[31190] | 1 | # This class based on arcinfo.pm
|
---|
| 2 | package oaiinfo;
|
---|
| 3 |
|
---|
| 4 | use constant INFO_STATUS_INDEX => 0;
|
---|
| 5 | use constant INFO_TIMESTAMP_INDEX => 1;
|
---|
| 6 |
|
---|
| 7 | use strict;
|
---|
| 8 |
|
---|
| 9 | use arcinfo;
|
---|
| 10 | use dbutil;
|
---|
| 11 |
|
---|
| 12 | # QUESTIONS:
|
---|
| 13 | # Should we use time or localtime(time) for timestamp? Just timestamp.
|
---|
| 14 | # What format should the timestamp be in, or is the basic format used by perl sufficient? Basic.
|
---|
| 15 |
|
---|
| 16 | # File format read in: OID <tab> Date-timestamp <tab> Deletion-Status
|
---|
| 17 |
|
---|
| 18 | # Deletion status can be:
|
---|
| 19 | # E = Doc with OID exists (has not been deleted from collection). Timestamp indicates last time of build
|
---|
| 20 | # D = Doc with OID has been deleted. Timestamp indicates time of deletion
|
---|
| 21 | # PD = Provisionally Deleted. Timestamp momentarily unaltered.
|
---|
| 22 |
|
---|
| 23 | # oaidb is "always incremental": always reflects the I/B/R/D status of archive info db,
|
---|
| 24 | # before the indexing step of the build phase that alters the I/B/R/D contents of archive info db.
|
---|
| 25 | # (I=index, B=been indexed, R=reindex; D=delete)
|
---|
| 26 |
|
---|
| 27 | sub new {
|
---|
| 28 | my $class = shift(@_);
|
---|
| 29 | my ($config_filename, $infodbtype) = @_;
|
---|
| 30 |
|
---|
| 31 | my $self = {
|
---|
| 32 | 'info'=>{} # map of {OID, array[deletion-status,timestamp]} pairs
|
---|
| 33 | };
|
---|
| 34 |
|
---|
| 35 | if(!defined $infodbtype) {
|
---|
| 36 | $infodbtype = &dbutil::get_default_infodb_type();
|
---|
| 37 | }
|
---|
| 38 | $infodbtype = "gdbm" if ($infodbtype eq "gdbm-txtgz");
|
---|
| 39 | $self->{'infodbtype'} = $infodbtype;
|
---|
| 40 |
|
---|
| 41 | # Create and store the db filenames we'll be working with (tmp and livedb)
|
---|
| 42 | my $etc_dir = &util::get_parent_folder($config_filename);
|
---|
| 43 |
|
---|
| 44 | my $perform_firsttime_init = 0;
|
---|
| 45 | $self->{'oaidb_live_filepath'} = &dbutil::get_infodb_file_path($infodbtype, "oai-inf", $etc_dir, $perform_firsttime_init);
|
---|
| 46 | $self->{'oaidb_tmp_filepath'} = &dbutil::get_infodb_file_path($infodbtype, "oai-inf-tmp", $etc_dir, $perform_firsttime_init);
|
---|
| 47 | $self->{'etc_dir'} = $etc_dir;
|
---|
| 48 | # print STDERR "############ LIVE DB: $self->{'oaidb_live_filepath'}\n";
|
---|
| 49 | # print STDERR "############ TMP DB: $self->{'oaidb_tmp_filepath'}\n";
|
---|
| 50 |
|
---|
| 51 | $self->{'oaidb_file_path'} = $self->{'oaidb_tmp_filepath'}; # db file we're working with
|
---|
| 52 |
|
---|
| 53 | return bless $self, $class;
|
---|
| 54 | }
|
---|
| 55 |
|
---|
| 56 | # this subroutine will work out the starting contents of the tmp-db (temporary oai db):
|
---|
| 57 | # whether it should start off empty, or with the contents of any existing live-db,
|
---|
| 58 | # or with the contents of any existing tmp-db.
|
---|
| 59 | sub init_tmpdb {
|
---|
| 60 | my $self = shift(@_);
|
---|
| 61 | my ($removeold, $have_manifest) = @_;
|
---|
| 62 |
|
---|
| 63 | # if we have a manifest file, then we pretend we are fully incremental for oaiinfo db.
|
---|
| 64 | # removeold implies proper full-rebuild, whereas keepold or incremental means incremental
|
---|
| 65 | if($have_manifest) { # if we have a manifest file, we're not doing removeold/full-rebuild either
|
---|
| 66 | $removeold = 0;
|
---|
| 67 | }
|
---|
| 68 |
|
---|
[31192] | 69 | my $do_pd_step = ($removeold) ? 1 : 0;
|
---|
| 70 | # if $removeold, then proper full rebuild, will carry out step where all E will be marked as PD
|
---|
| 71 | # else some kind of incremental build, won't do the extra PD pass
|
---|
| 72 | # which is the step marking existing OIDs (E) as PD (provisionally deleted)
|
---|
[31190] | 73 |
|
---|
| 74 | my $oaidb_live_filepath = $self->{'oaidb_live_filepath'};
|
---|
| 75 | my $oaidb_tmp_filepath = $self->{'oaidb_tmp_filepath'};
|
---|
| 76 | my $infodbtype = $self->{'infodbtype'};
|
---|
| 77 | # Note: the live db can only exist if the collection has been activated at least once before
|
---|
| 78 | my $livedb_exists = &FileUtils::fileExists($oaidb_live_filepath);
|
---|
| 79 | my $tmpdb_exists = &FileUtils::fileExists($oaidb_tmp_filepath);
|
---|
| 80 |
|
---|
| 81 | my $initdb = 0;
|
---|
| 82 |
|
---|
| 83 | # work out what operation we need to do
|
---|
| 84 | # work with empty tmpdb
|
---|
| 85 | # copy_livedb_to_tmpdb
|
---|
| 86 | # work with existing tmpdb (so existing tmpdb will be topped up)
|
---|
| 87 |
|
---|
| 88 | # make_contents_of_tmpdb_empty
|
---|
| 89 | # make_contents_of_tmpdb_that_of_livedb
|
---|
| 90 | # continue_working_with_tmpdb ("contents_of_tmpdb_is_tmpdb")
|
---|
| 91 |
|
---|
| 92 | # We're going to prepare the starting state of tmpdb next.
|
---|
| 93 | # It can start off empty, start off with the contents of livedb, or it can start off with the contents
|
---|
| 94 | # of the existing tmp db. Which of these three it is depends on the 3 factors: whether livedb exists,
|
---|
| 95 | # whether tmpdb exists and whether or not removeold is true.
|
---|
| 96 | # i.o.w. which of the 3 outcomes it is depends on the truth table built on the following 3 variables:
|
---|
| 97 | # LDB = LiveDB exists
|
---|
| 98 | # TDB = TmpDB exists
|
---|
| 99 | # RO = Removeold
|
---|
| 100 | # OUTCOMES:
|
---|
| 101 | # clean slate (create an empty tmpdb/make tmpdb empty)
|
---|
| 102 | # top up tmpDB (work with existing tmpdb)
|
---|
| 103 | # copy LiveDB to TmpDB (liveDB's contents become the contents of TmpDB, and we'll work with that)
|
---|
| 104 | #
|
---|
| 105 | # TRUTH TABLE:
|
---|
| 106 | # ---------------------------------------
|
---|
| 107 | # LDB TDB RO | Outcome
|
---|
| 108 | # ---------------------------------------
|
---|
| 109 | # 0 0 0 | clean-slate
|
---|
| 110 | # 0 0 1 | clean-slate
|
---|
| 111 | # 0 1 0 | top-up-tmpdb
|
---|
| 112 | # 0 1 1 | erase tmpdb, clean-slate
|
---|
| 113 | # 1 0 0 | copy livedb to tmpdb
|
---|
| 114 | # 1 0 1 | copy livedb to tmpdb
|
---|
| 115 | # 1 1 0 | top-up-tmpdb
|
---|
| 116 | # 1 1 1 | copy livedb to tmpd
|
---|
| 117 | # ---------------------------------------
|
---|
| 118 | #
|
---|
| 119 | # Dr Bainbridge worked out using Karnaugh maps that, from the above truth table:
|
---|
| 120 | # => clean-slate/empty-tmpdb = !LDB && (RO || !TDB)
|
---|
| 121 | # => top-up-tmpdb/work-with-existing-tmpdb = !RO && TDB
|
---|
| 122 | # => copy-livedb-to-tmpdb = LDB && (!TDB || RO)
|
---|
| 123 | # I had most of these tests, except that I hadn't (yet) merged the two clean slate instances
|
---|
| 124 | # of first-build-ever and make-contents-of-tmpdb-empty
|
---|
| 125 |
|
---|
| 126 | #my $first_build_ever = (!$livedb_exists && !$tmpdb_exists);
|
---|
| 127 | #my $make_contents_of_tmpdb_empty = (!$livedb_exists && $tmpdb_exists && $removeold);
|
---|
| 128 | # Karnaugh map allows merging $first_build_ever and $make_contents_of_tmpdb_empty above
|
---|
| 129 | # into: my $work_with_empty_tmpdb = (!$livedb_exists && (!$tmpdb_exists || $removeold));
|
---|
| 130 | my $work_with_empty_tmpdb = (!$livedb_exists && (!$tmpdb_exists || $removeold));
|
---|
| 131 | my $make_contents_of_tmpdb_that_of_livedb = ($livedb_exists && (!$tmpdb_exists || $removeold));
|
---|
| 132 | my $work_with_existing_tmpdb = ($tmpdb_exists && !$removeold);
|
---|
| 133 |
|
---|
| 134 | if($work_with_empty_tmpdb) { # we'll use an empty tmpdb
|
---|
[31192] | 135 |
|
---|
[31190] | 136 | # If importing the collection for the very first time, neither db exists,
|
---|
| 137 | # so create an empty tmpdb.
|
---|
| 138 | #
|
---|
| 139 | # We also create an empty tmpdb when livedb doesn't exist and $removeold is true.
|
---|
| 140 | # This can happen if we've never run activate (so no livedb),
|
---|
| 141 | # yet had done some import (and perhaps building) followed by a full re-import now.
|
---|
| 142 | # Since there was no activate and we're doing a removeold/full-rebuild now, can just
|
---|
| 143 | # work with a new tmpdb, even though one already existed, its contents can be wiped out.
|
---|
| 144 | # In such a scenario, we'll be deleting tmpdb. Then there will be no livedb nor any tmpdb
|
---|
| 145 | # any more, so same situation as if importing the very first time when no oaidb exists either.
|
---|
| 146 |
|
---|
| 147 | &dbutil::remove_db_file($self->{'infodbtype'}, $oaidb_tmp_filepath) if $tmpdb_exists; # remove the db file and any assoc files
|
---|
| 148 | $initdb = 1; # new tmpdb
|
---|
| 149 |
|
---|
| 150 | # if the oai db is created the first time, it's like incremental and
|
---|
| 151 | # "keepold" (keepold means "only add, don't reprocess existing"). So
|
---|
| 152 | # no need to do the special passes dealing with "provisional deletes".
|
---|
[31192] | 153 | $do_pd_step = 0;
|
---|
[31190] | 154 |
|
---|
| 155 | } elsif ($make_contents_of_tmpdb_that_of_livedb) {
|
---|
| 156 |
|
---|
| 157 | # If the livedb exists and we're doing a full rebuild ($removeold is true),
|
---|
| 158 | # copy livedb to tmp regardless of if tmpdb already exists.
|
---|
| 159 | # Or if the livedb exists and tmpdb doesn't exist, it doesn't matter
|
---|
| 160 | # if we're incremental or not: also copy live to tmp and work with tmp.
|
---|
| 161 |
|
---|
| 162 | # copy livedb to tmpdb
|
---|
| 163 | &dbutil::remove_db_file($self->{'infodbtype'}, $oaidb_tmp_filepath) if $tmpdb_exists; # remove the db file and any assoc files
|
---|
| 164 | &FileUtils::copyFiles($oaidb_live_filepath, $oaidb_tmp_filepath);
|
---|
| 165 |
|
---|
| 166 | $initdb = 0; # tmpdb exists, since we just copied livedb to tmpdb, so need to init existing tmpdb
|
---|
| 167 |
|
---|
| 168 | } else { # $work_with_existing_tmpdb, so we'll build on top of what's presently already in tmpdb
|
---|
| 169 | # (we'll be topping up the current tmpdb)
|
---|
| 170 |
|
---|
| 171 | # !$removeold, meaning incremental
|
---|
| 172 | # If incremental and have a tmpdb already, regardless of whether livedb exists,
|
---|
| 173 | # then work with the existing tmpdb file, as this means we've been
|
---|
| 174 | # importing (perhaps followed by building) repeatedly without activating the
|
---|
| 175 | # last time but want to maintain the (incremental) changes in tmpdb.
|
---|
| 176 |
|
---|
| 177 | $initdb = 0;
|
---|
| 178 |
|
---|
| 179 | } # Dr Bainbridge drew up Karnaugh maps on the truth table, which proved that all cases
|
---|
| 180 | # are indeed covered above, so don't need any other catch-all else here
|
---|
| 181 |
|
---|
| 182 | $self->{'oaidb_file_path'} = &dbutil::get_infodb_file_path($infodbtype, "oai-inf-tmp", $self->{'etc_dir'}, $initdb);
|
---|
| 183 | # final param follows jmt's $perform_firsttime_init in inexport.pm
|
---|
| 184 |
|
---|
| 185 | # print STDERR "@@@@@ oaidb: $self->{'oaidb_file_path'}\n";
|
---|
| 186 |
|
---|
[31192] | 187 | return $do_pd_step;
|
---|
[31190] | 188 | }
|
---|
| 189 |
|
---|
| 190 | sub get_filepath {
|
---|
| 191 | my $self = shift (@_);
|
---|
| 192 | return $self->{'oaidb_file_path'};
|
---|
| 193 | }
|
---|
| 194 |
|
---|
| 195 | sub import_stage {
|
---|
| 196 | my $self = shift (@_);
|
---|
| 197 | my ($removeold, $have_manifest) = @_;
|
---|
| 198 |
|
---|
[31191] | 199 | my $do_pd_step = $self->init_tmpdb($removeold, $have_manifest);
|
---|
[31192] | 200 | # returns 1 if the step to mark oaidb entries as PD is required
|
---|
| 201 | # if we're doing full rebuilding and it's NOT the first time creating the oai_inf db,
|
---|
| 202 | # then the tasks to do with PD (provisionally deleted) OAI OIDs should be carried out
|
---|
[31190] | 203 |
|
---|
| 204 | $self->load_info();
|
---|
| 205 | $self->print_info(); # DEBUGGING
|
---|
| 206 |
|
---|
| 207 | if ($do_pd_step) {
|
---|
| 208 | $self->mark_all_existing_as_provisionallydeleted();
|
---|
| 209 | $self->print_info(); # DEBUGGING
|
---|
| 210 |
|
---|
| 211 | # save to db file now that we're done
|
---|
| 212 | $self->save_info();
|
---|
| 213 | }
|
---|
| 214 |
|
---|
| 215 | }
|
---|
| 216 |
|
---|
| 217 | sub building_stage_before_indexing() {
|
---|
| 218 | my $self = shift (@_);
|
---|
| 219 | my ($archivedir) = @_;
|
---|
| 220 |
|
---|
| 221 | # load archive info db into memory
|
---|
| 222 | my $arcinfo_doc_filename = &dbutil::get_infodb_file_path($self->{'infodbtype'}, "archiveinf-doc", $archivedir);
|
---|
| 223 | my $arcinfo_src_filename = &dbutil::get_infodb_file_path($self->{'infodbtype'}, "archiveinf-src", $archivedir);
|
---|
| 224 | my $archive_info = new arcinfo ($self->{'infodbtype'});
|
---|
| 225 | $archive_info->load_info ($arcinfo_doc_filename);
|
---|
| 226 |
|
---|
| 227 | #my $started_from_scratch = &FileUtils::fileTest($self->{'oaidb_tmp_filepath'}, '-z'); # 1 if tmpdb is empty
|
---|
| 228 | # -z test for file is empty http://www.perlmonks.org/?node_id=927447
|
---|
| 229 |
|
---|
| 230 | # load the oaidb file's contents into memory.
|
---|
| 231 | $self->load_info();
|
---|
| 232 | $self->print_info(); # DEBUGGING
|
---|
| 233 |
|
---|
| 234 | # process all the index, reindex and delete operations as indicated in arcinfo,
|
---|
| 235 | # all the while ensuring all PDs are changed back to E for OIDs that exist in both arcinfo and oaiinfo db.
|
---|
| 236 |
|
---|
| 237 | foreach my $OID (keys $archive_info->{'info'}) {
|
---|
| 238 | my $arcinf_tuple = $archive_info->{'info'}->{$OID};
|
---|
| 239 | my $indexing_status = $arcinf_tuple->[arcinfo::INFO_STATUS_INDEX];
|
---|
| 240 | # use packageName::constant to refer to constants declared in another package,
|
---|
| 241 | # see http://perldoc.perl.org/constant.html
|
---|
| 242 |
|
---|
| 243 | print STDERR "######## OID: $OID - status: $indexing_status\n";
|
---|
| 244 |
|
---|
| 245 | if($indexing_status eq "I") {
|
---|
| 246 | $self->index($OID); # add new as E with current timestamp/or set existing as E with orig timestamp
|
---|
| 247 | } elsif($indexing_status eq "R") {
|
---|
| 248 | $self->reindex($OID); # update timestamp and ensure marked as E (if oid doesn't exist, add new)
|
---|
| 249 | } elsif($indexing_status eq "D") {
|
---|
| 250 | $self->delete($OID); # set as D with current timestamp
|
---|
| 251 | } elsif($indexing_status eq "B") { # B for "been indexed"
|
---|
| 252 | $self->been_indexed($OID); # will flip any PD to E if oid exists, else will add new entry for oid
|
---|
| 253 | # A new entry may be required if the collection had been built prior to turning this into
|
---|
| 254 | # an oaicollection. But what if we always maintain an oaidb? Still call $self->index() here.
|
---|
| 255 | } else {
|
---|
| 256 | print STDERR "### oaiinfo::building_stage_before_indexing(): Unrecognised indexing status $indexing_status\n";
|
---|
| 257 | }
|
---|
| 258 | }
|
---|
| 259 |
|
---|
| 260 | # once all docs processed, go through oaiiinfo db changing any PDs to D along with current timestamp
|
---|
| 261 | # to indicate that they're deleted
|
---|
| 262 | $self->mark_all_provisionallydeleted_as_deleted();
|
---|
| 263 | $self->print_info();
|
---|
| 264 |
|
---|
| 265 | # let's save to db file now that we're done
|
---|
| 266 | $self->save_info();
|
---|
| 267 |
|
---|
| 268 | }
|
---|
| 269 |
|
---|
| 270 | sub activate_collection { # move tmp db to live db
|
---|
| 271 | my $self = shift (@_);
|
---|
| 272 |
|
---|
| 273 | my $oaidb_live_filepath = $self->{'oaidb_live_filepath'};
|
---|
| 274 | my $oaidb_tmp_filepath = $self->{'oaidb_tmp_filepath'};
|
---|
| 275 |
|
---|
| 276 | my $livedb_exists = &FileUtils::fileExists($oaidb_live_filepath);
|
---|
| 277 | my $tmpdb_exists = &FileUtils::fileExists($oaidb_tmp_filepath);
|
---|
| 278 |
|
---|
| 279 | if($tmpdb_exists) {
|
---|
| 280 | if($livedb_exists) {
|
---|
| 281 | #&dbutil::remove_db_file($self->{'infodbtype'}, s$oaidb_live_filepath); # remove the db file and any assoc files
|
---|
| 282 | &dbutil::rename_db_file_to($self->{'infodbtype'}, $oaidb_live_filepath, $oaidb_live_filepath.".bak"); # rename the db file and any assoc files
|
---|
| 283 | }
|
---|
| 284 | #&FileUtils::moveFiles($oaidb_tmp_filepath, $oaidb_live_filepath);
|
---|
| 285 | &dbutil::rename_db_file_to($self->{'infodbtype'}, $oaidb_tmp_filepath, $oaidb_live_filepath); # rename the db file and any assoc files
|
---|
| 286 |
|
---|
| 287 | print STDERR "#### Should now have MOVED $self->{'oaidb_tmp_filepath'} to $self->{'oaidb_live_filepath'}\n";
|
---|
| 288 |
|
---|
| 289 | } else {
|
---|
| 290 | print STDERR "@@@@@ In oaiinfo::activate_collection():\n";
|
---|
| 291 | print STDERR "@@@@@ No tmpdb at $self->{'oaidb_tmp_filepath'}\n";
|
---|
| 292 | print STDERR "@@@@@ to make 'live' by moving to $self->{'oaidb_live_filepath'}.\n";
|
---|
| 293 | }
|
---|
| 294 | }
|
---|
| 295 |
|
---|
| 296 | ##################### SPECIFIC TO PD-STEP ####################
|
---|
| 297 |
|
---|
| 298 |
|
---|
| 299 | # mark all existing, E (non-deleted) OIDs as Provisionally Deleted (PD)
|
---|
| 300 | # this subroutine doesn't save to oai-inf.DB
|
---|
| 301 | # the caller should call save_info when they want to save to the db
|
---|
| 302 | sub mark_all_existing_as_provisionallydeleted {
|
---|
| 303 | my $self = shift (@_);
|
---|
| 304 |
|
---|
| 305 | print STDERR "@@@@@ oaiinfo::mark_all_E_as_PD(): Marking the E entries as PD\n";
|
---|
| 306 |
|
---|
| 307 | foreach my $OID (keys $self->{'info'}) {
|
---|
[31192] | 308 | my $OID_info = $self->{'info'}->{$OID};
|
---|
| 309 | my $curr_status = $OID_info->[INFO_STATUS_INDEX];
|
---|
| 310 | if($curr_status eq "E") {
|
---|
| 311 | $OID_info->[INFO_STATUS_INDEX] = "PD";
|
---|
[31190] | 312 | }
|
---|
| 313 | }
|
---|
| 314 | }
|
---|
| 315 |
|
---|
| 316 | # mark all OIDs that are Provisionally Deleted (PD) as deleted, and set to current timestamp
|
---|
| 317 | # To be called at end of build. Again, the caller should save to DB by calling save_info.
|
---|
| 318 | sub mark_all_provisionallydeleted_as_deleted {
|
---|
| 319 | my $self = shift (@_);
|
---|
| 320 |
|
---|
| 321 | print STDERR "@@@@@ oaiinfo::mark_all_PD_as_D(): Marking the PD entries as D\n";
|
---|
| 322 |
|
---|
| 323 | foreach my $OID (keys $self->{'info'}) {
|
---|
[31192] | 324 | my $OID_info = $self->{'info'}->{$OID};
|
---|
| 325 | my $curr_status = $OID_info->[INFO_STATUS_INDEX];
|
---|
| 326 | if($curr_status eq "PD") {
|
---|
[31190] | 327 | $self->set_info($OID, "D", $self->get_current_time());
|
---|
| 328 | }
|
---|
| 329 | }
|
---|
| 330 | }
|
---|
| 331 |
|
---|
| 332 |
|
---|
| 333 | ##################### GENERAL, NOT SPECIFIC TO PD-STEP ####################
|
---|
| 334 |
|
---|
| 335 | sub print_info {
|
---|
| 336 | my $self = shift (@_);
|
---|
| 337 |
|
---|
| 338 | print STDERR "###########################################################\n";
|
---|
| 339 | print STDERR "@@@@@ oaiinfo::print_info(): oaidb in memory contains: \n";
|
---|
| 340 |
|
---|
| 341 | foreach my $OID (keys $self->{'info'}) {
|
---|
| 342 | print STDERR "OID: $OID";
|
---|
| 343 | print STDERR " status: " . $self->{'info'}->{$OID}->[INFO_STATUS_INDEX];
|
---|
| 344 | print STDERR " time: " . $self->{'info'}->{$OID}->[INFO_TIMESTAMP_INDEX];
|
---|
| 345 | print STDERR "\n";
|
---|
| 346 | }
|
---|
| 347 |
|
---|
| 348 | print STDERR "###########################################################\n";
|
---|
| 349 | }
|
---|
| 350 |
|
---|
| 351 |
|
---|
| 352 | # Find the OID, if it exists, make its status=E for existing. Leave its timestamp alone.
|
---|
| 353 | # If the OID doesn't yet exist, add it as a new entry with status=E and with current timestamp.
|
---|
| 354 | sub index { # Add a new oid with current time and E. If the oid was already present, mark as E
|
---|
| 355 | my $self = shift (@_);
|
---|
| 356 | my ($OID) = @_;
|
---|
| 357 |
|
---|
| 358 | my $OID_info = $self->{'info'}->{$OID};
|
---|
| 359 |
|
---|
| 360 | if (defined $OID_info) { # if OID is present, this will change status back to E, timestamp unchanged
|
---|
| 361 | $OID_info->[INFO_STATUS_INDEX] = "E";
|
---|
| 362 |
|
---|
| 363 | } else { # if OID is not present, then it's now added as existing from current time on
|
---|
| 364 | $self->set_info($OID, "E", $self->get_current_time());
|
---|
| 365 | }
|
---|
| 366 | }
|
---|
| 367 |
|
---|
| 368 | # Upon reindexing a document with identifier OID, change its timestamp to current time
|
---|
| 369 | # if a new OID, then add as new entry with status=E and current timestamp
|
---|
| 370 | sub reindex { # update timestamp if oid is already present, if not (unlikely), add as new
|
---|
| 371 | my $self = shift (@_);
|
---|
| 372 | my ($OID) = @_;
|
---|
| 373 |
|
---|
| 374 | my $OID_info = $self->{'info'}->{$OID};
|
---|
| 375 | $self->set_info($OID, "E", $self->get_current_time()); # Takes care of 3 things:
|
---|
| 376 | # if OID exists, updates modified time to indicate the doc has been reindexed
|
---|
| 377 | # if OID exists, ensures any status=PD is flipped back to E for this OID doc (as we know it exists);
|
---|
| 378 | # if the OID doesn't yet exist, adds a new OID entry with status=E and current timestamp.
|
---|
| 379 |
|
---|
| 380 | }
|
---|
| 381 |
|
---|
| 382 | # Does the same as index():
|
---|
| 383 | # OIDs that have been indexed upon rebuild may still be new to the oaidb: GS2 collections
|
---|
| 384 | # are not OAI collections by default, unlike GS3 collections. Imagine rebuilding a (GS2) collection
|
---|
| 385 | # 5 times and then setting them to be an OAI collection. In that case, the doc OIDs in the collection
|
---|
| 386 | # may not be in the oaidb yet. Unless, we decide (as is the present case) to always maintain an oaidb
|
---|
| 387 | # (always creating an oaidb regardless of whether the collection has OAI support turned on or not).
|
---|
| 388 | sub been_indexed {
|
---|
| 389 | my $self = shift (@_);
|
---|
| 390 | my ($OID) = @_;
|
---|
| 391 |
|
---|
| 392 | $self->index($OID);
|
---|
| 393 | }
|
---|
| 394 |
|
---|
| 395 | # Upon deleting a document with identifier OID,
|
---|
| 396 | # set status to deleted and change its timestamp to current time
|
---|
| 397 | sub delete {
|
---|
| 398 | my $self = shift (@_);
|
---|
| 399 | my ($OID) = @_;
|
---|
| 400 |
|
---|
| 401 | # the following method will set to current time if no timestamp provided,
|
---|
| 402 | # But by explicit here, the code is easier to follow
|
---|
| 403 | $self->set_info($OID, "D", $self->get_current_time());
|
---|
| 404 |
|
---|
| 405 | }
|
---|
| 406 |
|
---|
| 407 | #############################################################
|
---|
| 408 | sub get_current_time {
|
---|
| 409 | my $self = shift (@_);
|
---|
| 410 | return time; # current time
|
---|
| 411 |
|
---|
| 412 | # localtime(time) returns an array of values (day, month, year, hour, min, seconds) or singular string
|
---|
| 413 | # return localtime; # same as localtime(time); # http://perldoc.perl.org/functions/localtime.html
|
---|
| 414 |
|
---|
| 415 | }
|
---|
| 416 |
|
---|
| 417 | sub _load_info_txt
|
---|
| 418 | {
|
---|
| 419 | my $self = shift (@_);
|
---|
| 420 | my ($filename) = @_;
|
---|
| 421 |
|
---|
| 422 | if (defined $filename && &FileUtils::fileExists($filename)) {
|
---|
| 423 | open (INFILE, $filename) ||
|
---|
| 424 | die "oaiinfo::load_info couldn't read $filename\n";
|
---|
| 425 |
|
---|
| 426 | my ($line, @lineparts);
|
---|
| 427 | while (defined ($line = <INFILE>)) {
|
---|
| 428 | $line =~ s/\cM|\cJ//g; # remove end-of-line characters
|
---|
| 429 | @lineparts = split ("\t", $line);
|
---|
| 430 | if (scalar(@lineparts) >= 2) {
|
---|
[31192] | 431 | $self->set_info (@lineparts);
|
---|
[31190] | 432 | }
|
---|
| 433 | }
|
---|
| 434 | close (INFILE);
|
---|
| 435 | }
|
---|
| 436 |
|
---|
| 437 | }
|
---|
| 438 |
|
---|
| 439 | sub _load_info_db
|
---|
| 440 | {
|
---|
| 441 | my $self = shift (@_);
|
---|
| 442 | my ($filename) = @_;
|
---|
| 443 |
|
---|
| 444 | my $infodb_map = {};
|
---|
| 445 |
|
---|
| 446 | &dbutil::read_infodb_file($self->{'infodbtype'}, $filename, $infodb_map);
|
---|
| 447 |
|
---|
| 448 | foreach my $oid ( keys $infodb_map ) {
|
---|
| 449 | my $vals = $infodb_map->{$oid};
|
---|
| 450 | # interested in oid, timestamp, deletion status
|
---|
| 451 |
|
---|
| 452 | my ($deletion_status) = ($vals=~/^<status>(.*)$/m);
|
---|
| 453 | my ($timestamp) = ($vals=~/^<timestamp>(.*)$/m);
|
---|
| 454 |
|
---|
[31192] | 455 | $self->set_info ($oid, $deletion_status, $timestamp);
|
---|
[31190] | 456 | }
|
---|
| 457 | }
|
---|
| 458 |
|
---|
| 459 | # if no filename is passed in (and you don't generally want to), then
|
---|
| 460 | # it tries to load in <collection>/etc/oai-inf.<db> if it exists
|
---|
| 461 | sub load_info {
|
---|
| 462 | my $self = shift (@_);
|
---|
| 463 | my ($filename) = @_;
|
---|
| 464 |
|
---|
| 465 | $self->{'info'} = {};
|
---|
| 466 |
|
---|
| 467 | $filename = $self->{'oaidb_file_path'} unless defined $filename;
|
---|
| 468 |
|
---|
| 469 | if (&FileUtils::fileExists($filename)) {
|
---|
| 470 | if ($filename =~ m/\.inf$/) {
|
---|
| 471 | $self->_load_info_txt($filename);
|
---|
| 472 | }
|
---|
| 473 | else {
|
---|
| 474 | $self->_load_info_db($filename);
|
---|
| 475 | }
|
---|
| 476 | }
|
---|
| 477 |
|
---|
| 478 | }
|
---|
| 479 |
|
---|
| 480 | sub _save_info_txt {
|
---|
| 481 | my $self = shift (@_);
|
---|
| 482 | my ($filename) = @_;
|
---|
| 483 |
|
---|
| 484 | my ($OID, $info);
|
---|
| 485 |
|
---|
| 486 | open (OUTFILE, ">$filename") ||
|
---|
| 487 | die "oaiinfo::save_info couldn't write $filename\n";
|
---|
| 488 |
|
---|
| 489 | foreach $info (@{$self->get_OID_list()}) {
|
---|
| 490 | if (defined $info) {
|
---|
| 491 | print OUTFILE join("\t", @$info), "\n";
|
---|
| 492 | }
|
---|
| 493 | }
|
---|
| 494 | close (OUTFILE);
|
---|
| 495 | }
|
---|
| 496 |
|
---|
| 497 | # if no filename is passed in (and you don't generally want to), then
|
---|
| 498 | # this subroutine tries to write to <collection>/etc/oai-inf.<db>.
|
---|
| 499 | sub _save_info_db {
|
---|
| 500 | my $self = shift (@_);
|
---|
| 501 | my ($filename) = @_;
|
---|
| 502 |
|
---|
| 503 | $filename = $self->{'oaidb_file_path'} unless defined $filename;
|
---|
| 504 | my $infodbtype = $self->{'infodbtype'};
|
---|
| 505 |
|
---|
| 506 | # write out again. Open file for overwriting, not appending.
|
---|
| 507 | # Then write out data structure $self->{'info'} that's been maintaining the data in-memory.
|
---|
| 508 | my $infodb_handle = &dbutil::open_infodb_write_handle($infodbtype, $filename);
|
---|
| 509 | foreach my $oid ( keys $self->{'info'} ) {
|
---|
| 510 | my $OID_info = $self->{'info'}->{$oid};
|
---|
| 511 | #my $val_hash = {
|
---|
| 512 | # "status" => $OID_info->[INFO_STATUS_INDEX],
|
---|
| 513 | # "timestamp" => $OID_info->[INFO_TIMESTAMP_INDEX]
|
---|
| 514 | #};
|
---|
| 515 |
|
---|
| 516 | #&dbutil::write_infodb_rawentry($infodbtype,$infodb_handle,$oid,$val_hash);
|
---|
| 517 |
|
---|
| 518 | my $val = "<status>".$OID_info->[INFO_STATUS_INDEX]."\n<timestamp>".$OID_info->[INFO_TIMESTAMP_INDEX]."\n";
|
---|
| 519 | &dbutil::write_infodb_rawentry($infodbtype,$infodb_handle,$oid,$val);
|
---|
| 520 | }
|
---|
| 521 | &dbutil::close_infodb_write_handle($infodbtype, $infodb_handle);
|
---|
| 522 | }
|
---|
| 523 |
|
---|
| 524 | sub save_info {
|
---|
| 525 | my $self = shift (@_);
|
---|
| 526 | my ($filename) = @_;
|
---|
| 527 |
|
---|
| 528 | if(defined $filename) {
|
---|
| 529 | if ($filename =~ m/(contents)|(\.inf)$/) {
|
---|
| 530 | $self->_save_info_txt($filename);
|
---|
| 531 | }
|
---|
| 532 | else {
|
---|
| 533 | $self->_save_info_db($filename);
|
---|
| 534 | }
|
---|
| 535 | } else {
|
---|
| 536 | $self->_save_info_db();
|
---|
| 537 | }
|
---|
| 538 | }
|
---|
| 539 |
|
---|
| 540 |
|
---|
| 541 | sub set_info { # sets existing or appends
|
---|
| 542 | my $self = shift (@_);
|
---|
| 543 | my ($OID, $del_status, $timestamp) = @_;
|
---|
| 544 | if(!defined $timestamp) { # get current date timestamp
|
---|
| 545 | $timestamp = $self->get_current_time();
|
---|
| 546 | }
|
---|
| 547 | $self->{'info'}->{$OID} = [$del_status, $timestamp];
|
---|
| 548 |
|
---|
| 549 | }
|
---|
| 550 |
|
---|
| 551 |
|
---|
| 552 | # returns a list of the form [[OID, timestamp, deletion_status], ...]
|
---|
| 553 | sub get_OID_list
|
---|
| 554 | {
|
---|
| 555 | my $self = shift (@_);
|
---|
| 556 |
|
---|
| 557 | my @list = ();
|
---|
| 558 |
|
---|
| 559 | foreach my $OID (keys $self->{'info'}) {
|
---|
| 560 | my $OID_info = $self->{'info'}->{$OID};
|
---|
| 561 |
|
---|
| 562 | push (@list, [$OID, $OID_info->[INFO_STATUS_INDEX],
|
---|
| 563 | $OID_info->[INFO_TIMESTAMP_INDEX]]);
|
---|
| 564 | }
|
---|
| 565 |
|
---|
| 566 | return \@list;
|
---|
| 567 | }
|
---|
| 568 |
|
---|
| 569 |
|
---|
| 570 | # returns the number of entries so far, including deleted ones
|
---|
| 571 | # http://stackoverflow.com/questions/1109095/how-can-i-find-the-number-of-keys-in-a-hash-in-perl
|
---|
| 572 | sub size {
|
---|
| 573 | my $self = shift (@_);
|
---|
| 574 | return (scalar keys $self->{'info'});
|
---|
| 575 | }
|
---|
| 576 |
|
---|
| 577 | 1;
|
---|