1 | ###########################################################################
|
---|
2 | #
|
---|
3 | # dbutil.pm -- utility functions for writing to different databases
|
---|
4 | # Copyright (C) 2008 DL Consulting Ltd
|
---|
5 | #
|
---|
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 | # This program is free software; you can redistribute it and/or modify
|
---|
11 | # it under the terms of the GNU General Public License as published by
|
---|
12 | # the Free Software Foundation; either version 2 of the License, or
|
---|
13 | # (at your option) any later version.
|
---|
14 | #
|
---|
15 | # This program is distributed in the hope that it will be useful,
|
---|
16 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
17 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
18 | # GNU General Public License for more details.
|
---|
19 | #
|
---|
20 | # You should have received a copy of the GNU General Public License
|
---|
21 | # along with this program; if not, write to the Free Software
|
---|
22 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
23 | #
|
---|
24 | ###########################################################################
|
---|
25 |
|
---|
26 | package dbutil;
|
---|
27 |
|
---|
28 | use strict;
|
---|
29 |
|
---|
30 |
|
---|
31 | sub open_infodb_write_handle
|
---|
32 | {
|
---|
33 | my $infodb_type = shift(@_);
|
---|
34 | my $infodb_file_path = shift(@_);
|
---|
35 |
|
---|
36 | if ($infodb_type eq "sqlite")
|
---|
37 | {
|
---|
38 | return &open_infodb_write_handle_sqlite($infodb_file_path);
|
---|
39 | }
|
---|
40 | elsif ($infodb_type eq "gdbm-txtgz")
|
---|
41 | {
|
---|
42 | return &open_infodb_write_handle_gdbm_txtgz($infodb_file_path);
|
---|
43 | }
|
---|
44 |
|
---|
45 | # Use GDBM if the infodb type is empty or not one of the values above
|
---|
46 | return &open_infodb_write_handle_gdbm($infodb_file_path);
|
---|
47 | }
|
---|
48 |
|
---|
49 |
|
---|
50 | sub close_infodb_write_handle
|
---|
51 | {
|
---|
52 | my $infodb_type = shift(@_);
|
---|
53 | my $infodb_handle = shift(@_);
|
---|
54 |
|
---|
55 | if ($infodb_type eq "sqlite")
|
---|
56 | {
|
---|
57 | return &close_infodb_write_handle_sqlite($infodb_handle);
|
---|
58 | }
|
---|
59 | elsif ($infodb_type eq "gdbm-txtgz")
|
---|
60 | {
|
---|
61 | return &close_infodb_write_handle_gdbm_txtgz($infodb_handle);
|
---|
62 | }
|
---|
63 |
|
---|
64 | # Use GDBM if the infodb type is empty or not one of the values above
|
---|
65 | &close_infodb_write_handle_gdbm($infodb_handle);
|
---|
66 | }
|
---|
67 |
|
---|
68 |
|
---|
69 | sub get_default_infodb_type
|
---|
70 | {
|
---|
71 | # The default is GDBM so everything works the same for existing collections
|
---|
72 | # To use something else, specify the "infodbtype" in the collection's collect.cfg file
|
---|
73 | return "gdbm";
|
---|
74 | }
|
---|
75 |
|
---|
76 |
|
---|
77 | sub get_infodb_file_path
|
---|
78 | {
|
---|
79 | my $infodb_type = shift(@_);
|
---|
80 | my $collection_name = shift(@_);
|
---|
81 | my $infodb_directory_path = shift(@_);
|
---|
82 |
|
---|
83 | if ($infodb_type eq "sqlite")
|
---|
84 | {
|
---|
85 | return &get_infodb_file_path_sqlite($collection_name, $infodb_directory_path);
|
---|
86 | }
|
---|
87 | elsif ($infodb_type eq "gdbm-txtgz")
|
---|
88 | {
|
---|
89 | return &get_infodb_file_path_gdbm_txtgz($collection_name, $infodb_directory_path);
|
---|
90 | }
|
---|
91 |
|
---|
92 | # Use GDBM if the infodb type is empty or not one of the values above
|
---|
93 | return &get_infodb_file_path_gdbm($collection_name, $infodb_directory_path);
|
---|
94 | }
|
---|
95 |
|
---|
96 |
|
---|
97 | sub read_infodb_file
|
---|
98 | {
|
---|
99 | my $infodb_type = shift(@_);
|
---|
100 | my $infodb_file_path = shift(@_);
|
---|
101 | my $infodb_map = shift(@_);
|
---|
102 |
|
---|
103 | if ($infodb_type eq "sqlite")
|
---|
104 | {
|
---|
105 | return &read_infodb_file_sqlite($infodb_file_path, $infodb_map);
|
---|
106 | }
|
---|
107 | elsif ($infodb_type eq "gdbm-txtgz")
|
---|
108 | {
|
---|
109 | return &read_infodb_file_gdbm_txtgz($infodb_file_path, $infodb_map);
|
---|
110 | }
|
---|
111 |
|
---|
112 | # Use GDBM if the infodb type is empty or not one of the values above
|
---|
113 | return &read_infodb_file_gdbm($infodb_file_path, $infodb_map);
|
---|
114 | }
|
---|
115 |
|
---|
116 |
|
---|
117 | sub write_infodb_entry
|
---|
118 | {
|
---|
119 | my $infodb_type = shift(@_);
|
---|
120 | my $infodb_handle = shift(@_);
|
---|
121 | my $infodb_key = shift(@_);
|
---|
122 | my $infodb_map = shift(@_);
|
---|
123 |
|
---|
124 | if ($infodb_type eq "sqlite")
|
---|
125 | {
|
---|
126 | return &write_infodb_entry_sqlite($infodb_handle, $infodb_key, $infodb_map);
|
---|
127 | }
|
---|
128 | elsif ($infodb_type eq "gdbm-txtgz")
|
---|
129 | {
|
---|
130 | return &write_infodb_entry_gdbm_txtgz($infodb_handle, $infodb_key, $infodb_map);
|
---|
131 | }
|
---|
132 |
|
---|
133 | # Use GDBM if the infodb type is empty or not one of the values above
|
---|
134 | return &write_infodb_entry_gdbm($infodb_handle, $infodb_key, $infodb_map);
|
---|
135 | }
|
---|
136 |
|
---|
137 |
|
---|
138 |
|
---|
139 | # -----------------------------------------------------------------------------
|
---|
140 | # GDBM TXT-GZ IMPLEMENTATION
|
---|
141 | # -----------------------------------------------------------------------------
|
---|
142 |
|
---|
143 | sub open_infodb_write_handle_gdbm_txtgz
|
---|
144 | {
|
---|
145 | # Keep infodb in GDBM neutral form => save data as compressed text file,
|
---|
146 | # read for txt2db to be run on it later (i.e. by the runtime system,
|
---|
147 | # first time the collection is ever accessed). This makes it easier
|
---|
148 | # distribute pre-built collections to various architectures.
|
---|
149 | #
|
---|
150 | # NB: even if two architectures are little endian (e.g. Intel and
|
---|
151 | # ARM procesors) GDBM does *not* guarantee that the database generated on
|
---|
152 | # one will work on the other
|
---|
153 |
|
---|
154 | my $infodb_file_path = shift(@_);
|
---|
155 |
|
---|
156 | # Greenstone ships with gzip for windows, on $PATH
|
---|
157 |
|
---|
158 | my $infodb_file_handle = undef;
|
---|
159 | if (!open($infodb_file_handle, "| gzip - > \"$infodb_file_path\""))
|
---|
160 | {
|
---|
161 | return undef;
|
---|
162 | }
|
---|
163 |
|
---|
164 | return $infodb_file_handle;
|
---|
165 | }
|
---|
166 |
|
---|
167 |
|
---|
168 | sub close_infodb_write_handle_gdbm_txtgz
|
---|
169 | {
|
---|
170 | my $infodb_handle = shift(@_);
|
---|
171 |
|
---|
172 | close($infodb_handle);
|
---|
173 | }
|
---|
174 |
|
---|
175 |
|
---|
176 | sub get_infodb_file_path_gdbm_txtgz
|
---|
177 | {
|
---|
178 | my $collection_name = shift(@_);
|
---|
179 | my $infodb_directory_path = shift(@_);
|
---|
180 |
|
---|
181 | my $infodb_file_name = &util::get_dirsep_tail($collection_name).".txt.gz";
|
---|
182 | return &util::filename_cat($infodb_directory_path, $infodb_file_name);
|
---|
183 | }
|
---|
184 |
|
---|
185 |
|
---|
186 | sub read_infodb_file_gdbm_txtgz
|
---|
187 | {
|
---|
188 | my $infodb_file_path = shift(@_);
|
---|
189 | my $infodb_map = shift(@_);
|
---|
190 |
|
---|
191 | my $cmd = "gzip --decompress \"$infodb_file_path\"";
|
---|
192 |
|
---|
193 | open (PIPEIN, "$cmd |")
|
---|
194 | || die "Error: Couldn't open pipe from gzip: $!\n $cmd\n";
|
---|
195 |
|
---|
196 | my $infodb_line = "";
|
---|
197 | my $infodb_key = "";
|
---|
198 | my $infodb_value = "";
|
---|
199 | while (defined ($infodb_line = <PIPEIN>))
|
---|
200 | {
|
---|
201 | if ($infodb_line =~ /^\[([^\]]+)\]$/)
|
---|
202 | {
|
---|
203 | $infodb_key = $1;
|
---|
204 | }
|
---|
205 | elsif ($infodb_line =~ /^-{70}$/)
|
---|
206 | {
|
---|
207 | $infodb_map->{$infodb_key} = $infodb_value;
|
---|
208 | $infodb_key = "";
|
---|
209 | $infodb_value = "";
|
---|
210 | }
|
---|
211 | else
|
---|
212 | {
|
---|
213 | $infodb_value .= $infodb_line;
|
---|
214 | }
|
---|
215 | }
|
---|
216 |
|
---|
217 | close (PIPEIN);
|
---|
218 | }
|
---|
219 |
|
---|
220 |
|
---|
221 | sub write_infodb_entry_gdbm_txtgz
|
---|
222 | {
|
---|
223 |
|
---|
224 | my $infodb_handle = shift(@_);
|
---|
225 | my $infodb_key = shift(@_);
|
---|
226 | my $infodb_map = shift(@_);
|
---|
227 |
|
---|
228 | print $infodb_handle "[$infodb_key]\n";
|
---|
229 | foreach my $infodb_value_key (keys(%$infodb_map))
|
---|
230 | {
|
---|
231 | foreach my $infodb_value (@{$infodb_map->{$infodb_value_key}})
|
---|
232 | {
|
---|
233 | if ($infodb_value =~ /-{70,}/)
|
---|
234 | {
|
---|
235 | # if value contains 70 or more hyphens in a row we need to escape them
|
---|
236 | # to prevent txt2db from treating them as a separator
|
---|
237 | $infodb_value =~ s/-/&\#045;/gi;
|
---|
238 | }
|
---|
239 | print $infodb_handle "<$infodb_value_key>" . $infodb_value . "\n";
|
---|
240 | }
|
---|
241 | }
|
---|
242 | print $infodb_handle '-' x 70, "\n";
|
---|
243 | }
|
---|
244 |
|
---|
245 |
|
---|
246 |
|
---|
247 | # -----------------------------------------------------------------------------
|
---|
248 | # GDBM IMPLEMENTATION
|
---|
249 | # -----------------------------------------------------------------------------
|
---|
250 |
|
---|
251 | sub open_infodb_write_handle_gdbm
|
---|
252 | {
|
---|
253 | my $infodb_file_path = shift(@_);
|
---|
254 |
|
---|
255 | my $txt2db_exe = &util::filename_cat($ENV{'GSDLHOME'},"bin",$ENV{'GSDLOS'}, "txt2db" . &util::get_os_exe());
|
---|
256 | my $infodb_file_handle = undef;
|
---|
257 | if (!-e "$txt2db_exe" || !open($infodb_file_handle, "| \"$txt2db_exe\" \"$infodb_file_path\""))
|
---|
258 | {
|
---|
259 | return undef;
|
---|
260 | }
|
---|
261 |
|
---|
262 | return $infodb_file_handle;
|
---|
263 | }
|
---|
264 |
|
---|
265 |
|
---|
266 | sub close_infodb_write_handle_gdbm
|
---|
267 | {
|
---|
268 | my $infodb_handle = shift(@_);
|
---|
269 |
|
---|
270 | close($infodb_handle);
|
---|
271 | }
|
---|
272 |
|
---|
273 |
|
---|
274 | sub get_infodb_file_path_gdbm
|
---|
275 | {
|
---|
276 | my $collection_name = shift(@_);
|
---|
277 | my $infodb_directory_path = shift(@_);
|
---|
278 |
|
---|
279 | my $infodb_file_extension = (&util::is_little_endian() ? ".ldb" : ".bdb");
|
---|
280 | my $infodb_file_name = &util::get_dirsep_tail($collection_name) . $infodb_file_extension;
|
---|
281 | return &util::filename_cat($infodb_directory_path, $infodb_file_name);
|
---|
282 | }
|
---|
283 |
|
---|
284 |
|
---|
285 | sub read_infodb_file_gdbm
|
---|
286 | {
|
---|
287 | my $infodb_file_path = shift(@_);
|
---|
288 | my $infodb_map = shift(@_);
|
---|
289 |
|
---|
290 | open (PIPEIN, "db2txt \"$infodb_file_path\" |") || die "couldn't open pipe from db2txt\n";
|
---|
291 | my $infodb_line = "";
|
---|
292 | my $infodb_key = "";
|
---|
293 | my $infodb_value = "";
|
---|
294 | while (defined ($infodb_line = <PIPEIN>))
|
---|
295 | {
|
---|
296 | if ($infodb_line =~ /^\[([^\]]+)\]$/)
|
---|
297 | {
|
---|
298 | $infodb_key = $1;
|
---|
299 | }
|
---|
300 | elsif ($infodb_line =~ /^-{70}$/)
|
---|
301 | {
|
---|
302 | $infodb_map->{$infodb_key} = $infodb_value;
|
---|
303 | $infodb_key = "";
|
---|
304 | $infodb_value = "";
|
---|
305 | }
|
---|
306 | else
|
---|
307 | {
|
---|
308 | $infodb_value .= $infodb_line;
|
---|
309 | }
|
---|
310 | }
|
---|
311 |
|
---|
312 | close (PIPEIN);
|
---|
313 | }
|
---|
314 |
|
---|
315 |
|
---|
316 | sub write_infodb_entry_gdbm
|
---|
317 | {
|
---|
318 | # With infodb_handle already set up, works the same as _gdbm_txtgz version
|
---|
319 | write_infodb_entry_gdbm_txtgz(@_);
|
---|
320 | }
|
---|
321 |
|
---|
322 |
|
---|
323 |
|
---|
324 | # -----------------------------------------------------------------------------
|
---|
325 | # SQLITE IMPLEMENTATION
|
---|
326 | # -----------------------------------------------------------------------------
|
---|
327 |
|
---|
328 | sub open_infodb_write_handle_sqlite
|
---|
329 | {
|
---|
330 | my $infodb_file_path = shift(@_);
|
---|
331 |
|
---|
332 | my $sqlite3_exe = &util::filename_cat($ENV{'GSDLHOME'},"bin",$ENV{'GSDLOS'}, "sqlite3" . &util::get_os_exe());
|
---|
333 | my $infodb_handle = undef;
|
---|
334 | if (!-e "$sqlite3_exe" || !open($infodb_handle, "| \"$sqlite3_exe\" \"$infodb_file_path\""))
|
---|
335 | {
|
---|
336 | return undef;
|
---|
337 | }
|
---|
338 |
|
---|
339 | print $infodb_handle "CREATE TABLE IF NOT EXISTS data (key TEXT PRIMARY KEY, value TEXT);\n";
|
---|
340 | print $infodb_handle "CREATE TABLE IF NOT EXISTS document_metadata (id INTEGER PRIMARY KEY, docOID TEXT, element TEXT, value TEXT);\n";
|
---|
341 |
|
---|
342 | # This is crucial for efficiency when importing large amounts of data
|
---|
343 | print $infodb_handle "CREATE INDEX IF NOT EXISTS dmd ON document_metadata(docOID);\n";
|
---|
344 |
|
---|
345 | # This is very important for efficiency, otherwise each command will be actioned one at a time
|
---|
346 | print $infodb_handle "BEGIN TRANSACTION;\n";
|
---|
347 |
|
---|
348 | return $infodb_handle;
|
---|
349 | }
|
---|
350 |
|
---|
351 |
|
---|
352 | sub close_infodb_write_handle_sqlite
|
---|
353 | {
|
---|
354 | my $infodb_handle = shift(@_);
|
---|
355 |
|
---|
356 | # Close the transaction we began after opening the file
|
---|
357 | print $infodb_handle "END TRANSACTION;\n";
|
---|
358 |
|
---|
359 | # This is crucial for efficient queries on the database!
|
---|
360 | print $infodb_handle "CREATE INDEX IF NOT EXISTS dme ON document_metadata(element);\n";
|
---|
361 |
|
---|
362 | close($infodb_handle);
|
---|
363 | }
|
---|
364 |
|
---|
365 |
|
---|
366 | sub get_infodb_file_path_sqlite
|
---|
367 | {
|
---|
368 | my $collection_name = shift(@_);
|
---|
369 | my $infodb_directory_path = shift(@_);
|
---|
370 |
|
---|
371 | my $infodb_file_extension = ".db";
|
---|
372 | my $infodb_file_name = &util::get_dirsep_tail($collection_name) . $infodb_file_extension;
|
---|
373 | return &util::filename_cat($infodb_directory_path, $infodb_file_name);
|
---|
374 | }
|
---|
375 |
|
---|
376 |
|
---|
377 | sub read_infodb_file_sqlite
|
---|
378 | {
|
---|
379 | my $infodb_file_path = shift(@_);
|
---|
380 | my $infodb_map = shift(@_);
|
---|
381 |
|
---|
382 | # !! TO IMPLEMENT
|
---|
383 | }
|
---|
384 |
|
---|
385 |
|
---|
386 | sub write_infodb_entry_sqlite
|
---|
387 | {
|
---|
388 | my $infodb_handle = shift(@_);
|
---|
389 | my $infodb_key = shift(@_);
|
---|
390 | my $infodb_map = shift(@_);
|
---|
391 |
|
---|
392 | # Add the key -> value mapping into the "data" table
|
---|
393 | my $infodb_entry_value = "";
|
---|
394 | foreach my $infodb_value_key (keys(%$infodb_map))
|
---|
395 | {
|
---|
396 | foreach my $infodb_value (@{$infodb_map->{$infodb_value_key}})
|
---|
397 | {
|
---|
398 | $infodb_entry_value .= "<$infodb_value_key>" . $infodb_value . "\n";
|
---|
399 | }
|
---|
400 | }
|
---|
401 |
|
---|
402 | my $safe_infodb_key = &sqlite_safe($infodb_key);
|
---|
403 | print $infodb_handle "INSERT OR REPLACE INTO data (key, value) VALUES ('" . $safe_infodb_key . "', '" . &sqlite_safe($infodb_entry_value) . "');\n";
|
---|
404 |
|
---|
405 | # If this infodb entry is for a document, add all the interesting document metadata to the
|
---|
406 | # "document_metadata" table (for use by the dynamic classifiers)
|
---|
407 | if ($infodb_key !~ /\./ && $infodb_entry_value =~ /\<doctype\>doc\n/)
|
---|
408 | {
|
---|
409 | print $infodb_handle "DELETE FROM document_metadata WHERE docOID='" . $safe_infodb_key . "';\n";
|
---|
410 |
|
---|
411 | foreach my $infodb_value_key (keys(%$infodb_map))
|
---|
412 | {
|
---|
413 | # We're not interested in most of the automatically added document metadata
|
---|
414 | next if ($infodb_value_key eq "archivedir" ||
|
---|
415 | $infodb_value_key eq "assocfilepath" ||
|
---|
416 | $infodb_value_key eq "childtype" ||
|
---|
417 | $infodb_value_key eq "contains" ||
|
---|
418 | $infodb_value_key eq "docnum" ||
|
---|
419 | $infodb_value_key eq "doctype" ||
|
---|
420 | $infodb_value_key eq "Encoding" ||
|
---|
421 | $infodb_value_key eq "FileSize" ||
|
---|
422 | $infodb_value_key eq "hascover" ||
|
---|
423 | $infodb_value_key eq "hastxt" ||
|
---|
424 | $infodb_value_key eq "lastmodified" ||
|
---|
425 | $infodb_value_key eq "metadataset" ||
|
---|
426 | $infodb_value_key eq "thistype" ||
|
---|
427 | $infodb_value_key =~ /^metadatafreq\-/ ||
|
---|
428 | $infodb_value_key =~ /^metadatalist\-/);
|
---|
429 |
|
---|
430 | foreach my $infodb_value (@{$infodb_map->{$infodb_value_key}})
|
---|
431 | {
|
---|
432 | print $infodb_handle "INSERT INTO document_metadata (docOID, element, value) VALUES ('" . $safe_infodb_key . "', '" . &sqlite_safe($infodb_value_key) . "', '" . &sqlite_safe($infodb_value) . "');\n";
|
---|
433 | }
|
---|
434 | }
|
---|
435 | }
|
---|
436 | }
|
---|
437 |
|
---|
438 |
|
---|
439 | sub sqlite_safe
|
---|
440 | {
|
---|
441 | my $value = shift(@_);
|
---|
442 |
|
---|
443 | # Escape any single quotes in the value
|
---|
444 | $value =~ s/\'/\'\'/g;
|
---|
445 |
|
---|
446 | return $value;
|
---|
447 | }
|
---|
448 |
|
---|
449 |
|
---|
450 | 1;
|
---|