| 1 | #! /usr/bin/perl |
| 2 | # |
| 3 | # parsedb.pl |
| 4 | # |
| 5 | # This script will parse a database with raw header information |
| 6 | # from a INN feed to a structured database. |
| 7 | # |
| 8 | # It is part of the NewsStats package. |
| 9 | # |
| 10 | # Copyright (c) 2013 Thomas Hochstein <thh@inter.net> |
| 11 | # |
| 12 | # It can be redistributed and/or modified under the same terms under |
| 13 | # which Perl itself is published. |
| 14 | |
| 15 | BEGIN { |
| 16 | our $VERSION = "0.01"; |
| 17 | use File::Basename; |
| 18 | # we're in .../bin, so our module is in ../lib |
| 19 | push(@INC, dirname($0).'/../lib'); |
| 20 | } |
| 21 | use strict; |
| 22 | use warnings; |
| 23 | |
| 24 | use NewsStats qw(:DEFAULT :TimePeriods :SQLHelper); |
| 25 | |
| 26 | use DBI; |
| 27 | use Getopt::Long qw(GetOptions); |
| 28 | Getopt::Long::config ('bundling'); |
| 29 | |
| 30 | use Encode qw/decode/; |
| 31 | use Mail::Address; |
| 32 | |
| 33 | ################################# Definitions ################################## |
| 34 | |
| 35 | # define header names with separate database fields |
| 36 | my %DBFields = ('date' => 'date', |
| 37 | 'references' => 'refs', |
| 38 | 'followup-to' => 'fupto', |
| 39 | 'from' => 'from_', |
| 40 | 'sender' => 'sender', |
| 41 | 'reply-to' => 'replyto', |
| 42 | 'subject' => 'subject', |
| 43 | 'organization' => 'organization', |
| 44 | 'lines' => 'linecount', |
| 45 | 'approved' => 'approved', |
| 46 | 'supersedes' => 'supersedes', |
| 47 | 'expires' => 'expires', |
| 48 | 'user-agent' => 'useragent', |
| 49 | 'x-newsreader' => 'xnewsreader', |
| 50 | 'x-mailer' => 'xmailer', |
| 51 | 'x-no-archive' => 'xnoarchive', |
| 52 | 'content-type' => 'contenttype', |
| 53 | 'content-transfer-encoding' => 'contentencoding', |
| 54 | 'cancel-lock' => 'cancellock', |
| 55 | 'injection-info' => 'injectioninfo', |
| 56 | 'x-trace' => 'xtrace', |
| 57 | 'nntp-posting-host' => 'postinghost'); |
| 58 | |
| 59 | # define field list for database |
| 60 | my @DBFields = qw/day mid refs date path newsgroups fupto from_ from_parsed |
| 61 | from_name from_address sender sender_parsed sender_name |
| 62 | sender_address replyto replyto_parsed replyto_name |
| 63 | replyto_address subject subject_parsed organization linecount |
| 64 | approved supersedes expires useragent xnewsreader xmailer |
| 65 | xnoarchive contenttype contentencoding cancellock injectioninfo |
| 66 | xtrace postinghost headers disregard/; |
| 67 | |
| 68 | ################################# Main program ################################# |
| 69 | |
| 70 | ### read commandline options |
| 71 | my ($OptDay,$OptDebug,$OptParseDB,$OptRawDB,$OptTest,$OptConfFile); |
| 72 | GetOptions ('d|day=s' => \$OptDay, |
| 73 | 'debug!' => \$OptDebug, |
| 74 | 'parsedb=s' => \$OptParseDB, |
| 75 | 'rawdb=s' => \$OptRawDB, |
| 76 | 't|test!' => \$OptTest, |
| 77 | 'conffile=s' => \$OptConfFile, |
| 78 | 'h|help' => \&ShowPOD, |
| 79 | 'V|version' => \&ShowVersion) or exit 1; |
| 80 | |
| 81 | ### read configuration |
| 82 | my %Conf = %{ReadConfig($OptConfFile)}; |
| 83 | |
| 84 | ### override configuration via commandline options |
| 85 | my %ConfOverride; |
| 86 | $ConfOverride{'DBTableRaw'} = $OptRawDB if $OptRawDB; |
| 87 | $ConfOverride{'DBTableParse'} = $OptParseDB if $OptParseDB; |
| 88 | &OverrideConfig(\%Conf,\%ConfOverride); |
| 89 | |
| 90 | ### get time period |
| 91 | ### and set $Period for output and expression for SQL 'WHERE' clause |
| 92 | my ($Period,$SQLWherePeriod) = &GetTimePeriod($OptDay,'day'); |
| 93 | # bail out if --month is invalid or "all" |
| 94 | &Bleat(2,"--day option has an invalid format - please use 'YYYY-MM-DD' or ". |
| 95 | "'YYYY-MM-DD:YYYY-MM-DD'!") if (!$Period or $Period eq 'all time'); |
| 96 | |
| 97 | ### init database |
| 98 | my $DBHandle = InitDB(\%Conf,1); |
| 99 | |
| 100 | ### get & write data |
| 101 | &Bleat(1,'Test mode. Database is not updated.') if $OptTest; |
| 102 | |
| 103 | # create $SQLWhereClause |
| 104 | my $SQLWhereClause = SQLBuildClause('where',$SQLWherePeriod,'NOT disregard'); |
| 105 | |
| 106 | # delete old data for current period |
| 107 | if (!$OptTest) { |
| 108 | print "----------- Deleting old data ... -----------\n" if $OptDebug; |
| 109 | my $DBQuery = $DBHandle->do(sprintf("DELETE FROM %s.%s %s", |
| 110 | $Conf{'DBDatabase'},$Conf{'DBTableParse'}, |
| 111 | $SQLWhereClause)) |
| 112 | or &Bleat(2,sprintf("Can't delete old parsed data for %s from %s.%s: ". |
| 113 | "$DBI::errstr\n",$Period, |
| 114 | $Conf{'DBDatabase'},$Conf{'DBTableParse'})); |
| 115 | }; |
| 116 | |
| 117 | # read from DBTableRaw |
| 118 | print "-------------- Reading data ... -------------\n" if $OptDebug; |
| 119 | my $DBQuery = $DBHandle->prepare(sprintf("SELECT id, day, mid, peer, path, ". |
| 120 | "newsgroups, headers, disregard ". |
| 121 | "FROM %s.%s %s", $Conf{'DBDatabase'}, |
| 122 | $Conf{'DBTableRaw'}, $SQLWhereClause)); |
| 123 | $DBQuery->execute() |
| 124 | or &Bleat(2,sprintf("Can't get data for %s from %s.%s: ". |
| 125 | "$DBI::errstr\n",$Period, |
| 126 | $Conf{'DBDatabase'},$Conf{'DBTableRaw'})); |
| 127 | |
| 128 | # set output and database connection to UTF-8 |
| 129 | # as we're going to write decoded header contents containing UTF-8 chars |
| 130 | binmode(STDOUT, ":utf8"); |
| 131 | $DBHandle->do("SET NAMES 'utf8'"); |
| 132 | |
| 133 | # create a list of supported encondings |
| 134 | my %LegalEncodings; |
| 135 | foreach (Encode->encodings()) { |
| 136 | $LegalEncodings{$_} = 1; |
| 137 | } |
| 138 | # parse data in a loop and write it out |
| 139 | print "-------------- Parsing data ... -------------\n" if $OptDebug; |
| 140 | while (my $HeadersR = $DBQuery->fetchrow_hashref) { |
| 141 | my %Headers = %{$HeadersR}; |
| 142 | |
| 143 | # parse $Headers{'headers'} ('headers' from DBTableRaw) |
| 144 | # remove empty lines (that should not even exist in a header!) |
| 145 | $Headers{'headers'} =~ s/\n\s*\n/\n/g; |
| 146 | # merge continuation lines |
| 147 | # from Perl Cookbook, 1st German ed. 1999, pg. 91 |
| 148 | $Headers{'headers'} =~ s/\n\s+/ /g; |
| 149 | # split headers in single lines |
| 150 | my $OtherHeaders; |
| 151 | for (split(/\n/,$Headers{'headers'})) { |
| 152 | # split header lines in header name and header content |
| 153 | my ($key,$value); |
| 154 | if ($_ =~ /:/) { |
| 155 | ($key,$value) = split(/:/,$_,2); |
| 156 | $key =~ s/\s*//; |
| 157 | $value =~ s/^\s*(.+)\s*$/$1/; |
| 158 | } else { |
| 159 | &Bleat(1,sprintf("Illegal header line in %s.%s id %s: %s", |
| 160 | $Conf{'DBDatabase'}, $Conf{'DBTableRaw'}, |
| 161 | $Headers{'id'},$_)); |
| 162 | next; |
| 163 | } |
| 164 | # check for empty (mandatory) fields from DBTableRaw |
| 165 | # and set them from $Headers{'headers', if necessary |
| 166 | if (lc($key) =~ /^(message-id|path|newsgroups)$/) { |
| 167 | my $HeaderName = lc($key); |
| 168 | $HeaderName = 'mid' if ($HeaderName eq 'message-id'); |
| 169 | if (!defined($Headers{$HeaderName}) or $Headers{$HeaderName} eq '') { |
| 170 | $Headers{$HeaderName} = $value; |
| 171 | &Bleat(1,sprintf("Taking missing %s from 'headers' in %s.%s id %s.", |
| 172 | $HeaderName, $Conf{'DBDatabase'}, $Conf{'DBTableRaw'}, |
| 173 | $Headers{'id'})); |
| 174 | } |
| 175 | } |
| 176 | # save each header, separate database fields in %Headers, |
| 177 | # the rest in $OtherHeaders (but not Message-ID, Path, Peer |
| 178 | # and Newsgroups as those do already exist) |
| 179 | if (defined($DBFields{lc($key)})) { |
| 180 | $Headers{$DBFields{lc($key)}} = $value; |
| 181 | } else { |
| 182 | $OtherHeaders .= sprintf("%s: %s\n",$key,$value) |
| 183 | if lc($key) !~ /^(message-id|path|peer|newsgroups)$/; |
| 184 | } |
| 185 | } |
| 186 | # replace old (now parsed) $Headers{'headers'} with remanining $OtherHeaders |
| 187 | chomp($OtherHeaders); |
| 188 | $Headers{'headers'} = $OtherHeaders; |
| 189 | |
| 190 | foreach ('from_','sender', 'replyto', 'subject') { |
| 191 | if ($Headers{$_}) { |
| 192 | my $HeaderName = $_; |
| 193 | $HeaderName =~ s/_$//; |
| 194 | # decode From: / Sender: / Reply-To: / Subject: |
| 195 | if ($Headers{$_} =~ /\?(B|Q)\?/) { |
| 196 | # check for legal encoding and decode |
| 197 | (my $Encoding) = $Headers{$_} =~ /\?([^?]+)\?(B|Q)\?/; |
| 198 | $Headers{$HeaderName.'_parsed'} = decode('MIME-Header',$Headers{$_}) |
| 199 | if (exists($LegalEncodings{$Encoding})); |
| 200 | } |
| 201 | # extract name(s) and mail(s) from From: / Sender: / Reply-To: |
| 202 | # in parsed form, if available |
| 203 | if ($_ ne 'subject') { |
| 204 | my @Address; |
| 205 | # start parser on header or parsed header |
| 206 | # @Address will have an array of Mail::Address objects, one for |
| 207 | # each name/mail (you can have more than one person in From:!) |
| 208 | if (defined($Headers{$HeaderName.'_parsed'})) { |
| 209 | @Address = Mail::Address->parse($Headers{$HeaderName.'_parsed'}); |
| 210 | } else { |
| 211 | @Address = Mail::Address->parse($Headers{$_}); |
| 212 | } |
| 213 | # split each Mail::Address object to @Names and @Adresses |
| 214 | my (@Names,@Adresses); |
| 215 | foreach (@Address) { |
| 216 | # take address part in @Addresses |
| 217 | push (@Adresses, $_->address()); |
| 218 | # take name part form "phrase", if there is one: |
| 219 | # From: My Name <addr@ess> (Comment) |
| 220 | # otherwise, take it from "comment": |
| 221 | # From: addr@ess (Comment) |
| 222 | # and push it in @Names |
| 223 | my ($Name); |
| 224 | $Name = $_->comment() unless $Name = $_->phrase; |
| 225 | $Name =~ s/^\((.+)\)$/$1/; |
| 226 | push (@Names, $Name); |
| 227 | } |
| 228 | # put all @Adresses and all @Names in %Headers as comma separated lists |
| 229 | $Headers{$HeaderName.'_address'} = join(', ',@Adresses); |
| 230 | $Headers{$HeaderName.'_name'} = join(', ',@Names); |
| 231 | } |
| 232 | } |
| 233 | } |
| 234 | |
| 235 | # order output for database entry: fill @SQLBindVars |
| 236 | print "-------------- Next entry:\n" if $OptDebug; |
| 237 | my @SQLBindVars; |
| 238 | foreach (@DBFields) { |
| 239 | if (defined($Headers{$_}) and $Headers{$_} ne '') { |
| 240 | push (@SQLBindVars,$Headers{$_}); |
| 241 | printf ("FOUND: %s -> %s\n",$_,$Headers{$_}) if $OptDebug; |
| 242 | } else { |
| 243 | push (@SQLBindVars,undef); |
| 244 | } |
| 245 | } |
| 246 | |
| 247 | # write data to DBTableParse |
| 248 | if (!$OptTest) { |
| 249 | print "-------------- Writing data ... -------------\n" if $OptDebug; |
| 250 | my $DBWrite = |
| 251 | $DBHandle->prepare(sprintf("INSERT INTO %s.%s (%s) VALUES (%s)", |
| 252 | $Conf{'DBDatabase'}, |
| 253 | $Conf{'DBTableParse'}, |
| 254 | # get field names from @DBFields |
| 255 | join(', ',@DBFields), |
| 256 | # create a list of '?' for each DBField |
| 257 | join(', ', |
| 258 | split(/ /,'? ' x scalar(@DBFields))) |
| 259 | )); |
| 260 | $DBWrite->execute(@SQLBindVars) |
| 261 | or &Bleat(2,sprintf("Can't write parsed data for %s to %s.%s: ". |
| 262 | "$DBI::errstr\n",$Period, |
| 263 | $Conf{'DBDatabase'},$Conf{'DBTableParse'})); |
| 264 | $DBWrite->finish; |
| 265 | } |
| 266 | }; |
| 267 | $DBQuery->finish; |
| 268 | |
| 269 | ### close handles |
| 270 | $DBHandle->disconnect; |
| 271 | |
| 272 | print "------------------- DONE! -------------------\n" if $OptDebug; |
| 273 | __END__ |
| 274 | |
| 275 | ################################ Documentation ################################# |
| 276 | |
| 277 | =head1 NAME |
| 278 | |
| 279 | parsedb - parse raw data and save it to a database |
| 280 | |
| 281 | =head1 SYNOPSIS |
| 282 | |
| 283 | B<parsedb> [B<-Vht>] [B<--day> I<YYYY-MM-DD> | I<YYYY-MM-DD:YYYY-MM-DD>] [B<--rawdb> I<database table>] [B<--parsedb> I<database table>] [B<--conffile> I<filename>] [B<--debug>] |
| 284 | |
| 285 | =head1 REQUIREMENTS |
| 286 | |
| 287 | See L<doc/README>. |
| 288 | |
| 289 | =head1 DESCRIPTION |
| 290 | |
| 291 | This script will parse raw, unstructured headers from a database table which is |
| 292 | fed from F<feedlog.pl> for a given time period and write its results to |
| 293 | nother database table with separate fields (columns) for most (or even all) |
| 294 | relevant headers. |
| 295 | |
| 296 | I<Subject:>, I<From:>, I<Sender:> and I<Reply-To:> will be parsed from MIME |
| 297 | encoded words to UTF-8 as needed while the unparsed copy is kept. From that |
| 298 | parsed copy, I<From:>, I<Sender:> and I<Reply-To:> will also be split into |
| 299 | separate name(s) and address(es) fields while the un-splitted copy is kept, |
| 300 | too. |
| 301 | |
| 302 | B<parsedb> should be run nightly from cron for yesterdays data so all |
| 303 | other scripts get current information. The time period to act on defaults to |
| 304 | yesterday, accordingly; you can assign another time period or a single day via |
| 305 | the B<--day> option (see below). |
| 306 | |
| 307 | =head2 Configuration |
| 308 | |
| 309 | B<parsedb> will read its configuration from F<newsstats.conf> |
| 310 | should be present in etc/ via Config::Auto or from a configuration file |
| 311 | submitted by the B<--conffile> option. |
| 312 | |
| 313 | See L<doc/INSTALL> for an overview of possible configuration options. |
| 314 | |
| 315 | You can override configuration options via the B<--rawdb> and |
| 316 | B<--parsedb> options, respectively. |
| 317 | |
| 318 | =head1 OPTIONS |
| 319 | |
| 320 | =over 3 |
| 321 | |
| 322 | =item B<-V>, B<--version> |
| 323 | |
| 324 | Print out version and copyright information and exit. |
| 325 | |
| 326 | =item B<-h>, B<--help> |
| 327 | |
| 328 | Print this man page and exit. |
| 329 | |
| 330 | =item B<--debug> |
| 331 | |
| 332 | Output (rather much) debugging information to STDOUT while processing. |
| 333 | |
| 334 | =item B<-t>, B<--test> |
| 335 | |
| 336 | Do not write results to database. You should use B<--debug> in |
| 337 | conjunction with B<--test> ... everything else seems a bit pointless. |
| 338 | |
| 339 | =item B<-d>, B<--day> I<YYYY-MM-DD[:YYYY-MM-DD]> |
| 340 | |
| 341 | Set processing period to a single day in YYYY-MM-DD format or to a time |
| 342 | period between two days in YYYY-MM-DD:YYYY-MM-DD format (two days, separated |
| 343 | by a colon). |
| 344 | |
| 345 | Defaults to yesterday. |
| 346 | |
| 347 | =item B<--rawdb> I<table> (raw data table) |
| 348 | |
| 349 | Override I<DBTableRaw> from F<newsstats.conf>. |
| 350 | |
| 351 | =item B<--parsedb> I<table> (parsed data table) |
| 352 | |
| 353 | Override I<DBTableParse> from F<newsstats.conf>. |
| 354 | |
| 355 | =item B<--conffile> I<filename> |
| 356 | |
| 357 | Load configuration from I<filename> instead of F<newsstats.conf>. |
| 358 | |
| 359 | =back |
| 360 | |
| 361 | =head1 INSTALLATION |
| 362 | |
| 363 | See L<doc/INSTALL>. |
| 364 | |
| 365 | =head1 EXAMPLES |
| 366 | |
| 367 | An example crontab entry: |
| 368 | |
| 369 | 0 1 * * * /path/to/bin/parsedb.pl |
| 370 | |
| 371 | Do a dry run for yesterday's data, showing results of processing: |
| 372 | |
| 373 | parsedb --debug --test | less |
| 374 | |
| 375 | =head1 FILES |
| 376 | |
| 377 | =over 4 |
| 378 | |
| 379 | =item F<bin/parsedb.pl> |
| 380 | |
| 381 | The script itself. |
| 382 | |
| 383 | =item F<lib/NewsStats.pm> |
| 384 | |
| 385 | Library functions for the NewsStats package. |
| 386 | |
| 387 | =item F<etc/newsstats.conf> |
| 388 | |
| 389 | Runtime configuration file. |
| 390 | |
| 391 | =back |
| 392 | |
| 393 | =head1 BUGS |
| 394 | |
| 395 | Please report any bugs or feature requests to the author or use the |
| 396 | bug tracker at L<http://bugs.th-h.de/>! |
| 397 | |
| 398 | =head1 SEE ALSO |
| 399 | |
| 400 | =over 2 |
| 401 | |
| 402 | =item - |
| 403 | |
| 404 | L<doc/README> |
| 405 | |
| 406 | =item - |
| 407 | |
| 408 | L<doc/INSTALL> |
| 409 | |
| 410 | =back |
| 411 | |
| 412 | This script is part of the B<NewsStats> package. |
| 413 | |
| 414 | =head1 AUTHOR |
| 415 | |
| 416 | Thomas Hochstein <thh@inter.net> |
| 417 | |
| 418 | =head1 COPYRIGHT AND LICENSE |
| 419 | |
| 420 | Copyright (c) 2013 Thomas Hochstein <thh@inter.net> |
| 421 | |
| 422 | This program is free software; you may redistribute it and/or modify it |
| 423 | under the same terms as Perl itself. |
| 424 | |
| 425 | =cut |