root/main/trunk/greenstone2/perllib/plugouts/GreenstoneSQLPlugout.pm @ 32542

Revision 32542, 15.0 KB (checked in by ak19, 9 months ago)

Instead of the docoid being stored in the docsql-<OID>.xml filename, all filenames produced are back to being docsql.xml, but the root element Archive now contains the doc oid as attribute: <Archive docoid="oid">

Line 
1###########################################################################
2#
3# GreenstoneSQLPlugout.pm -- plugout module for writing all or some the
4# Greenstone document format (metadata and/or fulltext) into a (My)SQL db.
5# The rest is then still written out by GreenstoneXMLPlugout as usual.
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
28package GreenstoneSQLPlugout;
29
30use strict;
31no strict 'refs';
32no strict 'subs';
33
34use GreenstoneXMLPlugout;
35use docprint;
36use gssql;
37
38use DBI; # the central package for this plugout
39
40
41# TODO: SIGTERM rollback and disconnect?
42# TODO Q: what about verbosity for debugging
43# TODO Q: introduced site_name param to plugins and plugouts. Did I do it right? And should they have hiddengli = "yes"
44
45# this plugout does not output xml to a file, but outputs rows into a mysql table
46sub BEGIN {
47    @GreenstoneSQLPlugout::ISA = ('GreenstoneXMLPlugout');
48}
49
50# NOTTODO: die() statements need to be replaced with premature_termination
51# which should ensure the GreenstoneXMLPlugin (group)'s stuff is closed and cleaned up SOMEHOW
52# It's fine: the die() stmts all take place before setting up the super class' begin
53
54# TODO Q: about build_mode: how to detect removeold
55# TODO: deal with -removeold and everything? Or type out instructions for user
56
57# TODO Q: what is "group" in GreenstoneXMLPlugout?
58# TODO Q: site_name only exists for GS3. What about GS2?
59
60
61my $process_mode_list =
62    [ { 'name' => "meta_only",
63        'desc' => "{GreenstoneSQLPlug.process_mode.meta_only}" },     
64      { 'name' => "text_only",
65        'desc' => "{GreenstoneSQLPlug.process_mode.text_only}" },
66      { 'name' => "all",
67        'desc' => "{GreenstoneSQLPlug.process_mode.all}" } ];
68
69# The following are the saveas.options:
70my $arguments = [
71    { 'name' => "process_mode",
72      'desc' => "{GreenstoneSQLPlug.process_mode}",
73      'type' => "enum",
74      'list' => $process_mode_list,
75      'deft' => "all",
76      'reqd' => "no",
77      'hiddengli' => "no"},
78    { 'name' => "db_driver",
79      'desc' => "{GreenstoneSQLPlug.db_driver}",
80      'type' => "string",   
81      'deft' => "mysql",
82      'reqd' => "yes"},
83    { 'name' => "db_client_user",
84      'desc' => "{GreenstoneSQLPlug.db_client_user}",
85      'type' => "string",   
86      'deft' => "root",
87      'reqd' => "yes"},
88    { 'name' => "db_client_pwd",
89      'desc' => "{GreenstoneSQLPlug.db_client_pwd}",
90      'type' => "string",
91      'deft' => "",
92      'reqd' => "yes"}, # pwd required?
93    { 'name' => "db_host",
94      'desc' => "{GreenstoneSQLPlug.db_host}",
95      'type' => "string",
96      'deft' => "127.0.0.1",
97      'reqd' => "yes"},
98    { 'name' => "db_encoding",
99      'desc' => "{GreenstoneSQLPlug.db_encoding}",
100    'type' => "string",
101      'deft' => "utf8",
102      'reqd' => "yes"}
103    ];
104
105my $options = { 'name'     => "GreenstoneSQLPlugout",
106        'desc'     => "{GreenstoneSQLPlugout.desc}",
107        'abstract' => "no",
108        'inherits' => "yes",
109        'args'     => $arguments };
110
111sub new {
112    my ($class) = shift (@_);
113    my ($plugoutlist, $inputargs,$hashArgOptLists) = @_;
114    push(@$plugoutlist, $class);
115
116    push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
117    push(@{$hashArgOptLists->{"OptList"}},$options);
118
119    my $self = new GreenstoneXMLPlugout($plugoutlist,$inputargs,$hashArgOptLists);
120   
121    if ($self->{'info_only'}) {
122        # don't worry about any options etc
123        return bless $self, $class;
124    }
125    print STDERR "***** GreenstoneSQLPlugout process mode = \"", $self->{'process_mode'}, "\"\n";
126   
127    return bless $self, $class;
128}
129
130# connect here and ensure all tables and databases exist
131sub begin {
132
133    my $self= shift (@_);
134
135    # The saveas.options
136    #print STDERR "@@@@ PLUGOUT db_pwd: " . $self->{'db_client_pwd'} . "\n";
137    #print STDERR "@@@@ user: " . $self->{'db_client_user'} . "\n";
138    #print STDERR "@@@@ db_host: " . $self->{'db_host'} . "\n";
139    #print STDERR "@@@@ db_enc: " . $self->{'db_encoding'} . "\n";
140    #print STDERR "@@@@ db_driver: " . $self->{'db_driver'} . "\n";
141    #print STDERR "@@@@ proc_mode: " . $self->{'process_mode'} . "\n";
142   
143    ########### TODO: these should be set from cmdline/GLI options to plugout #########
144
145    $self->{'build_mode'} = "removeold";
146
147    ############ LOAD NECESSARY OPTIONS ###########
148    #print "@@@ plugout SITE NAME: ". $self->{'site_name'} . "\n" if defined $self->{'site_name'};
149    #print STDERR "########## COLLECTION: ". $ENV{'GSDLCOLLECTION'}."\n";
150
151    print STDERR "***** GreenstoneSQLPlugout process mode = \"", $self->{'process_mode'}, "\"\n";
152
153    my $db_params = {
154    'collection_name' => $ENV{'GSDLCOLLECTION'},   
155    'db_encoding' => $self->{'db_encoding'}
156    };
157
158    my $gs_sql = new gssql($db_params);
159   
160    # try connecting to the mysql db, if that fails it will die
161    # so don't bother preparing GreenstoneXMLPlugout by calling superclass' begin()
162    if(!$gs_sql->connect_to_db({
163    'db_driver' => $self->{'db_driver'},
164    'db_client_user' => $self->{'db_client_user'},
165    'db_client_pwd' => $self->{'db_client_pwd'},
166    'db_host' => $self->{'db_host'}
167                   })
168    )
169    {
170    # This is fatal for the plugout, let's terminate here
171    # PrintError would already have displayed the warning message on connection fail   
172    die("Could not connect to db. Can't proceed.\n");
173    }
174   
175    my $db_name = $self->{'site_name'} || "greenstone2"; # one database per GS3 site, for GS2 the db is called greenstone2
176    my $build_mode = $self->{'build_mode'} || "removeold";
177    if(!$gs_sql->load_db_and_tables($db_name, $build_mode)) {
178   
179    # This is fatal for the plugout, let's terminate here after disconnecting again
180    # PrintError would already have displayed the warning message on load fail
181    $gs_sql->disconnect_from_db()
182        || warn("Unable to disconnect from database.\n");
183    die("Could not use db $db_name and/or prepare its tables. Can't proceed.\n");
184    }
185
186    # prepare the shared/common HANDLES to SQL insert statements that contain placeholders
187    # and which we will reuse repeatedly when actually executing the insert statements
188    my $proc_mode = $self->{'process_mode'};
189    if($proc_mode eq "all" || $proc_mode eq "meta_only" ) {
190    $self->{'metadata_prepared_insert_statement_handle'} = $gs_sql->prepare_insert_metadata_row_stmthandle();
191    }
192    if($proc_mode eq "all" || $proc_mode eq "text_only" ) {
193    $self->{'fulltxt_prepared_insert_statement_handle'} = $gs_sql->prepare_insert_fulltxt_row_stmthandle();
194    }
195
196    # store the DBI wrapper instance
197    $self->{'gs_sql'} = $gs_sql;
198   
199    print STDERR "#### Meta stmt: " . $self->{'metadata_prepared_insert_statement_handle'}->{'Statement'} . "\n";
200    print STDERR "#### Full stmt: " . $self->{'fulltxt_prepared_insert_statement_handle'}->{'Statement'} . "\n";
201   
202    # if setting up to work with sql db failed, we'd have terminated and wouldn't come up to here:
203    # won't bother preparing GreenstoneXMLPlugout by calling superclass' begin()
204    # finally, can call begin on super - important as doc.xml is opened as a group etc
205   
206    $self->SUPER::begin(@_);
207}
208
209# disconnect from database here, see inexport.pm
210sub end
211{
212    my $self = shift(@_);
213
214    # do the superclass stuff first, as any sql db failures should not prevent superclass cleanup
215    $self->SUPER::end(@_);   
216   
217    $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
218}
219
220# Produce files called docsql.xml instead of doc.xml
221sub get_doc_xml_filename {
222    my $self = shift (@_);
223    my ($doc_obj) = @_;
224   
225    return "docsql.xml";
226}
227
228# overriding to store doc OID as attribute of top level element: <Archive docoid="oid">
229sub output_xml_header {
230    my $self = shift (@_);
231    my ($outhandle, $doc_oid) = @_;
232
233    print $outhandle '<?xml version="1.0" encoding="utf-8" standalone="no"?>' . "\n";
234    print $outhandle "<!DOCTYPE Archive SYSTEM \"http://greenstone.org/dtd/Archive/1.0/Archive.dtd\">\n";
235    print $outhandle "<Archive docoid=\"$doc_oid\">\n";
236}
237 
238# TODO: check arc-inf.db for whether each entry is to be deleted/indexed/reindexed/been indexed
239sub saveas {
240    my $self = shift (@_);
241    my ($doc_obj, $doc_dir) = @_;
242
243#    print STDERR "\n\n@@@ In saveas\n\n";
244   
245    my $proc_mode = $self->{'process_mode'};
246   
247    # 1. pre save out and saving debug handle
248
249    # must call superclass (pre/post) saveas methods, as they handle assoc_files too
250    my ($docxml_outhandler, $output_file) = $self->SUPER::pre_saveas(@_);
251
252    $self->{'debug_outhandle'} = $docxml_outhandler if ($self->{'debug'}); # STDOUT if debug
253
254    # TODO: also set debugging in begin()? Then stmts creating db and tables also sent to debug out and not executed
255
256    # TODO: remove unused old_unused_saveas from GreenstoneXMLPlugout
257   
258   
259    # 2. overriding saving behaviour to do what the superclass does PLUS saving to sql db   
260
261    #NOTE: if proc_mode == all, then "breadcrumbs" go into both meta and txt elements of doc.xml:
262    # statements pointing viewer to the sql db for contents
263   
264    # write the INVERSE into doc.xml as to what is written to the db   
265    my $docxml_output_options = { 'output' => docprint::OUTPUT_NONE };
266    if($proc_mode eq "meta_only" ) { # since only meta to go into MySQL db, text will go into docxml
267    $docxml_output_options->{'output'} = docprint::OUTPUT_TEXT_ONLY;
268    } elsif($proc_mode eq "text_only" ) { # since only full text to go into MySQL db, meta will go into docxml
269    $docxml_output_options->{'output'} = docprint::OUTPUT_META_ONLY;
270    }
271   
272    # now we've prepared to write out whatever is meant to go into docxml
273    # and can do actual the steps superclass GreenstoneXMLPlugout carries out to write out docxml
274    # So: write out the doc xml file for the current document
275    my $section_text = &docprint::get_section_xml($doc_obj, $docxml_output_options);
276    print $docxml_outhandler $section_text;   
277   
278   
279    # We also write out whatever needs to go into the MySQL database
280    $self->write_meta_and_text($doc_obj);
281
282   
283    # 3. post save out
284    #$self->SUPER::post_saveas(@_);
285    $self->SUPER::post_saveas($doc_obj, $doc_dir, $docxml_outhandler, $output_file);
286   
287   
288    # database connection is closed in end() method
289    # so we don't open and close over and over for each doc during a single build
290}
291
292
293# write meta and/or text PER DOC out to DB
294sub write_meta_and_text {
295    my $self = shift (@_);
296    my ($doc_obj) = @_;
297    my $doc_oid = $doc_obj->get_OID(); # this method processes a single doc at a time, so it uses the same OID throughout
298    my $root_section = $doc_obj->get_top_section();
299
300    # 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)
301    my $metadata_table_sth = $self->{'metadata_prepared_insert_statement_handle'};
302    my $fulltxt_table_sth = $self->{'fulltxt_prepared_insert_statement_handle'};
303   
304    $self->recursive_write_meta_and_text($doc_obj, $doc_oid, $root_section, $metadata_table_sth, $fulltxt_table_sth);
305}
306
307# Perl: Reading or Writing to Another Program
308# https://nnc3.com/mags/Perl3/cookbook/ch16_05.htm
309sub recursive_write_meta_and_text {
310    my $self = shift (@_);
311    my ($doc_obj, $doc_oid, $section, $metadata_table_sth, $fulltxt_table_sth) = @_;   
312
313    # If section=ROOT, write "root" as section name into table
314    # doc->get_top_section() is the name of the doc root section, which is ""
315    my $section_name = ($section eq "") ? "root" : $section;
316   
317    my $section_ptr = $doc_obj->_lookup_section ($section);
318    return "" unless defined $section_ptr;
319
320    my $debug_out = $self->{'debug_outhandle'};
321#    print STDERR "#### Meta stmt: " . $metadata_table_sth->{'Statement'} . "\n";
322#    print STDERR "#### Full stmt: " . $fulltxt_table_sth->{'Statement'} . "\n";
323   
324    #my $proc_mode = $self->{'process_mode'};
325    #if($proc_mode eq "all" || $proc_mode eq "meta_only" ) {
326    if($metadata_table_sth) { # meta insert statement handle will be undef if not writing meta
327   
328    foreach my $data (@{$section_ptr->{'metadata'}}) {
329        my $meta_name = $data->[0];
330        # TODO: does it need to be stored escaped, as it requires unescaping when read back in
331        # from db (unlike for reading back in from doc.xml)
332        my $escaped_meta_value = &docprint::escape_text($data->[1]);
333
334        # Write out the current section's meta to collection db's METADATA table       
335       
336        # for each set of values to write to meta table, execute the prepared statement, filling in the values
337
338        if($self->{'debug'}) {
339        # just print the statement we were going to execute
340
341        print $debug_out $metadata_table_sth->{'Statement'} . "($doc_oid, $section_name, $meta_name, $escaped_meta_value)\n";
342        }
343        else {
344       
345        $metadata_table_sth->execute($doc_oid, $section_name, $meta_name, $escaped_meta_value)
346            || warn ("Unable to write metadata row to db:\n\tOID $doc_oid, section $section_name,\n\tmeta name: $meta_name, val: $escaped_meta_value");
347        # Execution failure will print out info anyway: since db connection sets PrintError
348        }
349    }
350    }
351   
352    #if($proc_mode eq "all" || $proc_mode eq "text_only" ) {
353    if($fulltxt_table_sth) { # fulltxt insert statement handle will be undef if not writing fulltxt
354
355    if($self->{'debug'}) {
356        # just print the statement we were going to execute, minus the fulltxt value
357        my $txt_repr = $section_ptr->{'text'} ? "<TXT>" : "NULL";
358        print $debug_out $fulltxt_table_sth->{'Statement'} . "($doc_oid, $section_name, $txt_repr)\n";
359    } else {
360        my $section_text = &docprint::escape_text($section_ptr->{'text'});
361       
362        # fulltxt column can be SQL NULL. undef value gets written out as NULL:
363        # https://stackoverflow.com/questions/12708633/which-one-represents-null-undef-or-empty-string
364       
365        # Write out the current section's text to collection db's FULLTeXT table
366        $fulltxt_table_sth->execute($doc_oid, $section_name, $section_text)
367        || warn ("Unable to write fulltxt row to db for row:\n\tOID $doc_oid, section $section_name");
368        # Execution failure will print out info anyway: since db connection sets PrintError
369    }
370    }
371   
372    # output all subsections: RECURSIVE CALL
373    foreach my $subsection (@{$section_ptr->{'subsection_order'}}) {
374    $self->recursive_write_meta_and_text($doc_obj, $doc_oid, "$section.$subsection", $metadata_table_sth, $fulltxt_table_sth);
375    }
376}
377
378
3791;
Note: See TracBrowser for help on using the browser.