[32518] | 1 | ###########################################################################
|
---|
| 2 | #
|
---|
[32527] | 3 | # GreenstoneSQLPlugout.pm -- plugout module for writing all or some the
|
---|
| 4 | # Greenstone document format (metadata and/or fulltext) into a (My)SQL db.
|
---|
[32526] | 5 | # The rest is then still written out by GreenstoneXMLPlugout as usual.
|
---|
[32518] | 6 | # A component of the Greenstone digital library software
|
---|
| 7 | # from the New Zealand Digital Library Project at the
|
---|
| 8 | # University of Waikato, New Zealand.
|
---|
| 9 | #
|
---|
| 10 | # Copyright (C) 2006 New Zealand Digital Library Project
|
---|
| 11 | #
|
---|
| 12 | # This program is free software; you can redistribute it and/or modify
|
---|
| 13 | # it under the terms of the GNU General Public License as published by
|
---|
| 14 | # the Free Software Foundation; either version 2 of the License, or
|
---|
| 15 | # (at your option) any later version.
|
---|
| 16 | #
|
---|
| 17 | # This program is distributed in the hope that it will be useful,
|
---|
| 18 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
| 19 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
| 20 | # GNU General Public License for more details.
|
---|
| 21 | #
|
---|
| 22 | # You should have received a copy of the GNU General Public License
|
---|
| 23 | # along with this program; if not, write to the Free Software
|
---|
| 24 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
| 25 | #
|
---|
| 26 | ###########################################################################
|
---|
| 27 |
|
---|
[32527] | 28 | package GreenstoneSQLPlugout;
|
---|
[32518] | 29 |
|
---|
| 30 | use strict;
|
---|
| 31 | no strict 'refs';
|
---|
| 32 | no strict 'subs';
|
---|
| 33 |
|
---|
[32520] | 34 | use GreenstoneXMLPlugout;
|
---|
[32518] | 35 | use docprint;
|
---|
[32529] | 36 | use gssql;
|
---|
[32518] | 37 |
|
---|
[32524] | 38 | use DBI; # the central package for this plugout
|
---|
| 39 |
|
---|
[32518] | 40 |
|
---|
[32521] | 41 | # TODO: SIGTERM rollback and disconnect?
|
---|
| 42 |
|
---|
| 43 |
|
---|
[32518] | 44 | # this plugout does not output xml to a file, but outputs rows into a mysql table
|
---|
| 45 | sub BEGIN {
|
---|
[32527] | 46 | @GreenstoneSQLPlugout::ISA = ('GreenstoneXMLPlugout');
|
---|
[32518] | 47 | }
|
---|
| 48 |
|
---|
[32529] | 49 | # NOTTODO: die() statements need to be replaced with premature_termination
|
---|
| 50 | # which should ensure the GreenstoneXMLPlugin (group)'s stuff is closed and cleaned up SOMEHOW
|
---|
| 51 | # It's fine: the die() stmts all take place before setting up the super class' begin
|
---|
[32520] | 52 |
|
---|
| 53 | # TODO: deal with -removeold and everything? Or type out instructions for user
|
---|
| 54 |
|
---|
[32518] | 55 | # TODO Q: what is "group" in GreenstoneXMLPlugout?
|
---|
[32520] | 56 | # TODO Q: site_name only exists for GS3. What about GS2?
|
---|
[32518] | 57 |
|
---|
| 58 | my $process_mode_list =
|
---|
| 59 | [ { 'name' => "meta_only",
|
---|
[32527] | 60 | 'desc' => "{GreenstoneSQLPlugout.process_mode.meta_only}" },
|
---|
[32518] | 61 | { 'name' => "text_only",
|
---|
[32527] | 62 | 'desc' => "{GreenstoneSQLPlugout.process_mode.text_only}" },
|
---|
[32518] | 63 | { 'name' => "all",
|
---|
[32527] | 64 | 'desc' => "{GreenstoneSQLPlugout.process_mode.all}" } ];
|
---|
[32518] | 65 |
|
---|
| 66 | my $arguments = [
|
---|
| 67 | { 'name' => "process_mode",
|
---|
[32527] | 68 | 'desc' => "{GreenstoneSQLPlugout.process_mode}",
|
---|
[32518] | 69 | 'type' => "enum",
|
---|
| 70 | 'list' => $process_mode_list,
|
---|
| 71 | 'deft' => "all",
|
---|
| 72 | 'reqd' => "no",
|
---|
| 73 | 'hiddengli' => "no"} ];
|
---|
| 74 |
|
---|
[32527] | 75 | my $options = { 'name' => "GreenstoneSQLPlugout",
|
---|
| 76 | 'desc' => "{GreenstoneSQLPlugout.desc}",
|
---|
[32518] | 77 | 'abstract' => "no",
|
---|
| 78 | 'inherits' => "yes",
|
---|
| 79 | 'args' => $arguments };
|
---|
| 80 |
|
---|
| 81 | sub new {
|
---|
| 82 | my ($class) = shift (@_);
|
---|
| 83 | my ($plugoutlist, $inputargs,$hashArgOptLists) = @_;
|
---|
| 84 | push(@$plugoutlist, $class);
|
---|
| 85 |
|
---|
| 86 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
---|
| 87 | push(@{$hashArgOptLists->{"OptList"}},$options);
|
---|
| 88 |
|
---|
| 89 | my $self = new GreenstoneXMLPlugout($plugoutlist,$inputargs,$hashArgOptLists);
|
---|
| 90 |
|
---|
| 91 | if ($self->{'info_only'}) {
|
---|
| 92 | # don't worry about any options etc
|
---|
| 93 | return bless $self, $class;
|
---|
| 94 | }
|
---|
[32527] | 95 | print STDERR "***** GreenstoneSQLPlugout process mode = \"", $self->{'process_mode'}, "\"\n";
|
---|
[32518] | 96 |
|
---|
| 97 | return bless $self, $class;
|
---|
| 98 | }
|
---|
| 99 |
|
---|
[32520] | 100 | # connect here and ensure all tables and databases exist
|
---|
| 101 | sub begin {
|
---|
| 102 |
|
---|
| 103 | my $self= shift (@_);
|
---|
| 104 |
|
---|
| 105 | ########### TODO: these should be set from cmdline/GLI options to plugout #########
|
---|
| 106 | $self->{'db_driver'} = "mysql";
|
---|
| 107 | $self->{'site_name'} = "localsite";
|
---|
[32521] | 108 | $self->{'db_client_user'} = "root";
|
---|
| 109 | $self->{'db_client_pwd'} = "6reenstone3";
|
---|
[32524] | 110 | $self->{'build_mode'} = "removeold";
|
---|
[32520] | 111 | #$self->{'db_host'} = "127.0.0.1";
|
---|
| 112 | #$self->{'db_encoding'} = "utf8";
|
---|
[32524] | 113 | #TODO: proc_mode is also a saveas option
|
---|
[32520] | 114 |
|
---|
| 115 | ############ LOAD NECESSARY OPTIONS ###########
|
---|
| 116 | print STDERR "########## COLLECTION: ". $ENV{'GSDLCOLLECTION'}."\n";
|
---|
[32529] | 117 | #$self->{'collection_name'} = $ENV{'GSDLCOLLECTION'};
|
---|
[32527] | 118 | print STDERR "***** GreenstoneSQLPlugout process mode = \"", $self->{'process_mode'}, "\"\n";
|
---|
[32529] | 119 |
|
---|
| 120 | my $db_params = {
|
---|
[32530] | 121 | 'collection_name' => $ENV{'GSDLCOLLECTION'},
|
---|
[32529] | 122 | 'db_encoding' => $self->{'db_encoding'}
|
---|
| 123 | #'db_name' => $self->{'site_name'},
|
---|
| 124 | #'build_mode' => $self->{'build_mode'},
|
---|
| 125 | };
|
---|
| 126 |
|
---|
| 127 | my $gs_sql = new gssql($db_params);
|
---|
[32520] | 128 |
|
---|
[32530] | 129 | # try connecting to the mysql db, if that fails it will die
|
---|
| 130 | # so don't bother preparing GreenstoneXMLPlugout by calling superclass' begin()
|
---|
| 131 | if(!$gs_sql->connect_to_db({
|
---|
| 132 | 'db_driver' => $self->{'db_driver'},
|
---|
| 133 | 'db_client_user' => $self->{'db_client_user'},
|
---|
| 134 | 'db_client_pwd' => $self->{'db_client_pwd'},
|
---|
| 135 | 'db_host' => $self->{'db_host'}
|
---|
| 136 | })
|
---|
| 137 | )
|
---|
| 138 | {
|
---|
[32520] | 139 | # This is fatal for the plugout, let's terminate here
|
---|
| 140 | # PrintError would already have displayed the warning message on connection fail
|
---|
| 141 | die("Could not connect to db. Can't proceed.\n");
|
---|
| 142 | }
|
---|
[32524] | 143 |
|
---|
| 144 | my $db_name = $self->{'site_name'} || "localsite"; # one database per GS3 site
|
---|
| 145 | my $build_mode = $self->{'build_mode'} || "removeold";
|
---|
[32529] | 146 | if(!$gs_sql->load_db_and_tables($db_name, $build_mode)) {
|
---|
[32524] | 147 |
|
---|
[32520] | 148 | # This is fatal for the plugout, let's terminate here
|
---|
| 149 | # PrintError would already have displayed the warning message on connection fail
|
---|
| 150 | die("Could not use db or prepare its tables. Can't proceed.\n");
|
---|
| 151 | }
|
---|
| 152 |
|
---|
| 153 | # prepare the shared/common HANDLES to SQL insert statements that contain placeholders
|
---|
[32521] | 154 | # and which we will reuse repeatedly when actually executing the insert statements
|
---|
[32520] | 155 | my $proc_mode = $self->{'process_mode'};
|
---|
| 156 | if($proc_mode eq "all" || $proc_mode eq "meta_only" ) {
|
---|
[32529] | 157 | $self->{'metadata_prepared_insert_statement_handle'} = $gs_sql->prepare_insert_metadata_row_stmthandle();
|
---|
[32520] | 158 | }
|
---|
| 159 | if($proc_mode eq "all" || $proc_mode eq "text_only" ) {
|
---|
[32529] | 160 | $self->{'fulltxt_prepared_insert_statement_handle'} = $gs_sql->prepare_insert_fulltxt_row_stmthandle();
|
---|
[32520] | 161 | }
|
---|
[32521] | 162 |
|
---|
[32529] | 163 | # store the DBI wrapper instance
|
---|
| 164 | $self->{'gs_sql'} = $gs_sql;
|
---|
| 165 |
|
---|
[32524] | 166 | print STDERR "#### Meta stmt: " . $self->{'metadata_prepared_insert_statement_handle'}->{'Statement'} . "\n";
|
---|
| 167 | print STDERR "#### Full stmt: " . $self->{'fulltxt_prepared_insert_statement_handle'}->{'Statement'} . "\n";
|
---|
| 168 |
|
---|
[32521] | 169 | # if setting up to work with sql db failed, we'd have terminated and wouldn't come up to here:
|
---|
| 170 | # won't bother preparing GreenstoneXMLPlugout by calling superclass' begin()
|
---|
| 171 | # finally, can call begin on super - important as doc.xml is opened as a group etc
|
---|
[32523] | 172 |
|
---|
[32526] | 173 | $self->SUPER::begin(@_);
|
---|
[32520] | 174 | }
|
---|
| 175 |
|
---|
| 176 | # disconnect from database here, see inexport.pm
|
---|
| 177 | sub end
|
---|
| 178 | {
|
---|
| 179 | my $self = shift(@_);
|
---|
| 180 |
|
---|
[32521] | 181 | # do the superclass stuff first, as any sql db failures should not prevent superclass cleanup
|
---|
[32526] | 182 | $self->SUPER::end(@_);
|
---|
[32521] | 183 |
|
---|
[32529] | 184 | $self->{'gs_sql'}->disconnect_from_db() || warn("Unable to disconnect from database " . $self->{'site_name'} . "\n"); # disconnect_from_db() will also issue a warning, but this may be clearer
|
---|
[32520] | 185 | }
|
---|
[32521] | 186 |
|
---|
[32533] | 187 | # produce files called docsql.xml instead of doc.xml
|
---|
| 188 | sub get_short_doc_file {
|
---|
| 189 | my $self = shift (@_);
|
---|
| 190 | my ($doc_dir) = @_;
|
---|
| 191 | return &FileUtils::filenameConcatenate($doc_dir, "docsql.xml");
|
---|
| 192 | }
|
---|
| 193 |
|
---|
| 194 | # produce files called docsql.xml instead of doc.xml
|
---|
| 195 | sub get_output_file {
|
---|
| 196 | my $self = shift (@_);
|
---|
| 197 | my ($doc_dir) = @_;
|
---|
| 198 | return &FileUtils::filenameConcatenate($self->{'output_dir'}, $doc_dir, "docsql.xml");
|
---|
| 199 | }
|
---|
[32521] | 200 |
|
---|
[32518] | 201 | # TODO: check arc-inf.db for whether each entry is to be deleted/indexed/reindexed/been indexed
|
---|
| 202 | sub saveas {
|
---|
| 203 | my $self = shift (@_);
|
---|
| 204 | my ($doc_obj, $doc_dir) = @_;
|
---|
| 205 |
|
---|
[32524] | 206 | print STDERR "\n\n@@@ In saveas\n\n";
|
---|
| 207 |
|
---|
[32522] | 208 | my $proc_mode = $self->{'process_mode'};
|
---|
| 209 |
|
---|
[32521] | 210 | # 1. pre save out and saving debug handle
|
---|
| 211 |
|
---|
[32523] | 212 | # must call superclass (pre/post) saveas methods, as they handle assoc_files too
|
---|
[32526] | 213 | my ($docxml_outhandler, $output_file) = $self->SUPER::pre_saveas(@_);
|
---|
[32523] | 214 |
|
---|
| 215 | $self->{'debug_outhandle'} = $docxml_outhandler if ($self->{'debug'}); # STDOUT if debug
|
---|
| 216 |
|
---|
| 217 | # TODO: also set debugging in begin()? Then stmts creating db and tables also sent to debug out and not executed
|
---|
| 218 |
|
---|
| 219 | # TODO: remove unused old_unused_saveas from GreenstoneXMLPlugout
|
---|
[32520] | 220 |
|
---|
[32521] | 221 |
|
---|
| 222 | # 2. overriding saving behaviour to do what the superclass does PLUS saving to sql db
|
---|
[32518] | 223 |
|
---|
[32523] | 224 | #NOTE: if proc_mode == all, then "breadcrumbs" go into both meta and txt elements of doc.xml:
|
---|
| 225 | # statements pointing viewer to the sql db for contents
|
---|
[32522] | 226 |
|
---|
[32523] | 227 | # write the INVERSE into doc.xml as to what is written to the db
|
---|
| 228 | my $docxml_output_options = { 'output' => docprint::OUTPUT_NONE };
|
---|
| 229 | if($proc_mode eq "meta_only" ) { # since only meta to go into MySQL db, text will go into docxml
|
---|
| 230 | $docxml_output_options->{'output'} = docprint::OUTPUT_TEXT_ONLY;
|
---|
| 231 | } elsif($proc_mode eq "text_only" ) { # since only full text to go into MySQL db, meta will go into docxml
|
---|
| 232 | $docxml_output_options->{'output'} = docprint::OUTPUT_META_ONLY;
|
---|
[32518] | 233 | }
|
---|
[32521] | 234 |
|
---|
[32523] | 235 | # now we've prepared to write out whatever is meant to go into docxml
|
---|
| 236 | # and can do actual the steps superclass GreenstoneXMLPlugout carries out to write out docxml
|
---|
| 237 | # So: write out the doc xml file for the current document
|
---|
| 238 | my $section_text = &docprint::get_section_xml($doc_obj, $docxml_output_options);
|
---|
| 239 | print $docxml_outhandler $section_text;
|
---|
| 240 |
|
---|
| 241 |
|
---|
[32518] | 242 | # We also write out whatever needs to go into the MySQL database
|
---|
| 243 | $self->write_meta_and_text($doc_obj);
|
---|
| 244 |
|
---|
[32520] | 245 |
|
---|
[32521] | 246 | # 3. post save out
|
---|
[32526] | 247 | #$self->SUPER::post_saveas(@_);
|
---|
| 248 | $self->SUPER::post_saveas($doc_obj, $doc_dir, $docxml_outhandler, $output_file);
|
---|
[32521] | 249 |
|
---|
[32523] | 250 |
|
---|
[32521] | 251 | # database connection is closed in end() method
|
---|
| 252 | # so we don't open and close over and over for each doc during a single build
|
---|
[32518] | 253 | }
|
---|
| 254 |
|
---|
| 255 |
|
---|
[32520] | 256 | # write meta and/or text PER DOC out to DB
|
---|
[32518] | 257 | sub write_meta_and_text {
|
---|
| 258 | my $self = shift (@_);
|
---|
| 259 | my ($doc_obj) = @_;
|
---|
[32531] | 260 | my $doc_oid = $doc_obj->get_OID(); # this method processes a single doc at a time, so it uses the same OID throughout
|
---|
[32518] | 261 | my $root_section = $doc_obj->get_top_section();
|
---|
| 262 |
|
---|
[32520] | 263 | # load the prepared INSERT statement handles for both tables (can be undef for any table depending on whether meta_only or txt_only are set)
|
---|
| 264 | my $metadata_table_sth = $self->{'metadata_prepared_insert_statement_handle'};
|
---|
| 265 | my $fulltxt_table_sth = $self->{'fulltxt_prepared_insert_statement_handle'};
|
---|
[32518] | 266 |
|
---|
[32524] | 267 | $self->recursive_write_meta_and_text($doc_obj, $doc_oid, $root_section, $metadata_table_sth, $fulltxt_table_sth);
|
---|
[32521] | 268 | }
|
---|
[32518] | 269 |
|
---|
| 270 | # Perl: Reading or Writing to Another Program
|
---|
| 271 | # https://nnc3.com/mags/Perl3/cookbook/ch16_05.htm
|
---|
| 272 | sub recursive_write_meta_and_text {
|
---|
[32520] | 273 | my $self = shift (@_);
|
---|
| 274 | my ($doc_obj, $doc_oid, $section, $metadata_table_sth, $fulltxt_table_sth) = @_;
|
---|
| 275 |
|
---|
| 276 | # If section=ROOT, write "root" as section name into table
|
---|
| 277 | # doc->get_top_section() is the name of the doc root section, which is ""
|
---|
| 278 | my $section_name = ($section eq "") ? "root" : $section;
|
---|
[32518] | 279 |
|
---|
| 280 | my $section_ptr = $doc_obj->_lookup_section ($section);
|
---|
| 281 | return "" unless defined $section_ptr;
|
---|
| 282 |
|
---|
[32520] | 283 | my $debug_out = $self->{'debug_outhandle'};
|
---|
[32531] | 284 | # print STDERR "#### Meta stmt: " . $metadata_table_sth->{'Statement'} . "\n";
|
---|
| 285 | # print STDERR "#### Full stmt: " . $fulltxt_table_sth->{'Statement'} . "\n";
|
---|
[32520] | 286 |
|
---|
| 287 | #my $proc_mode = $self->{'process_mode'};
|
---|
| 288 | #if($proc_mode eq "all" || $proc_mode eq "meta_only" ) {
|
---|
| 289 | if($metadata_table_sth) { # meta insert statement handle will be undef if not writing meta
|
---|
| 290 |
|
---|
| 291 | foreach my $data (@{$section_ptr->{'metadata'}}) {
|
---|
| 292 | my $meta_name = $data->[0];
|
---|
[32524] | 293 | my $escaped_meta_value = &docprint::escape_text($data->[1]);
|
---|
[32518] | 294 |
|
---|
[32520] | 295 | # Write out the current section's meta to collection db's METADATA table
|
---|
| 296 |
|
---|
| 297 | # for each set of values to write to meta table, execute the prepared statement, filling in the values
|
---|
| 298 |
|
---|
| 299 | if($self->{'debug'}) {
|
---|
| 300 | # just print the statement we were going to execute
|
---|
| 301 |
|
---|
[32524] | 302 | print $debug_out $metadata_table_sth->{'Statement'} . "($doc_oid, $section_name, $meta_name, $escaped_meta_value)\n";
|
---|
[32520] | 303 | }
|
---|
| 304 | else {
|
---|
[32533] | 305 |
|
---|
| 306 | $metadata_table_sth->execute($doc_oid, $section_name, $meta_name, $escaped_meta_value)
|
---|
| 307 | || warn ("Unable to write metadata row to db:\n\tOID $doc_oid, section $section_name,\n\tmeta name: $meta_name, val: $escaped_meta_value");
|
---|
[32520] | 308 | # Execution failure will print out info anyway: since db connection sets PrintError
|
---|
| 309 | }
|
---|
[32518] | 310 | }
|
---|
| 311 | }
|
---|
[32520] | 312 |
|
---|
| 313 | #if($proc_mode eq "all" || $proc_mode eq "text_only" ) {
|
---|
| 314 | if($fulltxt_table_sth) { # fulltxt insert statement handle will be undef if not writing fulltxt
|
---|
| 315 |
|
---|
| 316 | if($self->{'debug'}) {
|
---|
| 317 | # just print the statement we were going to execute, minus the fulltxt value
|
---|
| 318 | my $txt_repr = $section_ptr->{'text'} ? "<TXT>" : "NULL";
|
---|
[32524] | 319 | print $debug_out $fulltxt_table_sth->{'Statement'} . "($doc_oid, $section_name, $txt_repr)\n";
|
---|
[32520] | 320 | } else {
|
---|
[32524] | 321 | my $section_text = &docprint::escape_text($section_ptr->{'text'});
|
---|
[32520] | 322 |
|
---|
| 323 | # fulltxt column can be SQL NULL. undef value gets written out as NULL:
|
---|
| 324 | # https://stackoverflow.com/questions/12708633/which-one-represents-null-undef-or-empty-string
|
---|
| 325 |
|
---|
| 326 | # Write out the current section's text to collection db's FULLTeXT table
|
---|
[32533] | 327 | $fulltxt_table_sth->execute($doc_oid, $section_name, $section_text)
|
---|
| 328 | || warn ("Unable to write fulltxt row to db for row:\n\tOID $doc_oid, section $section_name");
|
---|
[32520] | 329 | # Execution failure will print out info anyway: since db connection sets PrintError
|
---|
| 330 | }
|
---|
[32518] | 331 | }
|
---|
| 332 |
|
---|
| 333 | # output all subsections: RECURSIVE CALL
|
---|
| 334 | foreach my $subsection (@{$section_ptr->{'subsection_order'}}) {
|
---|
[32531] | 335 | $self->recursive_write_meta_and_text($doc_obj, $doc_oid, "$section.$subsection", $metadata_table_sth, $fulltxt_table_sth);
|
---|
[32518] | 336 | }
|
---|
| 337 | }
|
---|
| 338 |
|
---|
| 339 |
|
---|
| 340 | 1;
|
---|