periods, so you can set the output format by using B<-o> (see
303 below). Captions can be added by setting the B<-c> switch.
307 B<groupstats> will read its configuration from F<newsstats.conf>
308 which should be present in the same directory via Config::Auto.
310 See doc/INSTALL for an overview of possible configuration options.
312 You can override configuration options via the B<-g> switch.
318 =item B<-V> (version)
320 Print out version and copyright information on B<yapfaq> and exit.
324 Print this man page and exit.
326 =item B<-m> I<YYYY-MM> (month)
328 Set processing period to a month in YYYY-MM format. Ignored if B<-p>
331 =item B<-p> I<YYYY-MM:YYYY-MM> (period)
333 Set processing period to a time period between two month, each in
334 YYYY-MM format, separated by a colon. Overrides B<-m>. Ignored if
339 Set no processing period (process whole database). Overrides B<-m>
342 =item B<-n> I<newsgroup(s)> (newsgroups)
344 Limit processing to a certain set of newsgroups. I<newsgroup(s)> can
345 be a single newsgroup name (de.alt.test), a newsgroup hierarchy
346 (de.alt.*) or a list of either of these, separated by colons, for
349 de.test:de.alt.test:de.newusers.*
351 =item B<-t> I<threshold> (threshold)
353 Only include newsgroups with more than I<threshold> postings per
354 month. Can be inverted by the B<-i> switch so that only newsgroups
355 with less than I<threshold> postings will be included.
357 This setting will be ignored if B<-l> or B<-b> is set.
359 =item B<-l> I<level> (level)
361 Only include newsgroups with more than I<level> postings per
362 month, every month during the whole reporting period. Can be inverted
363 by the B<-i> switch so that only newsgroups with less than I<level>
364 postings every single month will be included. Output will be ordered
365 by newsgroup name, followed by month.
367 This setting will be ignored if B<-b> is set. Overrides B<-t> and
368 can't be used together with B<-q>, B<-d> or B<-f>.
370 =item B<-b> I<n> (best of)
372 Create a list of the I<n> newsgroups with the most postings over the
373 whole reporting period. Can be inverted by the B<-i> switch so that a
374 list of the I<n> newsgroups with the least postings over the whole
375 period is generated. Output will be ordered by sum of postings.
377 Overrides B<-t> and B<-l> and can't be used together with B<-q>, B<-d>
378 or B<-f>. Output format is set to I<pretty> (see below).
382 Used in conjunction with B<-t>, B<-l> or B<-b> to set a lower
383 threshold or level or generate a "bottom list" instead of a top list.
385 =item B<-s> (sum per hierarchy level)
387 Include "virtual" groups for every hierarchy level in output, for
394 See the B<gatherstats> man page for details.
396 =item B<-o> I<output type> (output format)
398 Set output format. Default is I<dump>, consisting of an alphabetical
399 list of newsgroups, each on a new line, followed by the number of
400 postings in that month. This default format can't be used with time
401 periods of more than one month.
403 I<list> format is like I<dump>, but will print the month in front of
406 I<dumpgroup> format can only be use with a group list (see B<-n>) of
407 exactly one newsgroup and is like I<dump>, but will output months,
408 followed by the number of postings.
410 If you don't need easily parsable output, you'll mostly use I<pretty>
411 format, which will print a header for each new month and try to align
412 newsgroup names and posting counts. Usage of B<-b> will force this
415 =item B<-c> (captions)
417 Add captions to output (reporting period, newsgroups list, threshold
420 This setting will be ignored if B<-f> is set.
422 =item B<-q> (quantity of postings)
424 Sort by number of postings instead of by newsgroup names.
426 Cannot be used with B<-l> or B<-b>.
428 =item B<-d> (descending)
430 Change sort order to descending.
432 Cannot be used with B<-l> or B<-b>.
434 =item B<-f> I<filename template> (output file)
436 Save output to file instead of dumping it to STDOUT. B<groupstats>
437 will create one file for each month, with filenames composed by
438 adding year and month to the I<filename template>, for example
445 This setting will be ignored if B<-l> or B<-b> is set. Output format
446 is set to I<dump> (see above).
448 =item B<-g> I<table> (postings per group table)
450 Override I<DBTableGrps> from F<newsstats.conf>.
460 Show number of postings per group for lasth month in I<dump> format:
464 Show that report for January of 2010 and de.alt.* plus de.test,
465 including display of hierarchy levels:
467 groupstats -m 2010-01 -n de.alt.*:de.test -s
469 Show that report for the year of 2010 in I<pretty> format:
471 groupstats -p 2010-01:2010-12 -o pretty
473 Only show newsgroups with less than 30 postings last month, ordered
474 by number of postings, descending, in I<pretty> format:
476 groupstats -iqdt 30 -o pretty
478 Show top 10 for the first half-year of of 2010 in I<pretty> format:
480 groupstats -p 2010-01:2010-06 -b 10 -o pretty
482 Report all groups that had less than 30 postings every singele month
483 in the year of 2010 (I<pretty> format is forced)
485 groupstats -p 2010-01:2010-12 -il 30
491 =item F<groupstats.pl>
495 =item F<NewsStats.pm>
497 Library functions for the NewsStats package.
499 =item F<newsstats.conf>
501 Runtime configuration file for B<yapfaq>.
507 Please report any bugs or feature requests to the author or use the
508 bug tracker at L<http://bugs.th-h.de/>!
528 This script is part of the B<NewsStats> package.
532 Thomas Hochstein <thh@inter.net>
534 =head1 COPYRIGHT AND LICENSE
536 Copyright (c) 2010 Thomas Hochstein <thh@inter.net>
538 This program is free software; you may redistribute it and/or modify it
539 under the same terms as Perl itself.