root/main/trunk/greenstone2/perllib/plugouts/GreenstoneSQLPlugout.pm @ 32555

Revision 32555, 15.2 KB (checked in by ak19, 21 months ago)

1. In GreenstoneSQLPlugout, removeold is now paramterised (as are keepold, incremental, incremental_mode). 2. Deletion on incremental_build works. But there are more questions. Why are there 4 passes? What to do on reindexing and when to do it (should it happen during GS SQL plugout or plugin)?

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, instead of current situation of printing out upon debug set at the expense of writing to db
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 the metadata and/or fulltxt xml to a file,
46# but outputs rows into a mysql table for metadata and/or a table for fulltxt
47sub BEGIN {
48    @GreenstoneSQLPlugout::ISA = ('GreenstoneXMLPlugout');
49}
50
51# NOTTODO: die() statements need to be replaced with premature_termination
52# which should ensure the GreenstoneXMLPlugin (group)'s stuff is closed and cleaned up SOMEHOW
53# It's fine: the die() stmts all take place before setting up the super class' begin
54
55# TODO Q: about build_mode: how to detect removeold
56# TODO: deal with -removeold and everything? Or type out instructions for user
57
58# TODO Q: what is "group" in GreenstoneXMLPlugout?
59
60my $process_mode_list =
61    [ { 'name' => "meta_only",
62        'desc' => "{GreenstoneSQLPlug.process_mode.meta_only}" },     
63      { 'name' => "text_only",
64        'desc' => "{GreenstoneSQLPlug.process_mode.text_only}" },
65      { 'name' => "all",
66        'desc' => "{GreenstoneSQLPlug.process_mode.all}" } ];
67
68# The following are the saveas.options:
69my $arguments = [
70    { 'name' => "process_mode",
71      'desc' => "{GreenstoneSQLPlug.process_mode}",
72      'type' => "enum",
73      'list' => $process_mode_list,
74      'deft' => "all",
75      'reqd' => "no",
76      'hiddengli' => "no"},
77    { 'name' => "db_driver",
78      'desc' => "{GreenstoneSQLPlug.db_driver}",
79      'type' => "string",   
80      'deft' => "mysql",
81      'reqd' => "yes"},
82    { 'name' => "db_client_user",
83      'desc' => "{GreenstoneSQLPlug.db_client_user}",
84      'type' => "string",   
85      'deft' => "root",
86      'reqd' => "yes"},
87    { 'name' => "db_client_pwd",
88      'desc' => "{GreenstoneSQLPlug.db_client_pwd}",
89      'type' => "string",
90      'deft' => "",
91      'reqd' => "yes"}, # pwd required?
92    { 'name' => "db_host",
93      'desc' => "{GreenstoneSQLPlug.db_host}",
94      'type' => "string",
95      'deft' => "127.0.0.1",
96      'reqd' => "yes"},
97    { 'name' => "db_encoding",
98      'desc' => "{GreenstoneSQLPlug.db_encoding}",
99    'type' => "string",
100      'deft' => "utf8",
101      'reqd' => "yes"}
102    ];
103
104my $options = { 'name'     => "GreenstoneSQLPlugout",
105        'desc'     => "{GreenstoneSQLPlugout.desc}",
106        'abstract' => "no",
107        'inherits' => "yes",
108        'args'     => $arguments };
109
110sub new {
111    my ($class) = shift (@_);
112    my ($plugoutlist, $inputargs,$hashArgOptLists) = @_;
113    push(@$plugoutlist, $class);
114
115    push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
116    push(@{$hashArgOptLists->{"OptList"}},$options);
117
118    my $self = new GreenstoneXMLPlugout($plugoutlist,$inputargs,$hashArgOptLists);
119   
120    if ($self->{'info_only'}) {
121        # don't worry about any options etc
122        return bless $self, $class;
123    }
124    print STDERR "***** GreenstoneSQLPlugout process mode = \"", $self->{'process_mode'}, "\"\n";
125   
126    return bless $self, $class;
127}
128
129# connect here and ensure all tables and databases exist
130sub begin {
131
132    my $self= shift (@_);
133
134    # The saveas.options
135    #print STDERR "@@@@ PLUGOUT db_pwd: " . $self->{'db_client_pwd'} . "\n";
136    #print STDERR "@@@@ user: " . $self->{'db_client_user'} . "\n";
137    #print STDERR "@@@@ db_host: " . $self->{'db_host'} . "\n";
138    #print STDERR "@@@@ db_enc: " . $self->{'db_encoding'} . "\n";
139    #print STDERR "@@@@ db_driver: " . $self->{'db_driver'} . "\n";
140    #print STDERR "@@@@ proc_mode: " . $self->{'process_mode'} . "\n";
141
142    ############ LOAD NECESSARY OPTIONS ###########
143    #print "@@@ plugout SITE NAME: ". $self->{'site_name'} . "\n" if defined $self->{'site_name'};
144    #print STDERR "########## COLLECTION: ". $ENV{'GSDLCOLLECTION'}."\n";
145
146    print STDERR "***** GreenstoneSQLPlugout process mode = \"", $self->{'process_mode'}, "\"\n";
147
148    my $db_params = {
149    'collection_name' => $ENV{'GSDLCOLLECTION'},   
150    'db_encoding' => $self->{'db_encoding'}
151    };
152
153    my $gs_sql = new gssql($db_params);
154   
155    # try connecting to the mysql db, if that fails it will die
156    # so don't bother preparing GreenstoneXMLPlugout by calling superclass' begin()
157    if(!$gs_sql->connect_to_db({
158    'db_driver' => $self->{'db_driver'},
159    'db_client_user' => $self->{'db_client_user'},
160    'db_client_pwd' => $self->{'db_client_pwd'},
161    'db_host' => $self->{'db_host'}
162                   })
163    )
164    {
165    # This is fatal for the plugout, let's terminate here
166    # PrintError would already have displayed the warning message on connection fail   
167    die("Could not connect to db. Can't proceed.\n");
168    }
169   
170    my $db_name = $self->{'site_name'} || "greenstone2"; # one database per GS3 site, for GS2 the db is called greenstone2
171    my $build_mode = ($self->{'removeold'}) ? "removeold" : "incremental";
172    print STDERR "@@@@@@@@@@@@ remove_old: $build_mode\n";
173   
174    if(!$gs_sql->load_db_and_tables($db_name, $build_mode)) {
175   
176    # This is fatal for the plugout, let's terminate here after disconnecting again
177    # PrintError would already have displayed the warning message on load fail
178    $gs_sql->disconnect_from_db()
179        || warn("Unable to disconnect from database.\n");
180    die("Could not use db $db_name and/or prepare its tables. Can't proceed.\n");
181    }
182
183    # prepare the shared/common HANDLES to SQL insert statements that contain placeholders
184    # and which we will reuse repeatedly when actually executing the insert statements
185    my $proc_mode = $self->{'process_mode'};
186    if($proc_mode eq "all" || $proc_mode eq "meta_only" ) {
187    $self->{'metadata_prepared_insert_statement_handle'} = $gs_sql->prepare_insert_metadata_row_stmthandle();
188    }
189    if($proc_mode eq "all" || $proc_mode eq "text_only" ) {
190    $self->{'fulltxt_prepared_insert_statement_handle'} = $gs_sql->prepare_insert_fulltxt_row_stmthandle();
191    }
192
193    # store the DBI wrapper instance
194    $self->{'gs_sql'} = $gs_sql;
195   
196    print STDERR "#### Meta stmt: " . $self->{'metadata_prepared_insert_statement_handle'}->{'Statement'} . "\n";
197    print STDERR "#### Full stmt: " . $self->{'fulltxt_prepared_insert_statement_handle'}->{'Statement'} . "\n";
198   
199    # if setting up to work with sql db failed, we'd have terminated and wouldn't come up to here:
200    # won't bother preparing GreenstoneXMLPlugout by calling superclass' begin()
201    # finally, can call begin on super - important as doc.xml is opened as a group etc
202   
203    $self->SUPER::begin(@_);
204}
205
206# disconnect from database here, see inexport.pm
207sub end
208{
209    my $self = shift(@_);
210
211    # do the superclass stuff first, as any sql db failures should not prevent superclass cleanup
212    $self->SUPER::end(@_);   
213   
214    $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
215}
216
217# Produce files called docsql.xml instead of doc.xml
218sub get_doc_xml_filename {
219    my $self = shift (@_);
220    my ($doc_obj) = @_;
221   
222    return "docsql.xml";
223}
224
225# overriding to store doc OID as attribute of top level element: <Archive docoid="oid">
226sub output_xml_header {
227    my $self = shift (@_);
228    my ($outhandle, $doc_oid) = @_;
229
230    print $outhandle '<?xml version="1.0" encoding="utf-8" standalone="no"?>' . "\n";
231    print $outhandle "<!DOCTYPE Archive SYSTEM \"http://greenstone.org/dtd/Archive/1.0/Archive.dtd\">\n";
232    print $outhandle "<Archive docoid=\"$doc_oid\">\n";
233}
234 
235# TODO: check arc-inf.db for whether each entry is to be deleted/indexed/reindexed/been indexed?
236# That's only for indexing, not for this step which only generates the content in archives dir
237sub saveas {
238    my $self = shift (@_);
239    my ($doc_obj, $doc_dir) = @_;
240
241#    print STDERR "\n\n@@@ In saveas\n\n";
242   
243    my $proc_mode = $self->{'process_mode'};
244   
245    # 1. pre save out and saving debug handle
246
247    # must call superclass (pre/post) saveas methods, as they handle assoc_files too
248    my ($docxml_outhandler, $output_file) = $self->SUPER::pre_saveas(@_);
249
250    $self->{'debug_outhandle'} = $docxml_outhandler if ($self->{'debug'}); # STDOUT if debug
251
252    # TODO: also set debugging in begin()? Then stmts creating db and tables also sent to debug out and not executed
253   
254   
255    # 2. overriding saving behaviour to do what the superclass does (writing out doc.xml files,
256    # under new name of docsql.xml, with breadcrumbs pointing to sql db) PLUS saving to sql db   
257
258    # NOTE: if proc_mode == all, then "breadcrumbs" (statements pointing viewer to the sql db
259    # for contents) go into both meta and txt elements of doc.xml (docsql.xml specifically):
260   
261    # write the INVERSE into doc.xml as to what is written to the SQL db   
262    my $docxml_output_options = { 'output' => docprint::OUTPUT_NONE };
263    if($proc_mode eq "meta_only" ) { # since only meta to go into MySQL db, text will go into docxml
264    $docxml_output_options->{'output'} = docprint::OUTPUT_TEXT_ONLY;
265    } elsif($proc_mode eq "text_only" ) { # since only full text to go into MySQL db, meta will go into docxml
266    $docxml_output_options->{'output'} = docprint::OUTPUT_META_ONLY;
267    }
268   
269    # now we've prepared to write out whatever is meant to go into docxml
270    # and can do actual the steps superclass GreenstoneXMLPlugout carries out to write out docxml
271    # So: write out the doc xml file, "docsql.xml", for the current document
272    my $section_text = &docprint::get_section_xml($doc_obj, $docxml_output_options);
273    print $docxml_outhandler $section_text;   
274   
275   
276    # We also write out whatever needs to go into the MySQL database
277    $self->write_meta_and_text($doc_obj);
278
279   
280    # 3. post save out
281    #$self->SUPER::post_saveas(@_);
282    $self->SUPER::post_saveas($doc_obj, $doc_dir, $docxml_outhandler, $output_file);
283   
284   
285    # database connection is closed in end() method
286    # so we don't open and close over and over for each doc during a single build
287}
288
289
290# write meta and/or text PER DOC out to DB
291sub write_meta_and_text {
292    my $self = shift (@_);
293    my ($doc_obj) = @_;
294    my $doc_oid = $doc_obj->get_OID(); # this method processes a single doc at a time, so it uses the same OID throughout
295    my $root_section = $doc_obj->get_top_section();
296
297    # 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)
298    my $metadata_table_sth = $self->{'metadata_prepared_insert_statement_handle'};
299    my $fulltxt_table_sth = $self->{'fulltxt_prepared_insert_statement_handle'};
300   
301    $self->recursive_write_meta_and_text($doc_obj, $doc_oid, $root_section, $metadata_table_sth, $fulltxt_table_sth);
302}
303
304# Perl: Reading or Writing to Another Program
305# https://nnc3.com/mags/Perl3/cookbook/ch16_05.htm
306sub recursive_write_meta_and_text {
307    my $self = shift (@_);
308    my ($doc_obj, $doc_oid, $section, $metadata_table_sth, $fulltxt_table_sth) = @_;   
309
310    # If section=ROOT, write "root" as section name into table
311    # doc->get_top_section() is the name of the doc root section, which is ""
312    my $section_name = ($section eq "") ? "root" : $section;
313   
314    my $section_ptr = $doc_obj->_lookup_section ($section);
315    return "" unless defined $section_ptr;
316
317    my $debug_out = $self->{'debug_outhandle'};
318#    print STDERR "#### Meta stmt: " . $metadata_table_sth->{'Statement'} . "\n";
319#    print STDERR "#### Full stmt: " . $fulltxt_table_sth->{'Statement'} . "\n";
320   
321    #my $proc_mode = $self->{'process_mode'};
322    #if($proc_mode eq "all" || $proc_mode eq "meta_only" ) {
323    if($metadata_table_sth) { # meta insert statement handle will be undef if not writing meta
324   
325    foreach my $data (@{$section_ptr->{'metadata'}}) {
326        my $meta_name = $data->[0];
327        # TODO: does it need to be stored escaped, as it requires unescaping when read back in
328        # from db (unlike for reading back in from doc.xml)
329        my $escaped_meta_value = &docprint::escape_text($data->[1]);
330
331        # Write out the current section's meta to collection db's METADATA table       
332       
333        # for each set of values to write to meta table, execute the prepared statement, filling in the values
334
335        if($self->{'debug'}) {
336        # just print the statement we were going to execute
337
338        print $debug_out $metadata_table_sth->{'Statement'} . "($doc_oid, $section_name, $meta_name, $escaped_meta_value)\n";
339        }
340        else {
341       
342        $metadata_table_sth->execute($doc_oid, $section_name, $meta_name, $escaped_meta_value)
343            || warn ("Unable to write metadata row to db:\n\tOID $doc_oid, section $section_name,\n\tmeta name: $meta_name, val: $escaped_meta_value");
344        # Execution failure will print out info anyway: since db connection sets PrintError
345        }
346    }
347    }
348   
349    #if($proc_mode eq "all" || $proc_mode eq "text_only" ) {
350    if($fulltxt_table_sth) { # fulltxt insert statement handle will be undef if not writing fulltxt
351
352    if($self->{'debug'}) {
353        # just print the statement we were going to execute, minus the fulltxt value
354        my $txt_repr = $section_ptr->{'text'} ? "<TXT>" : "NULL";
355        print $debug_out $fulltxt_table_sth->{'Statement'} . "($doc_oid, $section_name, $txt_repr)\n";
356    } else {
357        my $section_text = &docprint::escape_text($section_ptr->{'text'});
358       
359        # fulltxt column can be SQL NULL. undef value gets written out as NULL:
360        # https://stackoverflow.com/questions/12708633/which-one-represents-null-undef-or-empty-string
361       
362        # Write out the current section's text to collection db's FULLTeXT table
363        $fulltxt_table_sth->execute($doc_oid, $section_name, $section_text)
364        || warn ("Unable to write fulltxt row to db for row:\n\tOID $doc_oid, section $section_name");
365        # Execution failure will print out info anyway: since db connection sets PrintError
366    }
367    }
368   
369    # output all subsections: RECURSIVE CALL
370    foreach my $subsection (@{$section_ptr->{'subsection_order'}}) {
371    $self->recursive_write_meta_and_text($doc_obj, $doc_oid, "$section.$subsection", $metadata_table_sth, $fulltxt_table_sth);
372    }
373}
374
375
3761;
Note: See TracBrowser for help on using the browser.