You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@labs.apache.org by co...@apache.org on 2007/01/20 22:21:48 UTC
svn commit: r498171 - in /labs/pulse/report-email: ./ listex-report.pl
Author: coar
Date: Sat Jan 20 13:21:47 2007
New Revision: 498171
URL: http://svn.apache.org/viewvc?view=rev&rev=498171
Log:
Scripts for dealing with queries (command line or email)
Added:
labs/pulse/report-email/
labs/pulse/report-email/listex-report.pl (with props)
Added: labs/pulse/report-email/listex-report.pl
URL: http://svn.apache.org/viewvc/labs/pulse/report-email/listex-report.pl?view=auto&rev=498171
==============================================================================
--- labs/pulse/report-email/listex-report.pl (added)
+++ labs/pulse/report-email/listex-report.pl Sat Jan 20 13:21:47 2007
@@ -0,0 +1,438 @@
+#! /usr/local/bin/perl -w
+#
+# Extract the body of a mail message and treat each line in it as a
+# regex for matching mailing lists on which to report.
+#
+use strict;
+use DBD::mysql;
+use DBI;
+use Getopt::Long;
+use Mail::Mailer;
+use MIME::Parser;
+use Text::Abbrev;
+use Text::Shellwords;
+
+my %options;
+my $PROG;
+my $msg = '';
+my $batch = 0;
+my $debug = 0;
+my $mail2stdout = 0;
+
+my $verbs = abbrev qw( debug
+ end
+ help
+ list
+ moderated-by
+ );
+
+Getopt::Long::Configure('bundling');
+Getopt::Long::Configure('auto_abbrev');
+Getopt::Long::Configure('pass_through');
+GetOptions(\%options,
+ 'batch|r' => \$batch,
+ 'exec|e=s' => \$PROG,
+ 'test|R' => \$mail2stdout,
+ 'verbose|v+' => \$debug,
+ );
+
+my @patterns;
+my $parser;
+my $entity;
+my $from;
+my $abort = 0;
+
+debug("Debug level=$debug");
+#
+# See if we're doing a command line, or reading from a mail message.
+#
+if (! $batch) {
+ #
+ # Mail message. Trap any error text and include it in the output.
+ #
+ $SIG{__WARN__} = sub {
+ $msg .= $_[0];
+ };
+ debug(4, 'Operating from email input');
+ #
+ # Okey, get the message. If we can't, just die messily -- there's
+ # no way to include anything in a reply.
+ #
+ $parser = new MIME::Parser;
+ $parser->output_to_core(10000000);
+ $parser->output_dir('/tmp/');
+ eval {
+ $parser->tmp_recycling(0);
+ };
+ $entity = $parser->read(\*STDIN) or do {
+ print STDERR $msg;
+ die("Couldn't get message from stdin!");
+ };
+
+ #
+ # Get the sender (to whom the response will be sent), and the subject.
+ #
+ $from = ($entity->get('Reply-to')
+ || $entity->get('From')
+ || $entity->get('Sender'));
+ $from =~ s§[\r\n\s]+]$§§g;
+ $from = normalise_eddress($from);
+ if ($from !~ /\@apache.org$/i) {
+ $msg .= "Access restricted.\n\n"
+ . "You are not authorised to access this information.\n";
+ $abort = 1;
+ }
+ elsif ((my $nparts = $entity->parts()) > 1) {
+ $msg .= "Cannot handle complex or multipart messages. Please send\n"
+ . "normal, unsigned, unencoded plaintext. Your message\n"
+ . "had $nparts parts/attachments.\n";
+ $abort = 1;
+ }
+ else {
+ #
+ # Get the body.
+ #
+ my $body = $entity->bodyhandle()->as_string();
+ @patterns = split(m§[\r\n]+§, $body);
+ }
+}
+else {
+ my $command_line = join(' ', @ARGV);
+ debug(4, "Operating from command-line input: '$command_line'");
+ @patterns = ($command_line);
+}
+
+#
+# Ready to make the doughnuts..
+#
+debug('here we go');
+if (! $abort) {
+ if (@patterns) {
+ for (@patterns) {
+ debug(3, "Pattern input: '$_'");
+ }
+ $msg .= fetchdata(\@patterns);
+ }
+ else {
+ $msg .= "\nNo command nor regex.\n";
+ debug('No pattern input!');
+ }
+}
+
+#
+# Okey, time to issue our report.
+#
+if ($batch) {
+ print $msg;
+}
+else {
+ #
+ # We received it via mail, so reply the same way.
+ #
+ my $subject = $entity->get('Subject');
+ $subject =~ s§[\r\n\s]+]$§§g;
+ $subject =~ s§^(?:(?:Re|Sv|Aw|Antwort):\s*)+§§gi;
+ my $refs = $entity->get('References');
+ my $msgid = $entity->get('Message-ID');
+
+ my $reply = new Mail::Mailer('sendmail');
+ my %header = ('Subject' => "Re: $subject",
+ 'To' => $from,
+ 'Bcc' => 'Ken.Coar@Golux.Com',
+ 'From' => ('Apache Mailing List Info '
+ . '<do...@Apache.Org>'),
+ 'Content-Type' => 'text/plain',
+ 'MIME-Version' => '1.0');
+ if ($msgid) {
+ chomp($refs) if ($refs);
+ chomp($msgid) if ($msgid);
+ $refs = $msgid . ($refs ? ", $refs" : '');
+ $header{'In-Reply-To'} = $msgid;
+ $header{'References'} = $refs;
+ }
+ #
+ # Are we actually returning the reply, or just testing?
+ #
+ if (! $mail2stdout) {
+ $reply->open(\%header);
+ print $reply $msg;
+ $reply->close();
+ }
+ elsif ($debug) {
+ print $msg;
+ }
+}
+exit(0);
+
+#
+# Turn a possibly fancy eddress into the basic user@domain string.
+#
+sub normalise_eddress {
+ my $email = shift;
+ my @address_bits = Mail::Address->parse($email);
+ my $envaddr = shift(@address_bits);
+ $email = $envaddr->address();
+ $email =~ s:\@(.*):\@\L$1\E:g;
+ return $email;
+}
+
+sub debug {
+ my $text = shift;
+ my $level = 1;
+ if ($text =~ /^\d+$/) {
+ $level = $text;
+ $text = shift;
+ }
+ if ($debug >= $level) {
+ chomp($msg);
+ chomp($text);
+ $msg .= "\nDEBUG: $text\n";
+ }
+}
+
+#
+# Get the info from the database and build the reply.
+#
+sub fetchdata {
+ my ($rpatterns) = @_;
+ my @patterns = @{$rpatterns};
+ my $msg = '';
+ #
+ # Now connect to the database.
+ #
+ debug(4, 'Connecting to the database');
+ my $dsn = 'DBI:mysql:database=<dbname>:host=<dbfqhn>';
+ my $dbx = DBI->connect($dsn, '<username>', '<password>',
+ {
+ RaiseError => 0,
+ PrintError => 0
+ });
+ if (! $dbx) {
+ debug(2, "Error connecting to the database");
+ }
+ #
+ # Find out what the latest data import timestamp is.
+ #
+ debug(4, 'Determining date of most recent load');
+ my $sql = ('SELECT asof FROM loadinfo '
+ . 'ORDER BY asof DESC '
+ . 'LIMIT 1 ');
+ my $dbq = $dbx->prepare($sql);
+ if (! $dbq->execute()) {
+ debug(2, "Error fetching last load date");
+ }
+ my $loadinfo = $dbq->fetchrow_hashref();
+ $dbq->finish();
+ my $loaddate = $loadinfo->{asof};
+ debug(4, "Last loaded at '$loaddate'");
+
+ #
+ # Start constructing the reply.
+ #
+ $msg .= 'Results of your quer'
+ . (@patterns == 1 ? 'y' : 'ies')
+ . " (based on information dated $loaddate):\n";
+
+ #
+ # Now loop through and try to turn each one into a command.
+ #
+ foreach my $pattern (@patterns) {
+ debug(3, "Pattern loop: '$pattern'");
+ $pattern =~ s/^\s+(.*)/$1/;
+ next if (! $pattern);
+ last if ($pattern =~ /^--\s$/i);
+ $pattern =~ s/\s+$//;
+ my @tokens = split(/\s+/, $pattern);
+ my $verb = $tokens[0];
+ #
+ # Make a note of what we're going to try to parse as a command or RE.
+ #
+ $msg .= "\n[*] $pattern";
+ my $isverb = $verbs->{lc($verb)};
+ if (defined($isverb) && ($isverb eq 'debug')) {
+ if ((@tokens > 1) && ($tokens[1] =~ /^\d+$/)) {
+ $debug = $tokens[1];
+ }
+ else {
+ $debug++;
+ }
+ debug("Debug level now $debug");
+ next;
+ }
+ debug(2, "DSN=$dsn");
+ my @words = shellwords($pattern);
+ debug('Shellwords => ("' . join('","', @words) . '")');
+ my $field = undef;
+ #
+ # The command parser operates on @ARGV, so save a copy and
+ # replace it with our own command. Actually, a null command;
+ # turn the verb into a parsable option by prefixing it with '-'.
+ #
+ my @argv_temp = @ARGV;
+ @ARGV = @words;
+ if (defined($isverb)) {
+ $ARGV[0] = $isverb;
+ }
+ last if (defined($isverb) && ($isverb eq 'end'));
+ $ARGV[0] = '--' . $ARGV[0];
+ GetOptions('list' => sub { $field = 'list'; },
+ 'moderated-by' => sub { $field = 'moderators'; },
+ 'help' => sub { $field = 'help'; });
+ #
+ # If the first word was recognisably one of our verbs, take
+ # any leftover string as the expressions to use.
+ #
+ if ($field || defined($isverb)) {
+ @words = @ARGV;
+ }
+ else {
+ #
+ # Anything that isn't one of our verbs we treat as a regex for
+ # an implicit 'list' verb.
+ #
+ $field = 'list';
+ }
+ debug(2, '@words = "' . join('" "', @words) . '"');
+ if (defined($isverb) && ($isverb eq 'help')) {
+ $msg .= helpout(@words);
+ next;
+ }
+ #
+ # We're actually only interested in the first word. For now.
+ # Stick the actual script command line arguments back where they
+ # belong.
+ #
+ $pattern = $words[0];
+ @ARGV = @argv_temp;
+# $pattern =~ s/^\s*(.*)?\s*/$1/g if ($pattern !~ /^\s*$/);
+ #
+ # We now have a regex and a field name. Ask the database about
+ # them.
+ #
+ $sql = ('SELECT DISTINCT list FROM list '
+ . "WHERE ($field REGEXP '$pattern') "
+ . "AND (asof = '$loaddate') "
+ . 'ORDER BY domain ASC, list ASC ');
+ debug(5, 'Using SQL: ' . $sql);
+ $dbq = $dbx->prepare($sql);
+ $dbq->execute();
+ my $lists = $dbq->fetchall_arrayref();
+ $dbq->finish();
+ $msg .= "\nLists with $field matching '$pattern'";
+ if ((! $lists) || (! @{$lists})) {
+ debug(4, 'No matching lists');
+ $msg .= ":\n None.\n";
+ }
+ else {
+ $msg .= ' (' . scalar(@{$lists}) . "):\n";
+ #
+ # Okey, one or more lists actually matched. Report on all
+ # matches. First, collect all the list names together so
+ # we can loop through them.
+ #
+ debug(4, scalar(@{$lists}) . ' matching lists');
+ my @listname;
+ for (@{$lists}) {
+ push(@listname, $_->[0]);
+ }
+ #
+ # Now get the particulars.
+ #
+ foreach my $listname (@listname) {
+ $sql = ('SELECT * FROM list '
+ . "WHERE (list = '$listname') "
+ . "AND (asof = '$loaddate') ");
+ $dbq = $dbx->prepare($sql);
+ $dbq->execute();
+ my $listinfo = $dbq->fetchrow_hashref();
+ $dbq->finish();
+ my $n = $listinfo->{count};
+ my $dn = $listinfo->{dcount} || 0;
+ my $mods = $listinfo->{moderators};
+ $mods =~ s/,/, /g;
+ $msg .= " $listname\n"
+ . " Subscribers: $n\n"
+ . ' Digest readers: '
+ . ($dn < 0 ? 'no digest' : $dn) . "\n"
+ . " Moderators: $mods\n"
+ . "\n";
+ }
+ }
+ }
+ #
+ # Gone through 'em all; let us leave this place.
+ #
+ $dbx->disconnect();
+ return $msg;
+}
+
+sub helpout {
+ my (@words) = @_;
+ my $msg = '';
+ debug(4, 'helpout(<' . join('>,<', @words) . '>)');
+ if (@words > 0) {
+ my $isverb = $verbs->{lc($words[0])};
+ if ($isverb eq 'moderated-by') {
+ $msg .= "\nHELP MODERATED-BY\n"
+ . "\n moderated-by [regex]\n"
+ . "\n The first argument is treated as a regular "
+ . "expression, and"
+ . "\n all lists which have at least one moderator "
+ . "whose eddress matches"
+ . "\n the pattern will be included.\n";
+ }
+ elsif ($isverb eq 'debug') {
+ $msg .= "\nHELP DEBUG\n"
+ . "\n debug\n"
+ . "\n Increments the debug level each time it's"
+ . "\n encountered. Deep Magic, don't mess with it.\n";
+ }
+ elsif ($isverb eq 'list') {
+ $msg .= "\nHELP LIST\n"
+ . "\n list [regex]\n"
+ . "\n The first argument is treated as a regular "
+ . "expression, and"
+ . "\n all lists with names that match the pattern "
+ . "the pattern will be included.\n"
+ . "\n Note that you may omit the 'list' verb and "
+ . "just specify a regular expression.\n";
+ }
+ elsif ($isverb eq 'help') {
+ $msg .= "\nHELP HELP\n"
+ . "\n You're kidding, right?\n";
+ }
+ elsif ($isverb eq 'end') {
+ $msg .= "\nHELP END\n"
+ . "\n end\n"
+ . "\n This command signals the end of the commands"
+ . "\n in the message. All text after this will be"
+ . "\n ignored.\n";
+ }
+ }
+ else {
+ $msg .= "\nHelp for this utility:\n"
+ . "\nSend a message to the query address with one or"
+ . "\nmore of the following:\n"
+ . "\n help"
+ . "\n this message\n"
+ . "\n moderated-by [eddress-pattern]"
+ . "\n info about all lists with matching moderators\n"
+ . "\n [list] list-pattern"
+ . "\n info about all matching lists\n"
+ . "\n end"
+ . "\n End parsing of the message as commands. Text"
+ . "\n appearing after this verb is ignored.\n"
+ . "\nAnything that doesn't begin with one of these"
+ . "\nkeywords (or an abbreviation thereof) will be"
+ . "\ntreated as a Perl regex -- as though 'list' was"
+ . "\nthe command.\n";
+ }
+ return $msg;
+}
+#
+# Local Variables:
+# mode: cperl
+# tab-width: 4
+# indent-tabs-mode: nil
+# End:
+#
Propchange: labs/pulse/report-email/listex-report.pl
------------------------------------------------------------------------------
svn:executable = *
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@labs.apache.org
For additional commands, e-mail: commits-help@labs.apache.org