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 |
|
---|
28 | package GreenstoneSQLPlugout;
|
---|
29 |
|
---|
30 | use strict;
|
---|
31 | no strict 'refs';
|
---|
32 | no strict 'subs';
|
---|
33 |
|
---|
34 | use GreenstoneXMLPlugout;
|
---|
35 | use docprint;
|
---|
36 | use gsmysql;
|
---|
37 |
|
---|
38 | use DBI; # the central package for this plugout
|
---|
39 |
|
---|
40 |
|
---|
41 | # This plugout does not output the metadata and/or fulltxt xml to a file,
|
---|
42 | # but outputs rows into a MySQL db table for metadata and/or a db table for fulltxt
|
---|
43 |
|
---|
44 | sub BEGIN {
|
---|
45 | @GreenstoneSQLPlugout::ISA = ('GreenstoneXMLPlugout');
|
---|
46 | }
|
---|
47 |
|
---|
48 |
|
---|
49 | my $process_mode_list =
|
---|
50 | [ { 'name' => "meta_only",
|
---|
51 | 'desc' => "{GreenstoneSQLPlug.process_mode.meta_only}" },
|
---|
52 | { 'name' => "text_only",
|
---|
53 | 'desc' => "{GreenstoneSQLPlug.process_mode.text_only}" },
|
---|
54 | { 'name' => "all",
|
---|
55 | 'desc' => "{GreenstoneSQLPlug.process_mode.all}" } ];
|
---|
56 |
|
---|
57 | my $rollback_on_cancel_list =
|
---|
58 | [ { 'name' => "true",
|
---|
59 | 'desc' => "{GreenstoneSQLPlug.rollback_on_cancel}" },
|
---|
60 | { 'name' => "false",
|
---|
61 | 'desc' => "{GreenstoneSQLPlug.rollbacl_on_cancel}" } ];
|
---|
62 |
|
---|
63 | # The following are the saveas.options:
|
---|
64 | my $arguments = [
|
---|
65 | { 'name' => "process_mode",
|
---|
66 | 'desc' => "{GreenstoneSQLPlug.process_mode}",
|
---|
67 | 'type' => "enum",
|
---|
68 | 'list' => $process_mode_list,
|
---|
69 | 'deft' => "all",
|
---|
70 | 'reqd' => "no",
|
---|
71 | 'hiddengli' => "no"},
|
---|
72 | { 'name' => "rollback_on_cancel",
|
---|
73 | 'desc' => "{GreenstoneSQLPlug.rollback_on_cancel}",
|
---|
74 | 'type' => "enum",
|
---|
75 | 'list' => $rollback_on_cancel_list,
|
---|
76 | 'deft' => "false", # better default than true
|
---|
77 | 'reqd' => "no",
|
---|
78 | 'hiddengli' => "no"},
|
---|
79 | { 'name' => "db_driver",
|
---|
80 | 'desc' => "{GreenstoneSQLPlug.db_driver}",
|
---|
81 | 'type' => "string",
|
---|
82 | 'deft' => "mysql",
|
---|
83 | 'reqd' => "yes"},
|
---|
84 | { 'name' => "db_client_user",
|
---|
85 | 'desc' => "{GreenstoneSQLPlug.db_client_user}",
|
---|
86 | 'type' => "string",
|
---|
87 | 'deft' => "root",
|
---|
88 | 'reqd' => "yes"},
|
---|
89 | { 'name' => "db_client_pwd",
|
---|
90 | 'desc' => "{GreenstoneSQLPlug.db_client_pwd}",
|
---|
91 | 'type' => "string",
|
---|
92 | 'deft' => "",
|
---|
93 | 'reqd' => "no"}, # pwd not required: can create mysql accounts without pwd
|
---|
94 | { 'name' => "db_host",
|
---|
95 | 'desc' => "{GreenstoneSQLPlug.db_host}",
|
---|
96 | 'type' => "string",
|
---|
97 | 'deft' => "127.0.0.1", # localhost doesn't work for us, but 127.0.0.1 works. See gsmysql.pm
|
---|
98 | 'reqd' => "yes"},
|
---|
99 | { 'name' => "db_port",
|
---|
100 | 'desc' => "{GreenstoneSQLPlug.db_port}",
|
---|
101 | 'type' => "string", # NOTE: make this int? No default for port, since it's not a required connection param
|
---|
102 | 'reqd' => "no"}
|
---|
103 | ];
|
---|
104 |
|
---|
105 | my $options = { 'name' => "GreenstoneSQLPlugout",
|
---|
106 | 'desc' => "{GreenstoneSQLPlugout.desc}",
|
---|
107 | 'abstract' => "no",
|
---|
108 | 'inherits' => "yes",
|
---|
109 | 'args' => $arguments };
|
---|
110 |
|
---|
111 | ##### This entire class is called only during import.pl #####
|
---|
112 |
|
---|
113 | ##### Overridden methods #####
|
---|
114 |
|
---|
115 | sub new {
|
---|
116 | my ($class) = shift (@_);
|
---|
117 | my ($plugoutlist, $inputargs,$hashArgOptLists) = @_;
|
---|
118 | push(@$plugoutlist, $class);
|
---|
119 |
|
---|
120 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
---|
121 | push(@{$hashArgOptLists->{"OptList"}},$options);
|
---|
122 |
|
---|
123 | my $self = new GreenstoneXMLPlugout($plugoutlist,$inputargs,$hashArgOptLists);
|
---|
124 |
|
---|
125 | if ($self->{'info_only'}) {
|
---|
126 | # don't worry about any options etc
|
---|
127 | return bless $self, $class;
|
---|
128 | }
|
---|
129 |
|
---|
130 | return bless $self, $class;
|
---|
131 | }
|
---|
132 |
|
---|
133 | # connect here and ensure all tables and databases exist
|
---|
134 | sub begin {
|
---|
135 |
|
---|
136 | my $self= shift (@_);
|
---|
137 |
|
---|
138 | my $db_params = {
|
---|
139 | 'collection_name' => $ENV{'GSDLCOLLECTION'},
|
---|
140 | 'verbosity' => $self->{'verbosity'} || 0
|
---|
141 |
|
---|
142 | };
|
---|
143 |
|
---|
144 | my $gs_sql = new gsmysql($db_params);
|
---|
145 |
|
---|
146 | # if autocommit is set, there's no rollback support
|
---|
147 | my $autocommit = ($self->{'rollback_on_cancel'} eq "false") ? 1 : 0;
|
---|
148 |
|
---|
149 | # try connecting to the mysql db, die if that fails
|
---|
150 | # So don't bother preparing GreenstoneXMLPlugout by calling superclass' begin() yet
|
---|
151 | if(!$gs_sql->connect_to_db({
|
---|
152 | 'db_driver' => $self->{'db_driver'},
|
---|
153 | 'db_client_user' => $self->{'db_client_user'},
|
---|
154 | 'db_client_pwd' => $self->{'db_client_pwd'},
|
---|
155 | 'db_host' => $self->{'db_host'},
|
---|
156 | 'db_port' => $self->{'db_port'}, # undef by default, can leave as is
|
---|
157 | 'autocommit' => $autocommit
|
---|
158 | })
|
---|
159 | )
|
---|
160 | {
|
---|
161 | # This is fatal for the plugout, let's terminate here
|
---|
162 | # PrintError would already have displayed the warning message on connection fail
|
---|
163 | die("Could not connect to db. Can't proceed.\n");
|
---|
164 | }
|
---|
165 |
|
---|
166 | my $db_name = $self->{'site'} || "greenstone2"; # one database per GS3 site, for GS2 the db is called greenstone2
|
---|
167 | my $proc_mode = $self->{'process_mode'};
|
---|
168 |
|
---|
169 |
|
---|
170 | my $success = $gs_sql->use_db($db_name);
|
---|
171 |
|
---|
172 | if($success && $proc_mode ne "text_only") {
|
---|
173 | $success = $gs_sql->ensure_meta_table_exists();
|
---|
174 | }
|
---|
175 | if($success && $proc_mode ne "meta_only") {
|
---|
176 | $success = $gs_sql->ensure_fulltxt_table_exists();
|
---|
177 | }
|
---|
178 |
|
---|
179 | if(!$success) {
|
---|
180 | # This is fatal for the plugout, let's terminate here after disconnecting again
|
---|
181 | # PrintError would already have displayed the warning message on load fail
|
---|
182 | # And on die() perl will call gsmysql destroy which will ensure a disconnect() from db
|
---|
183 | die("Could not use db $db_name and/or prepare its tables. Can't proceed.\n");
|
---|
184 | }
|
---|
185 |
|
---|
186 | # store the DBI wrapper instance
|
---|
187 | $self->{'gs_sql'} = $gs_sql;
|
---|
188 |
|
---|
189 |
|
---|
190 | # If setting up connection to sql db failed, we'd have terminated and wouldn't come up to here
|
---|
191 | # and wouldn't have bothered preparing GreenstoneXMLPlugout by calling superclass' begin().
|
---|
192 | # Finally, can call begin on super - important as doc.xml is opened as a group etc
|
---|
193 |
|
---|
194 | $self->SUPER::begin(@_);
|
---|
195 | }
|
---|
196 |
|
---|
197 | # disconnect from database here, see inexport.pm
|
---|
198 | sub end
|
---|
199 | {
|
---|
200 | my $self = shift(@_);
|
---|
201 |
|
---|
202 | # do the superclass stuff first, as any sql db failures should not prevent superclass cleanup
|
---|
203 | $self->SUPER::end(@_);
|
---|
204 |
|
---|
205 | # Important to call finished():
|
---|
206 | # it will disconnect from db if this is the last gsmysql instance,
|
---|
207 | # and it will commit to db before disconnecting if rollbback_on_cancel turned on
|
---|
208 | $self->{'gs_sql'}->finished();
|
---|
209 | delete $self->{'gs_sql'}; # key gs_sql no longer exists, not just the value being undef
|
---|
210 | }
|
---|
211 |
|
---|
212 | # Produce files called docsql.xml instead of doc.xml
|
---|
213 | sub get_doc_xml_filename {
|
---|
214 | my $self = shift (@_);
|
---|
215 | my ($doc_obj) = @_;
|
---|
216 |
|
---|
217 | return "docsql.xml";
|
---|
218 | }
|
---|
219 |
|
---|
220 | # overriding to store doc OID as attribute of top level element: <Archive docoid="oid">
|
---|
221 | sub output_xml_header {
|
---|
222 | my $self = shift (@_);
|
---|
223 | my ($outhandle, $doc_oid) = @_;
|
---|
224 |
|
---|
225 | print $outhandle '<?xml version="1.0" encoding="utf-8" standalone="no"?>' . "\n";
|
---|
226 | print $outhandle "<!DOCTYPE Archive SYSTEM \"http://greenstone.org/dtd/Archive/1.0/Archive.dtd\">\n";
|
---|
227 | print $outhandle "<Archive docoid=\"$doc_oid\">\n";
|
---|
228 | }
|
---|
229 |
|
---|
230 | # saveas() only generates the content in archives dir and in the SQL database
|
---|
231 | sub saveas {
|
---|
232 | my $self = shift (@_);
|
---|
233 | my ($doc_obj, $doc_dir) = @_;
|
---|
234 |
|
---|
235 | my $proc_mode = $self->{'process_mode'};
|
---|
236 |
|
---|
237 | # 1. pre save out and saving debug handle
|
---|
238 |
|
---|
239 | # must call superclass (pre/post) saveas methods, as they handle assoc_files too
|
---|
240 | my ($docxml_outhandler, $output_file) = $self->SUPER::pre_saveas(@_);
|
---|
241 |
|
---|
242 | $self->{'debug_outhandle'} = $docxml_outhandler if ($self->{'debug'}); # STDOUT if debug
|
---|
243 |
|
---|
244 | # 2. overriding saving behaviour to do what the superclass does (writing out doc.xml files,
|
---|
245 | # under new name of docsql.xml, with breadcrumbs pointing to sql db) PLUS saving to sql db
|
---|
246 |
|
---|
247 | # NOTE: if proc_mode == all, then "breadcrumbs" (statements pointing viewer to the sql db
|
---|
248 | # for contents) go into both meta and txt elements of doc.xml (docsql.xml specifically):
|
---|
249 |
|
---|
250 | # write the INVERSE into doc.xml as to what is written to the SQL db
|
---|
251 | my $docxml_output_options = { 'output' => docprint::OUTPUT_NONE };
|
---|
252 | if($proc_mode eq "meta_only" ) { # since only meta to go into MySQL db, text will go into docxml
|
---|
253 | $docxml_output_options->{'output'} = docprint::OUTPUT_TEXT_ONLY;
|
---|
254 | } elsif($proc_mode eq "text_only" ) { # since only full text to go into MySQL db, meta will go into docxml
|
---|
255 | $docxml_output_options->{'output'} = docprint::OUTPUT_META_ONLY;
|
---|
256 | }
|
---|
257 |
|
---|
258 | # now we've prepared to write out whatever is meant to go into docxml
|
---|
259 | # and can do actual the steps superclass GreenstoneXMLPlugout carries out to write out docxml
|
---|
260 | # So: write out the doc xml file, "docsql.xml", for the current document
|
---|
261 | my $section_text = &docprint::get_section_xml($doc_obj, $docxml_output_options);
|
---|
262 | print $docxml_outhandler $section_text;
|
---|
263 |
|
---|
264 |
|
---|
265 | # We also write out whatever needs to go into the MySQL database
|
---|
266 | $self->write_meta_and_text($doc_obj);
|
---|
267 |
|
---|
268 |
|
---|
269 | # 3. post save out
|
---|
270 | $self->SUPER::post_saveas($doc_obj, $doc_dir, $docxml_outhandler, $output_file);
|
---|
271 |
|
---|
272 |
|
---|
273 | # database connection is closed once, in end() method
|
---|
274 | # We're not opening and closing over and over for each doc during a single build
|
---|
275 | }
|
---|
276 |
|
---|
277 | ##### New methods, not inherited #####
|
---|
278 |
|
---|
279 | # write meta and/or text PER DOC out to DB
|
---|
280 | sub write_meta_and_text {
|
---|
281 | my $self = shift (@_);
|
---|
282 | my ($doc_obj) = @_;
|
---|
283 | my $doc_oid = $doc_obj->get_OID(); # this method processes a single doc at a time, so it uses the same OID throughout
|
---|
284 | my $root_section = $doc_obj->get_top_section();
|
---|
285 |
|
---|
286 | $self->recursive_write_meta_and_text($doc_obj, $doc_oid, $root_section);
|
---|
287 | }
|
---|
288 |
|
---|
289 | sub recursive_write_meta_and_text {
|
---|
290 | my $self = shift (@_);
|
---|
291 | my ($doc_obj, $doc_oid, $section) = @_;
|
---|
292 |
|
---|
293 | # If section=ROOT, write "root" as section name into table
|
---|
294 | # doc->get_top_section() is the name of the doc root section, which is ""
|
---|
295 | my $section_name = ($section eq "") ? "root" : $section;
|
---|
296 |
|
---|
297 | my $section_ptr = $doc_obj->_lookup_section ($section);
|
---|
298 | return "" unless defined $section_ptr;
|
---|
299 |
|
---|
300 | my $debug_out = $self->{'debug_outhandle'};
|
---|
301 |
|
---|
302 | my $gs_sql = $self->{'gs_sql'};
|
---|
303 | my $proc_mode = $self->{'process_mode'};
|
---|
304 | if($proc_mode eq "all" || $proc_mode eq "meta_only" ) {
|
---|
305 |
|
---|
306 | foreach my $data (@{$section_ptr->{'metadata'}}) {
|
---|
307 | my $meta_name = $data->[0];
|
---|
308 |
|
---|
309 | # Treat db like a text file instead of an html/xml file: don't need to escape text
|
---|
310 | # going into it, unlike with doc(sql).xml
|
---|
311 | my $meta_value = $data->[1];
|
---|
312 |
|
---|
313 | # Write out the current section's meta to collection db's METADATA table
|
---|
314 |
|
---|
315 | # For each set of values to write to meta table, this next method call will
|
---|
316 | # efficiently execute an insert SQL statement (using a prepared insert statement),
|
---|
317 | # filling in the values
|
---|
318 | # OR if debugging, then it will print the SQL insert statement but not execute it
|
---|
319 | # (a behaviour following what the GS XML Plugout superclass does on debug)
|
---|
320 |
|
---|
321 | $gs_sql->insert_row_into_metadata_table($doc_oid, $section_name, $meta_name, $meta_value, $self->{'debug'});
|
---|
322 | }
|
---|
323 | }
|
---|
324 |
|
---|
325 |
|
---|
326 | if($proc_mode eq "all" || $proc_mode eq "text_only" ) {
|
---|
327 |
|
---|
328 | # See above, no need to html-escape for db
|
---|
329 | my $section_text = $section_ptr->{'text'};
|
---|
330 |
|
---|
331 | # fulltxt column can be SQL NULL. undef value for $section_text gets written out as NULL:
|
---|
332 | # https://stackoverflow.com/questions/12708633/which-one-represents-null-undef-or-empty-string
|
---|
333 | # The following will do the SQL insertion
|
---|
334 | # or if debug, the following will print the SQL insert stmt without executing it
|
---|
335 | $gs_sql->insert_row_into_fulltxt_table($doc_oid, $section_name, \$section_text, $self->{'debug'});
|
---|
336 |
|
---|
337 | }
|
---|
338 |
|
---|
339 | # output all subsections: RECURSIVE CALL
|
---|
340 | foreach my $subsection (@{$section_ptr->{'subsection_order'}}) {
|
---|
341 | $self->recursive_write_meta_and_text($doc_obj, $doc_oid, "$section.$subsection");
|
---|
342 | }
|
---|
343 | }
|
---|
344 |
|
---|
345 |
|
---|
346 | 1;
|
---|