| 1 | ###########################################################################
|
|---|
| 2 | #
|
|---|
| 3 | # DBPlugin.pm -- plugin to import records from a database
|
|---|
| 4 | #
|
|---|
| 5 | # A component of the Greenstone digital library software
|
|---|
| 6 | # from the New Zealand Digital Library Project at the
|
|---|
| 7 | # University of Waikato, New Zealand.
|
|---|
| 8 | #
|
|---|
| 9 | # Copyright (C) 2003 New Zealand Digital Library Project
|
|---|
| 10 | #
|
|---|
| 11 | # This program is free software; you can redistribute it and/or modify
|
|---|
| 12 | # it under the terms of the GNU General Public License as published by
|
|---|
| 13 | # the Free Software Foundation; either version 2 of the License, or
|
|---|
| 14 | # (at your option) any later version.
|
|---|
| 15 | #
|
|---|
| 16 | # This program is distributed in the hope that it will be useful,
|
|---|
| 17 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|---|
| 18 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|---|
| 19 | # GNU General Public License for more details.
|
|---|
| 20 | #
|
|---|
| 21 | # You should have received a copy of the GNU General Public License
|
|---|
| 22 | # along with this program; if not, write to the Free Software
|
|---|
| 23 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
|---|
| 24 | #
|
|---|
| 25 | ###########################################################################
|
|---|
| 26 |
|
|---|
| 27 | # A plugin that imports records from a database. This uses perl's DBI module,
|
|---|
| 28 | # which includes back-ends for mysql, postgresql, comma separated values (CSV),
|
|---|
| 29 | # MS Excel, ODBC, sybase, etc... Extra modules may need to be installed to
|
|---|
| 30 | # use this. See <GSDLHOME>/etc/packages/example.dbi for an example config file.
|
|---|
| 31 | #
|
|---|
| 32 |
|
|---|
| 33 | # Written by John McPherson for the NZDL project
|
|---|
| 34 | # Mar, Apr 2003
|
|---|
| 35 |
|
|---|
| 36 | package DBPlugin;
|
|---|
| 37 |
|
|---|
| 38 | use strict;
|
|---|
| 39 | no strict 'refs'; # allow variable as a filehandle
|
|---|
| 40 |
|
|---|
| 41 | use AutoExtractMetadata;
|
|---|
| 42 | use unicode;
|
|---|
| 43 |
|
|---|
| 44 | sub BEGIN {
|
|---|
| 45 | @DBPlugin::ISA = ('AutoExtractMetadata');
|
|---|
| 46 | }
|
|---|
| 47 |
|
|---|
| 48 | my $arguments =
|
|---|
| 49 | [ { 'name' => "process_exp",
|
|---|
| 50 | 'desc' => "{BasePlugin.process_exp}",
|
|---|
| 51 | 'type' => "regexp",
|
|---|
| 52 | 'deft' => &get_default_process_exp(),
|
|---|
| 53 | 'reqd' => "no" }];
|
|---|
| 54 |
|
|---|
| 55 | my $options = { 'name' => "DBPlugin",
|
|---|
| 56 | 'desc' => "{DBPlugin.desc}",
|
|---|
| 57 | 'abstract' => "no",
|
|---|
| 58 | 'inherits' => "yes",
|
|---|
| 59 | 'args' => $arguments };
|
|---|
| 60 |
|
|---|
| 61 | sub new {
|
|---|
| 62 | my ($class) = shift (@_);
|
|---|
| 63 | my ($pluginlist,$inputargs,$hashArgOptLists) = @_;
|
|---|
| 64 | push(@$pluginlist, $class);
|
|---|
| 65 |
|
|---|
| 66 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
|---|
| 67 | push(@{$hashArgOptLists->{"OptList"}},$options);
|
|---|
| 68 |
|
|---|
| 69 | my $self = new AutoExtractMetadata($pluginlist, $inputargs, $hashArgOptLists);
|
|---|
| 70 |
|
|---|
| 71 | return bless $self, $class;
|
|---|
| 72 | }
|
|---|
| 73 |
|
|---|
| 74 | sub get_default_process_exp {
|
|---|
| 75 | my $self = shift (@_);
|
|---|
| 76 |
|
|---|
| 77 | return q^(?i)\.dbi$^;
|
|---|
| 78 | }
|
|---|
| 79 |
|
|---|
| 80 | sub read {
|
|---|
| 81 | my $self = shift (@_);
|
|---|
| 82 | my ($pluginfo, $base_dir, $file, $block_hash, $metadata, $processor, $maxdocs,$total_count,$gli) = @_;
|
|---|
| 83 |
|
|---|
| 84 | #see if we can handle the passed file...
|
|---|
| 85 | my ($filename_full_path, $filename_no_path) = &util::get_full_filenames($base_dir, $file);
|
|---|
| 86 | return undef unless $self->can_process_this_file($filename_full_path);
|
|---|
| 87 |
|
|---|
| 88 | my $outhandle = $self->{'outhandle'};
|
|---|
| 89 | my $verbosity = $self->{'verbosity'};
|
|---|
| 90 |
|
|---|
| 91 | print STDERR "<Processing n='$file' p='DBPlugin'>\n" if ($gli);
|
|---|
| 92 | print $outhandle "DBPlugin: processing $file\n"
|
|---|
| 93 | if $self->{'verbosity'} > 1;
|
|---|
| 94 |
|
|---|
| 95 | require DBI; # database independent stuff
|
|---|
| 96 |
|
|---|
| 97 | # calculate the document hash, for document ids
|
|---|
| 98 | my $hash="0";
|
|---|
| 99 |
|
|---|
| 100 |
|
|---|
| 101 | # default options - may be overridden by config file
|
|---|
| 102 | my $language=undef;
|
|---|
| 103 | my $encoding=undef;
|
|---|
| 104 | my $dbplug_debug=0;
|
|---|
| 105 | my $username='';
|
|---|
| 106 | my $password='';
|
|---|
| 107 |
|
|---|
| 108 | # these settings must be set by the config file:
|
|---|
| 109 | my $db=undef;
|
|---|
| 110 |
|
|---|
| 111 | # get id of pages from "nonempty", get latest version number from
|
|---|
| 112 | # "recent", and then get pagename from "page" and content from "version" !
|
|---|
| 113 |
|
|---|
| 114 | my $sql_query_prime = undef ;
|
|---|
| 115 | my $sql_query = undef ;
|
|---|
| 116 |
|
|---|
| 117 | my %db_to_greenstone_fields=();
|
|---|
| 118 | my %callbacks=();
|
|---|
| 119 |
|
|---|
| 120 |
|
|---|
| 121 | # read in config file.
|
|---|
| 122 | if (!open (CONF, $filename_full_path)) {
|
|---|
| 123 | print $outhandle "DBPlugin: can't read $filename_full_path: $!\n";
|
|---|
| 124 | return 0;
|
|---|
| 125 | }
|
|---|
| 126 |
|
|---|
| 127 | my $line;
|
|---|
| 128 | my $statement="";
|
|---|
| 129 | my $callback="";
|
|---|
| 130 | while (defined($line=<CONF>)) {
|
|---|
| 131 | chomp $line;
|
|---|
| 132 | $line .= " "; # for multi-line statements - don't conjoin!
|
|---|
| 133 | $line =~ s/\s*\#.*$//mg; # remove comments
|
|---|
| 134 | $statement .= $line;
|
|---|
| 135 |
|
|---|
| 136 | if ($line =~ /^\}\s*$/ && $callback) { # ends the callback
|
|---|
| 137 | $callback .= $statement ; $statement = "";
|
|---|
| 138 | # try to check that the function is "safe"
|
|---|
| 139 | if ($callback =~ /\b(?:system|open|pipe|readpipe|qx|kill|eval|do|use|require|exec|fork)\b/ ||
|
|---|
| 140 | $callback =~ /[\`]|\|\-/) {
|
|---|
| 141 | # no backticks or functions that start new processes allowed
|
|---|
| 142 | print $outhandle "DBPlugin: bad function in callback\n";
|
|---|
| 143 | return 0;
|
|---|
| 144 | }
|
|---|
| 145 | $callback =~ s/sub (\w+?)_callback/sub/;
|
|---|
| 146 | my $fieldname = $1;
|
|---|
| 147 | my $ret = eval "\$callbacks{'$fieldname'} = $callback ; 1";
|
|---|
| 148 | if (!defined($ret)) {
|
|---|
| 149 | print $outhandle "DBPlugin: error eval'ing callback: $@\n";
|
|---|
| 150 | exit(1);
|
|---|
| 151 | }
|
|---|
| 152 | $callback="";
|
|---|
| 153 | print $outhandle "DBPlugin: callback registered for '$fieldname'\n"
|
|---|
| 154 | if $dbplug_debug;
|
|---|
| 155 | } elsif ($callback) {
|
|---|
| 156 | # add this line to the callback function
|
|---|
| 157 | $callback .= $statement;
|
|---|
| 158 | $statement = "";
|
|---|
| 159 | } elsif ($statement =~ m/;\s*$/) { # ends with ";"
|
|---|
| 160 | # check that it is safe
|
|---|
| 161 | # assignment
|
|---|
| 162 | if ($statement =~ m~(\$\w+)\s* = \s*
|
|---|
| 163 | (\d # digits
|
|---|
| 164 | | ".*?(?<!\\)" # " up to the next " not preceded by a \
|
|---|
| 165 | | '.*?(?<!\\)' # ' up to the next ' not preceded by a \
|
|---|
| 166 | )\s*;~x || # /x means ignore comments and whitespace in rx
|
|---|
| 167 | $statement =~ m~(\%\w+)\s*=\s*(\([\w\s\"\',:=>]+\))\s*;~ ) {
|
|---|
| 168 | # evaluate the assignment, return 1 on success "
|
|---|
| 169 | if (!eval "$1=$2; 1") {
|
|---|
| 170 | my $err=$@;
|
|---|
| 171 | chomp $err;
|
|---|
| 172 | $err =~ s/\.$//; # remove a trailing .
|
|---|
| 173 | print $outhandle "DBPlugin: error evaluating `$statement'\n";
|
|---|
| 174 | print $outhandle " $err (in $filename_full_path)\n";
|
|---|
| 175 | return 0; # there was an error reading the config file
|
|---|
| 176 | }
|
|---|
| 177 | } elsif ($statement =~ /sub \w+_callback/) {
|
|---|
| 178 | # this is the start of a callback function definition
|
|---|
| 179 | $callback = $statement;
|
|---|
| 180 | $statement = "";
|
|---|
| 181 | } else {
|
|---|
| 182 | print $outhandle "DBPlugin: skipping statement `$statement'\n";
|
|---|
| 183 | }
|
|---|
| 184 | $statement = "";
|
|---|
| 185 | }
|
|---|
| 186 | }
|
|---|
| 187 | close CONF;
|
|---|
| 188 |
|
|---|
| 189 |
|
|---|
| 190 | if (!defined($db)) {
|
|---|
| 191 | print $outhandle "DBPlugin: error: $filename_full_path does not specify a db!\n";
|
|---|
| 192 | return 0;
|
|---|
| 193 | }
|
|---|
| 194 | if (!defined($sql_query)) {
|
|---|
| 195 | print $outhandle "DBPlugin: error: no SQL query specified!\n";
|
|---|
| 196 | return 0;
|
|---|
| 197 | }
|
|---|
| 198 | # connect to database
|
|---|
| 199 | my $dbhandle=DBI->connect($db, $username, $password);
|
|---|
| 200 |
|
|---|
| 201 | if (!defined($dbhandle)) {
|
|---|
| 202 | die "DBPlugin: could not connect to database, exiting.\n";
|
|---|
| 203 | }
|
|---|
| 204 | if (defined($dbplug_debug) && $dbplug_debug==1) {
|
|---|
| 205 | print $outhandle "DBPlugin (debug): connected ok\n";
|
|---|
| 206 | }
|
|---|
| 207 |
|
|---|
| 208 | my $statement_hand;
|
|---|
| 209 |
|
|---|
| 210 | # The user gave 2 sql statements to execute?
|
|---|
| 211 | if ($sql_query_prime) {
|
|---|
| 212 | $statement_hand=$dbhandle->prepare($sql_query_prime);
|
|---|
| 213 | $statement_hand->execute;
|
|---|
| 214 | if ($statement_hand->err) {
|
|---|
| 215 | print $outhandle "Error: " . $statement_hand->errstr . "\n";
|
|---|
| 216 | return undef;
|
|---|
| 217 | }
|
|---|
| 218 | }
|
|---|
| 219 |
|
|---|
| 220 |
|
|---|
| 221 | $statement_hand=$dbhandle->prepare($sql_query);
|
|---|
| 222 | $statement_hand->execute;
|
|---|
| 223 | if ($statement_hand->err) {
|
|---|
| 224 | print $outhandle "Error:" . $statement_hand->errstr . "\n";
|
|---|
| 225 | return undef;
|
|---|
| 226 | }
|
|---|
| 227 |
|
|---|
| 228 | # get the array-ref for the field names and cast it to array
|
|---|
| 229 | my @field_names;
|
|---|
| 230 | @field_names=@{ $statement_hand->{NAME} };
|
|---|
| 231 |
|
|---|
| 232 | foreach my $fieldname (@field_names) {
|
|---|
| 233 | if (defined($db_to_greenstone_fields{$fieldname})) {
|
|---|
| 234 | if (defined($dbplug_debug) && $dbplug_debug==1) {
|
|---|
| 235 | print $outhandle "DBPlugin (debug): mapping db field "
|
|---|
| 236 | . "'$fieldname' to "
|
|---|
| 237 | . $db_to_greenstone_fields{$fieldname} . "\n";
|
|---|
| 238 | }
|
|---|
| 239 | $fieldname=$db_to_greenstone_fields{$fieldname};
|
|---|
| 240 | }
|
|---|
| 241 | }
|
|---|
| 242 |
|
|---|
| 243 | # get rows
|
|---|
| 244 |
|
|---|
| 245 | my $count = 0;
|
|---|
| 246 | my @row_array;
|
|---|
| 247 |
|
|---|
| 248 | @row_array=$statement_hand->fetchrow_array; # fetchrow_hashref?
|
|---|
| 249 |
|
|---|
| 250 | while (scalar(@row_array)) {
|
|---|
| 251 | if (defined($dbplug_debug) && $dbplug_debug==1) {
|
|---|
| 252 | print $outhandle "DBPlugin (debug): retrieved a row from query\n";
|
|---|
| 253 | }
|
|---|
| 254 |
|
|---|
| 255 | # create a new document
|
|---|
| 256 | my $doc_obj = new doc ($filename_full_path, "indexed_doc");
|
|---|
| 257 | $doc_obj->set_OIDtype ($processor->{'OIDtype'}, $processor->{'OIDmetadata'});
|
|---|
| 258 | my $cursection = $doc_obj->get_top_section();
|
|---|
| 259 |
|
|---|
| 260 | # if $language not set in config file, will use BasePlugin's default
|
|---|
| 261 | if (defined($language)) {
|
|---|
| 262 | $doc_obj->add_utf8_metadata($cursection, "Language", $language);
|
|---|
| 263 | }
|
|---|
| 264 | # if $encoding not set in config file, will use BasePlugin's default
|
|---|
| 265 | if (defined($encoding)) {
|
|---|
| 266 | # allow some common aliases
|
|---|
| 267 | if ($encoding =~ m/^utf[-_]8$/i) {$encoding="utf8"}
|
|---|
| 268 | $encoding =~ s/-/_/g; # greenstone uses eg iso_8859_1
|
|---|
| 269 | $doc_obj->add_utf8_metadata($cursection, "Encoding", $encoding);
|
|---|
| 270 | }
|
|---|
| 271 | $self->set_Source_metadata($doc_obj, $db, $encoding);
|
|---|
| 272 |
|
|---|
| 273 | if ($self->{'cover_image'}) {
|
|---|
| 274 | $self->associate_cover_image($doc_obj, $filename_full_path);
|
|---|
| 275 | }
|
|---|
| 276 | $doc_obj->add_utf8_metadata($doc_obj->get_top_section(), "Plugin", "$self->{'plugin_type'}");
|
|---|
| 277 |
|
|---|
| 278 | $doc_obj->add_metadata($doc_obj->get_top_section(), "FileFormat", "DB");
|
|---|
| 279 |
|
|---|
| 280 | # include any metadata passed in from previous plugins
|
|---|
| 281 | # note that this metadata is associated with the top level section
|
|---|
| 282 | $self->extra_metadata ($doc_obj, $cursection,
|
|---|
| 283 | $metadata);
|
|---|
| 284 |
|
|---|
| 285 | # do any automatic metadata extraction
|
|---|
| 286 | $self->auto_extract_metadata ($doc_obj);
|
|---|
| 287 |
|
|---|
| 288 | my $unique_id=undef;
|
|---|
| 289 |
|
|---|
| 290 | foreach my $fieldname (@field_names) {
|
|---|
| 291 | my $fielddata=shift @row_array;
|
|---|
| 292 |
|
|---|
| 293 | if (! defined($fielddata) ) {
|
|---|
| 294 | next; # this field was "" or NULL
|
|---|
| 295 | }
|
|---|
| 296 | # use the specified encoding, defaulting to utf-8
|
|---|
| 297 | if (defined($encoding) && $encoding ne "ascii"
|
|---|
| 298 | && $encoding ne "utf8") {
|
|---|
| 299 | $fielddata=&unicode::unicode2utf8(
|
|---|
| 300 | &unicode::convert2unicode($encoding, \$fielddata)
|
|---|
| 301 | );
|
|---|
| 302 | }
|
|---|
| 303 | # see if we have a ****_callback() function defined
|
|---|
| 304 | if (exists $callbacks{$fieldname}) {
|
|---|
| 305 | my $funcptr = $callbacks{$fieldname};
|
|---|
| 306 | $fielddata = &$funcptr($fielddata);
|
|---|
| 307 | }
|
|---|
| 308 |
|
|---|
| 309 | if ($fieldname eq "text") {
|
|---|
| 310 | # add as document text
|
|---|
| 311 | $fielddata=~s@<@<@g;
|
|---|
| 312 | $fielddata=~s@>@>@g; # for xml protection...
|
|---|
| 313 | $fielddata=~s@_@\\_@g; # for macro language protection...
|
|---|
| 314 | $doc_obj->add_utf8_text($cursection, $fielddata);
|
|---|
| 315 | } elsif ($fieldname eq "Identifier") {
|
|---|
| 316 | # use as greenstone's unique record id
|
|---|
| 317 | if ($fielddata =~ /^\d+$/) {
|
|---|
| 318 | # don't allow IDs that are completely numeric
|
|---|
| 319 | $unique_id="id" . $fielddata;
|
|---|
| 320 | } else {
|
|---|
| 321 | $unique_id=$fielddata;
|
|---|
| 322 | }
|
|---|
| 323 | } else {
|
|---|
| 324 | # add as document metadata
|
|---|
| 325 | $fielddata=~s/\[/[/g;
|
|---|
| 326 | $fielddata=~s/\]/]/g;
|
|---|
| 327 | $doc_obj->add_utf8_metadata($cursection,
|
|---|
| 328 | $fieldname, $fielddata);
|
|---|
| 329 |
|
|---|
| 330 | }
|
|---|
| 331 | }
|
|---|
| 332 |
|
|---|
| 333 |
|
|---|
| 334 | if (!defined $unique_id) {
|
|---|
| 335 | $doc_obj->set_OID();
|
|---|
| 336 | my $id = $doc_obj->get_OID();
|
|---|
| 337 | $doc_obj->set_OID($id."s$count");
|
|---|
| 338 | } else {
|
|---|
| 339 | # use our id from the database...
|
|---|
| 340 | $doc_obj->set_OID($unique_id);
|
|---|
| 341 | }
|
|---|
| 342 |
|
|---|
| 343 |
|
|---|
| 344 | # process the document
|
|---|
| 345 | $processor->process($doc_obj);
|
|---|
| 346 |
|
|---|
| 347 | $count++;
|
|---|
| 348 |
|
|---|
| 349 | # get next row
|
|---|
| 350 | @row_array=$statement_hand->fetchrow_array; # fetchrow_hashref?
|
|---|
| 351 | } # end of row_array is not empty
|
|---|
| 352 |
|
|---|
| 353 | # check "$sth->err" if empty array for error
|
|---|
| 354 | if ($statement_hand->err) {
|
|---|
| 355 | print $outhandle "DBPlugin: received error: \"" .
|
|---|
| 356 | $statement_hand->errstr . "\"\n";
|
|---|
| 357 | }
|
|---|
| 358 |
|
|---|
| 359 | # clean up connection to database
|
|---|
| 360 | $statement_hand->finish();
|
|---|
| 361 | $dbhandle->disconnect();
|
|---|
| 362 |
|
|---|
| 363 | # num of input files, rather than documents created?
|
|---|
| 364 | $self->{'num_processed'}++;
|
|---|
| 365 |
|
|---|
| 366 | if (defined($dbplug_debug) && $dbplug_debug==1) {
|
|---|
| 367 | print $outhandle "DBPlugin: imported $count DB records as documents.\n";
|
|---|
| 368 | }
|
|---|
| 369 | $count;
|
|---|
| 370 | }
|
|---|
| 371 |
|
|---|
| 372 | 1;
|
|---|
