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 gssql; |
---|
37 | |
---|
38 | use 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 | #X TODO Q: introduced site param to plugins and plugouts. Did I do it right? And should they have hiddengli = "yes". No longer a param |
---|
44 | # Did I do the pass by ref in docprint's escape and unescape textref functions correctly, and how they're called here? |
---|
45 | # Any more optimisation I can do around this? |
---|
46 | |
---|
47 | # this plugout does not output the metadata and/or fulltxt xml to a file, |
---|
48 | # but outputs rows into a mysql table for metadata and/or a table for fulltxt |
---|
49 | sub BEGIN { |
---|
50 | @GreenstoneSQLPlugout::ISA = ('GreenstoneXMLPlugout'); |
---|
51 | } |
---|
52 | |
---|
53 | # + NOTTODO: die() statements need to be replaced with premature_termination |
---|
54 | # which should ensure the GreenstoneXMLPlugin (group)'s stuff is closed and cleaned up SOMEHOW |
---|
55 | # It's fine: the die() stmts all take place before setting up the super class' begin |
---|
56 | |
---|
57 | # + TODO Q: about build_mode: how to detect removeold. Now handled by |
---|
58 | # GreenstoneSQLPlugout::remove_all(), which is inherited from a base plugin. |
---|
59 | # + TODO: deal with -removeold and everything? Or type out instructions for user |
---|
60 | |
---|
61 | # + TODO Q: what is "group" in GreenstoneXMLPlugout? |
---|
62 | |
---|
63 | my $process_mode_list = |
---|
64 | [ { 'name' => "meta_only", |
---|
65 | 'desc' => "{GreenstoneSQLPlug.process_mode.meta_only}" }, |
---|
66 | { 'name' => "text_only", |
---|
67 | 'desc' => "{GreenstoneSQLPlug.process_mode.text_only}" }, |
---|
68 | { 'name' => "all", |
---|
69 | 'desc' => "{GreenstoneSQLPlug.process_mode.all}" } ]; |
---|
70 | |
---|
71 | my $rollback_on_cancel_list = |
---|
72 | [ { 'name' => "true", |
---|
73 | 'desc' => "{GreenstoneSQLPlug.rollback_on_cancel}" }, |
---|
74 | { 'name' => "false", |
---|
75 | 'desc' => "{GreenstoneSQLPlug.rollbacl_on_cancel}" } ]; |
---|
76 | |
---|
77 | # The following are the saveas.options: |
---|
78 | my $arguments = [ |
---|
79 | { 'name' => "process_mode", |
---|
80 | 'desc' => "{GreenstoneSQLPlug.process_mode}", |
---|
81 | 'type' => "enum", |
---|
82 | 'list' => $process_mode_list, |
---|
83 | 'deft' => "all", |
---|
84 | 'reqd' => "no", |
---|
85 | 'hiddengli' => "no"}, |
---|
86 | { 'name' => "rollback_on_cancel", |
---|
87 | 'desc' => "{GreenstoneSQLPlug.rollback_on_cancel}", |
---|
88 | 'type' => "enum", |
---|
89 | 'list' => $rollback_on_cancel_list, |
---|
90 | 'deft' => "false", # TODO Q: what's the better default? If "true", any memory concerns? |
---|
91 | 'reqd' => "no", |
---|
92 | 'hiddengli' => "no"}, |
---|
93 | { 'name' => "db_driver", |
---|
94 | 'desc' => "{GreenstoneSQLPlug.db_driver}", |
---|
95 | 'type' => "string", |
---|
96 | 'deft' => "mysql", |
---|
97 | 'reqd' => "yes"}, |
---|
98 | { 'name' => "db_client_user", |
---|
99 | 'desc' => "{GreenstoneSQLPlug.db_client_user}", |
---|
100 | 'type' => "string", |
---|
101 | 'deft' => "root", |
---|
102 | 'reqd' => "yes"}, |
---|
103 | { 'name' => "db_client_pwd", |
---|
104 | 'desc' => "{GreenstoneSQLPlug.db_client_pwd}", |
---|
105 | 'type' => "string", |
---|
106 | 'deft' => "", |
---|
107 | 'reqd' => "no"}, # pwd required? NO. |
---|
108 | { 'name' => "db_host", |
---|
109 | 'desc' => "{GreenstoneSQLPlug.db_host}", |
---|
110 | 'type' => "string", |
---|
111 | 'deft' => "127.0.0.1", |
---|
112 | 'reqd' => "yes"}, |
---|
113 | { 'name' => "db_port", |
---|
114 | 'desc' => "{GreenstoneSQLPlug.db_port}", |
---|
115 | 'type' => "string", # NOTE: make this int? No default for port, since it's not a required connection param |
---|
116 | 'reqd' => "no"} |
---|
117 | ]; |
---|
118 | |
---|
119 | my $options = { 'name' => "GreenstoneSQLPlugout", |
---|
120 | 'desc' => "{GreenstoneSQLPlugout.desc}", |
---|
121 | 'abstract' => "no", |
---|
122 | 'inherits' => "yes", |
---|
123 | 'args' => $arguments }; |
---|
124 | |
---|
125 | ##### This entire class is called only during import.pl ##### |
---|
126 | |
---|
127 | ##### Overridden methods ##### |
---|
128 | |
---|
129 | sub new { |
---|
130 | my ($class) = shift (@_); |
---|
131 | my ($plugoutlist, $inputargs,$hashArgOptLists) = @_; |
---|
132 | push(@$plugoutlist, $class); |
---|
133 | |
---|
134 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments}); |
---|
135 | push(@{$hashArgOptLists->{"OptList"}},$options); |
---|
136 | |
---|
137 | my $self = new GreenstoneXMLPlugout($plugoutlist,$inputargs,$hashArgOptLists); |
---|
138 | |
---|
139 | if ($self->{'info_only'}) { |
---|
140 | # don't worry about any options etc |
---|
141 | return bless $self, $class; |
---|
142 | } |
---|
143 | #print STDERR "***** GreenstoneSQLPlugout process mode = \"", $self->{'process_mode'}, "\"\n"; |
---|
144 | |
---|
145 | return bless $self, $class; |
---|
146 | } |
---|
147 | |
---|
148 | # connect here and ensure all tables and databases exist |
---|
149 | sub begin { |
---|
150 | |
---|
151 | my $self= shift (@_); |
---|
152 | |
---|
153 | # The saveas.options |
---|
154 | #print STDERR "@@@@ PLUGOUT db_pwd: " . $self->{'db_client_pwd'} . "\n"; |
---|
155 | #print STDERR "@@@@ user: " . $self->{'db_client_user'} . "\n"; |
---|
156 | #print STDERR "@@@@ db_host: " . $self->{'db_host'} . "\n"; |
---|
157 | #print STDERR "@@@@ db_driver: " . $self->{'db_driver'} . "\n"; |
---|
158 | #print STDERR "@@@@ proc_mode: " . $self->{'process_mode'} . "\n"; |
---|
159 | |
---|
160 | ############ LOAD NECESSARY OPTIONS ########### |
---|
161 | #print "@@@ plugout SITE NAME: ". $self->{'site'} . "\n" if defined $self->{'site'}; |
---|
162 | #print STDERR "########## COLLECTION: ". $ENV{'GSDLCOLLECTION'}."\n"; |
---|
163 | |
---|
164 | print STDERR "***** GreenstoneSQLPlugout process mode = \"", $self->{'process_mode'}, "\"\n"; |
---|
165 | |
---|
166 | my $db_params = { |
---|
167 | 'collection_name' => $ENV{'GSDLCOLLECTION'}, |
---|
168 | 'verbosity' => $self->{'verbosity'} || 0 |
---|
169 | |
---|
170 | }; |
---|
171 | |
---|
172 | my $gs_sql = new gssql($db_params); |
---|
173 | |
---|
174 | # if autocommit is set, there's no rollback support |
---|
175 | my $autocommit = ($self->{'rollback_on_cancel'} eq "false") ? 1 : 0; |
---|
176 | |
---|
177 | # try connecting to the mysql db, die if that fails |
---|
178 | # So don't bother preparing GreenstoneXMLPlugout by calling superclass' begin() yet |
---|
179 | if(!$gs_sql->connect_to_db({ |
---|
180 | 'db_driver' => $self->{'db_driver'}, |
---|
181 | 'db_client_user' => $self->{'db_client_user'}, |
---|
182 | 'db_client_pwd' => $self->{'db_client_pwd'}, |
---|
183 | 'db_host' => $self->{'db_host'}, |
---|
184 | 'db_port' => $self->{'db_port'}, # undef by default, can leave as is |
---|
185 | 'autocommit' => $autocommit |
---|
186 | }) |
---|
187 | ) |
---|
188 | { |
---|
189 | # This is fatal for the plugout, let's terminate here |
---|
190 | # PrintError would already have displayed the warning message on connection fail |
---|
191 | die("Could not connect to db. Can't proceed.\n"); |
---|
192 | } |
---|
193 | |
---|
194 | #die("@@@@ TEST. Connected successfully. Testing gssql::destructor.\n"); # WORKS |
---|
195 | |
---|
196 | my $db_name = $self->{'site'} || "greenstone2"; # one database per GS3 site, for GS2 the db is called greenstone2 |
---|
197 | my $proc_mode = $self->{'process_mode'}; |
---|
198 | |
---|
199 | |
---|
200 | my $success = $gs_sql->use_db($db_name); |
---|
201 | |
---|
202 | if($success && $proc_mode ne "text_only") { |
---|
203 | ##print STDERR "@@@@ Ensuring meta table exists\n"; |
---|
204 | $success = $gs_sql->ensure_meta_table_exists(); |
---|
205 | } |
---|
206 | if($success && $proc_mode ne "meta_only") { |
---|
207 | ##print STDERR "@@@@ Ensuring fulltxt table exists\n"; |
---|
208 | $success = $gs_sql->ensure_fulltxt_table_exists(); |
---|
209 | } |
---|
210 | |
---|
211 | if(!$success) { |
---|
212 | # This is fatal for the plugout, let's terminate here after disconnecting again |
---|
213 | # PrintError would already have displayed the warning message on load fail |
---|
214 | # And on die() perl will call gssql destroy which will ensure a disconnect() from db |
---|
215 | #$gs_sql->force_disconnect_from_db(); # disconnect_from_db() will issue a warning on error |
---|
216 | die("Could not use db $db_name and/or prepare its tables. Can't proceed.\n"); |
---|
217 | } |
---|
218 | |
---|
219 | # store the DBI wrapper instance |
---|
220 | $self->{'gs_sql'} = $gs_sql; |
---|
221 | |
---|
222 | |
---|
223 | # if setting up to work with sql db failed, we'd have terminated and wouldn't come up to here: |
---|
224 | # won't bother preparing GreenstoneXMLPlugout by calling superclass' begin() |
---|
225 | # finally, can call begin on super - important as doc.xml is opened as a group etc |
---|
226 | |
---|
227 | $self->SUPER::begin(@_); |
---|
228 | } |
---|
229 | |
---|
230 | # disconnect from database here, see inexport.pm |
---|
231 | sub end |
---|
232 | { |
---|
233 | my $self = shift(@_); |
---|
234 | |
---|
235 | # do the superclass stuff first, as any sql db failures should not prevent superclass cleanup |
---|
236 | $self->SUPER::end(@_); |
---|
237 | |
---|
238 | # Important to call finished(): |
---|
239 | # it will disconnect from db if this is the last gssql instance, |
---|
240 | # and it will commit to db before disconnecting if rollbback_on_cancel turned on |
---|
241 | $self->{'gs_sql'}->finished(); |
---|
242 | delete $self->{'gs_sql'}; # key gs_sql no longer exists, not just the value being undef |
---|
243 | } |
---|
244 | |
---|
245 | # Produce files called docsql.xml instead of doc.xml |
---|
246 | sub get_doc_xml_filename { |
---|
247 | my $self = shift (@_); |
---|
248 | my ($doc_obj) = @_; |
---|
249 | |
---|
250 | return "docsql.xml"; |
---|
251 | } |
---|
252 | |
---|
253 | # overriding to store doc OID as attribute of top level element: <Archive docoid="oid"> |
---|
254 | sub output_xml_header { |
---|
255 | my $self = shift (@_); |
---|
256 | my ($outhandle, $doc_oid) = @_; |
---|
257 | |
---|
258 | print $outhandle '<?xml version="1.0" encoding="utf-8" standalone="no"?>' . "\n"; |
---|
259 | print $outhandle "<!DOCTYPE Archive SYSTEM \"http://greenstone.org/dtd/Archive/1.0/Archive.dtd\">\n"; |
---|
260 | print $outhandle "<Archive docoid=\"$doc_oid\">\n"; |
---|
261 | } |
---|
262 | |
---|
263 | # + X TODO: check arc-inf.db for whether each entry is to be deleted/indexed/reindexed/been indexed? |
---|
264 | # That's only for indexing, not for this step which only generates the content in archives dir |
---|
265 | sub saveas { |
---|
266 | my $self = shift (@_); |
---|
267 | my ($doc_obj, $doc_dir) = @_; |
---|
268 | |
---|
269 | my $proc_mode = $self->{'process_mode'}; |
---|
270 | |
---|
271 | # 1. pre save out and saving debug handle |
---|
272 | |
---|
273 | # must call superclass (pre/post) saveas methods, as they handle assoc_files too |
---|
274 | my ($docxml_outhandler, $output_file) = $self->SUPER::pre_saveas(@_); |
---|
275 | |
---|
276 | $self->{'debug_outhandle'} = $docxml_outhandler if ($self->{'debug'}); # STDOUT if debug |
---|
277 | |
---|
278 | # TODO: also set debugging in begin()? Then stmts creating db and tables also sent to debug out and not executed |
---|
279 | |
---|
280 | |
---|
281 | # 2. overriding saving behaviour to do what the superclass does (writing out doc.xml files, |
---|
282 | # under new name of docsql.xml, with breadcrumbs pointing to sql db) PLUS saving to sql db |
---|
283 | |
---|
284 | # NOTE: if proc_mode == all, then "breadcrumbs" (statements pointing viewer to the sql db |
---|
285 | # for contents) go into both meta and txt elements of doc.xml (docsql.xml specifically): |
---|
286 | |
---|
287 | # write the INVERSE into doc.xml as to what is written to the SQL db |
---|
288 | my $docxml_output_options = { 'output' => docprint::OUTPUT_NONE }; |
---|
289 | if($proc_mode eq "meta_only" ) { # since only meta to go into MySQL db, text will go into docxml |
---|
290 | $docxml_output_options->{'output'} = docprint::OUTPUT_TEXT_ONLY; |
---|
291 | } elsif($proc_mode eq "text_only" ) { # since only full text to go into MySQL db, meta will go into docxml |
---|
292 | $docxml_output_options->{'output'} = docprint::OUTPUT_META_ONLY; |
---|
293 | } |
---|
294 | |
---|
295 | # now we've prepared to write out whatever is meant to go into docxml |
---|
296 | # and can do actual the steps superclass GreenstoneXMLPlugout carries out to write out docxml |
---|
297 | # So: write out the doc xml file, "docsql.xml", for the current document |
---|
298 | my $section_text = &docprint::get_section_xml($doc_obj, $docxml_output_options); |
---|
299 | print $docxml_outhandler $section_text; |
---|
300 | |
---|
301 | |
---|
302 | # We also write out whatever needs to go into the MySQL database |
---|
303 | $self->write_meta_and_text($doc_obj); |
---|
304 | |
---|
305 | |
---|
306 | # 3. post save out |
---|
307 | $self->SUPER::post_saveas($doc_obj, $doc_dir, $docxml_outhandler, $output_file); |
---|
308 | |
---|
309 | |
---|
310 | # database connection is closed in end() method |
---|
311 | # so we don't open and close over and over for each doc during a single build |
---|
312 | } |
---|
313 | |
---|
314 | ##### New methods, not inherited ##### |
---|
315 | |
---|
316 | # write meta and/or text PER DOC out to DB |
---|
317 | sub write_meta_and_text { |
---|
318 | my $self = shift (@_); |
---|
319 | my ($doc_obj) = @_; |
---|
320 | my $doc_oid = $doc_obj->get_OID(); # this method processes a single doc at a time, so it uses the same OID throughout |
---|
321 | my $root_section = $doc_obj->get_top_section(); |
---|
322 | |
---|
323 | $self->recursive_write_meta_and_text($doc_obj, $doc_oid, $root_section); |
---|
324 | } |
---|
325 | |
---|
326 | # Perl: Reading or Writing to Another Program |
---|
327 | # https://nnc3.com/mags/Perl3/cookbook/ch16_05.htm |
---|
328 | sub recursive_write_meta_and_text { |
---|
329 | my $self = shift (@_); |
---|
330 | my ($doc_obj, $doc_oid, $section) = @_; |
---|
331 | |
---|
332 | # If section=ROOT, write "root" as section name into table |
---|
333 | # doc->get_top_section() is the name of the doc root section, which is "" |
---|
334 | my $section_name = ($section eq "") ? "root" : $section; |
---|
335 | |
---|
336 | my $section_ptr = $doc_obj->_lookup_section ($section); |
---|
337 | return "" unless defined $section_ptr; |
---|
338 | |
---|
339 | my $debug_out = $self->{'debug_outhandle'}; |
---|
340 | |
---|
341 | my $gs_sql = $self->{'gs_sql'}; |
---|
342 | my $proc_mode = $self->{'process_mode'}; |
---|
343 | if($proc_mode eq "all" || $proc_mode eq "meta_only" ) { |
---|
344 | |
---|
345 | foreach my $data (@{$section_ptr->{'metadata'}}) { |
---|
346 | my $meta_name = $data->[0]; |
---|
347 | # TODO: does it need to be stored escaped, as it requires unescaping when read back in |
---|
348 | # from db (unlike for reading back in from doc.xml) |
---|
349 | my $escaped_meta_value = &docprint::escape_text($data->[1]); |
---|
350 | |
---|
351 | # Write out the current section's meta to collection db's METADATA table |
---|
352 | |
---|
353 | # For each set of values to write to meta table, this next method call will |
---|
354 | # efficiently execute an insert SQL statement (using a prepared insert statement), |
---|
355 | # filling in the values |
---|
356 | # OR if debugging, then it will print the SQL insert statement but not execute it |
---|
357 | |
---|
358 | $gs_sql->insert_row_into_metadata_table($doc_oid, $section_name, $meta_name, $escaped_meta_value, $self->{'debug'}); |
---|
359 | } |
---|
360 | } |
---|
361 | |
---|
362 | |
---|
363 | if($proc_mode eq "all" || $proc_mode eq "text_only" ) { |
---|
364 | |
---|
365 | my $section_textref = &docprint::escape_textref(\$section_ptr->{'text'}); |
---|
366 | |
---|
367 | # fulltxt column can be SQL NULL. undef value gets written out as NULL: |
---|
368 | # https://stackoverflow.com/questions/12708633/which-one-represents-null-undef-or-empty-string |
---|
369 | # The following will do the SQL insertion |
---|
370 | # or if debug, the following will print the SQL insert stmt without executing it |
---|
371 | $gs_sql->insert_row_into_fulltxt_table($doc_oid, $section_name, $section_textref, $self->{'debug'}); |
---|
372 | |
---|
373 | } |
---|
374 | |
---|
375 | # output all subsections: RECURSIVE CALL |
---|
376 | foreach my $subsection (@{$section_ptr->{'subsection_order'}}) { |
---|
377 | $self->recursive_write_meta_and_text($doc_obj, $doc_oid, "$section.$subsection"); |
---|
378 | } |
---|
379 | } |
---|
380 | |
---|
381 | |
---|
382 | 1; |
---|