matching enhancements -- allow matching rule to be changed on the fly

Enhancement to store the matching rule associated with an
import batch and to allow the current matching rule in
effect to be changed and the duplicate detection redone.

Signed-off-by: Chris Cormack <crc@liblime.com>
Signed-off-by: Joshua Ferraro <jmf@liblime.com>

diff --git a/C4/ImportBatch.pm b/C4/ImportBatch.pm
index 05dae67..dfd0506 100644 (file)
--- a/C4/ImportBatch.pm
+++ b/C4/ImportBatch.pm
@@ -21,7 +21,6 @@ use strict;
 use C4::Context;
 use C4::Koha;
 use C4::Biblio;
-use C4::Matcher;
 require Exporter;
 
 
@@ -69,6 +68,8 @@ use C4::ImportBatch;
     SetImportBatchStatus
     GetImportBatchOverlayAction
     SetImportBatchOverlayAction
+    GetImportBatchMatcher
+    SetImportBatchMatcher
     GetImportRecordOverlayStatus
     SetImportRecordOverlayStatus
     GetImportRecordStatus
@@ -390,12 +391,16 @@ sub BatchFindBibDuplicates {
             &$progress_callback($rec_num);
         }
         my $marc_record = MARC::Record->new_from_usmarc($rowref->{'marc'});
-        my @matches = $matcher->get_matches($marc_record, $max_matches);
+        my @matches = ();
+        if (defined $matcher) {
+            @matches = $matcher->get_matches($marc_record, $max_matches);
+        }
         if (scalar(@matches) > 0) {
             $num_with_matches++;
             SetImportRecordMatches($rowref->{'import_record_id'}, @matches);
             SetImportRecordOverlayStatus($rowref->{'import_record_id'}, 'auto_match');
         } else {
+            SetImportRecordMatches($rowref->{'import_record_id'}, ());
             SetImportRecordOverlayStatus($rowref->{'import_record_id'}, 'no_match');
         }
     }
@@ -842,6 +847,49 @@ sub SetImportBatchOverlayAction {
 
 }
 
+=head2 GetImportBatchMatcher
+
+=over 4
+
+my $matcher_id = GetImportBatchMatcher($batch_id);
+
+=back
+
+=cut
+
+sub GetImportBatchMatcher {
+    my ($batch_id) = @_;
+
+    my $dbh = C4::Context->dbh;
+    my $sth = $dbh->prepare("SELECT matcher_id FROM import_batches WHERE import_batch_id = ?");
+    $sth->execute($batch_id);
+    my ($matcher_id) = $sth->fetchrow_array();
+    $sth->finish();
+    return $matcher_id;
+
+}
+
+
+=head2 SetImportBatchMatcher
+
+=over 4
+
+SetImportBatchMatcher($batch_id, $new_matcher_id);
+
+=back
+
+=cut
+
+sub SetImportBatchMatcher {
+    my ($batch_id, $new_matcher_id) = @_;
+
+    my $dbh = C4::Context->dbh;
+    my $sth = $dbh->prepare("UPDATE import_batches SET matcher_id = ? WHERE import_batch_id = ?");
+    $sth->execute($new_matcher_id, $batch_id);
+    $sth->finish();
+
+}
+
 =head2 GetImportRecordOverlayStatus
 
 =over 4

diff --git a/C4/Matcher.pm b/C4/Matcher.pm
index 16f42c2..505ff27 100644 (file)
--- a/C4/Matcher.pm
+++ b/C4/Matcher.pm
@@ -166,6 +166,7 @@ sub fetch {
     my $sth = $dbh->prepare_cached("SELECT * FROM marc_matchers WHERE matcher_id = ?");
     $sth->execute($id);
     my $row = $sth->fetchrow_hashref;
+    $sth->finish();
     return undef unless defined $row;
 
     my $self = {};

diff --git a/installer/data/mysql/kohastructure.sql b/installer/data/mysql/kohastructure.sql
index 739e08d..3d3b6a2 100644 (file)
--- a/installer/data/mysql/kohastructure.sql
+++ b/installer/data/mysql/kohastructure.sql
@@ -846,6 +846,7 @@ CREATE TABLE `ethnicity` (
 DROP TABLE IF EXISTS `import_batches`;
 CREATE TABLE `import_batches` (
   `import_batch_id` int(11) NOT NULL auto_increment,
+  `matcher_id` int(11) default NULL,
   `template_id` int(11) default NULL,
   `branchcode` varchar(10) default NULL,
   `num_biblios` int(11) NOT NULL default 0,

diff --git a/installer/data/mysql/updatedatabase.pl b/installer/data/mysql/updatedatabase.pl
index 6242034..688c262 100755 (executable)
--- a/installer/data/mysql/updatedatabase.pl
+++ b/installer/data/mysql/updatedatabase.pl
@@ -673,6 +673,13 @@ if (C4::Context->preference("Version") < TransformToNum($DBversion)) {
 }
 
 
+$DBversion = "3.00.00.029";
+if (C4::Context->preference("Version") < TransformToNum($DBversion)) {
+       $dbh->do("ALTER TABLE `import_batches` ADD `matcher_id` int(11) NULL AFTER `import_batch_id`");
+    print "Upgrade to $DBversion done (adding matcher_id to import_batches)\n";
+    SetVersion ($DBversion);
+}
+
 =item DropAllForeignKeys($table)
 
   Drop all foreign keys of the table $table

diff --git a/koha-tmpl/intranet-tmpl/prog/en/modules/tools/manage-marc-import.tmpl b/koha-tmpl/intranet-tmpl/prog/en/modules/tools/manage-marc-import.tmpl
index 4d05c06..0153499 100644 (file)
--- a/koha-tmpl/intranet-tmpl/prog/en/modules/tools/manage-marc-import.tmpl
+++ b/koha-tmpl/intranet-tmpl/prog/en/modules/tools/manage-marc-import.tmpl
@@ -37,6 +37,36 @@
 <p><label>Comments</label> <!-- TMPL_VAR name="comments" --></p>
 <p><label>Staged</label> <!-- TMPL_VAR name="upload_timestamp" --></p>
 <p><label>Status</label> <!-- TMPL_VAR name="import_status" --></p>
+<p><label>Matching rule applied</label> 
+<!-- TMPL_IF name="current_matcher_id" -->
+  <!-- TMPL_VAR name="current_matcher_code" --> (<!-- TMPL_VAR name="current_matcher_description" -->)
+<!-- TMPL_ELSE -->
+  No matching rule in effect
+<!-- /TMPL_IF -->
+<!-- TMPL_IF name="can_commit" -->
+<p>
+  <form action="<!-- TMPL_VAR name="script_name" -->" method="post">
+    <input type="hidden" name="op" value="redo-matching" />
+    <input type="hidden" name="import_batch_id" value="<!-- TMPL_VAR name="import_batch_id" -->" />
+    <input type="hidden" name="current_matcher_id" value="<!-- TMPL_VAR name="current_matcher_id" -->" />
+    <select name="new_matcher_id" id="new_matcher_id">
+       <option value="">Do not look for matching records</option> 
+       <!-- TMPL_LOOP name="available_matchers" -->
+          <!-- TMPL_IF name="selected" -->
+          <option value="<!-- TMPL_VAR name="matcher_id" -->" selected="selected">
+             <!-- TMPL_VAR name="code" --> (<!-- TMPL_VAR name="description" -->)
+          </option>
+          <!-- TMPL_ELSE -->
+          <option value="<!-- TMPL_VAR name="matcher_id" -->">
+            <!-- TMPL_VAR name="code" --> (<!-- TMPL_VAR name="description" -->)
+          </option>
+          <!-- /TMPL_IF -->
+       <!-- /TMPL_LOOP -->
+    </select>
+    <input type="submit" class="button" value="Apply different matching rule" />
+  </form>
+</p>
+<!-- /TMPL_IF -->
 <p>
   <!-- TMPL_IF name="can_commit" -->
   <form action="<!-- TMPL_VAR name="script_name" -->" method="post">
@@ -72,6 +102,15 @@
   <tr><td>Number of records ignored</td><td><!-- TMPL_VAR name="num_ignored" --></td></tr>
   </table>
   <!-- /TMPL_IF -->
+  <!-- TMPL_IF name="rematch_attempted" -->
+    <!-- TMPL_IF name="rematch_failed" -->
+      <p><span class="problem">Failed to apply different matching rule</span></p>
+    <!-- TMPL_ELSE -->
+      <p><span class="problem">Applied different matching rule.  Number of records matched now
+       <!-- TMPL_VAR name="num_with_matches" -->
+      </span></p>
+    <!-- /TMPL_IF -->
+  <!-- /TMPL_IF -->
 <!-- /TMPL_IF -->
 <br />
 

diff --git a/kohaversion.pl b/kohaversion.pl
index 6aa0995..c18be35 100644 (file)
--- a/kohaversion.pl
+++ b/kohaversion.pl
@@ -8,7 +8,7 @@
 #        and is automatically called by Auth.pm when needed.
 
 sub kohaversion {
-    return "3.00.00.028";
+    return "3.00.00.029";
 }
 
 1;

diff --git a/tools/manage-marc-import.pl b/tools/manage-marc-import.pl
index 6832eb7..1eb41dd 100755 (executable)
--- a/tools/manage-marc-import.pl
+++ b/tools/manage-marc-import.pl
@@ -72,12 +72,40 @@ if ($op eq "") {
     import_biblios_list($template, $import_batch_id, $offset, $results_per_page);
 } elsif ($op eq "clean-batch") {
     ;
+} elsif ($op eq "redo-matching") {
+    my $new_matcher_id = $input->param('new_matcher_id');
+    my $current_matcher_id = $input->param('current_matcher_id');
+    redo_matching($template, $import_batch_id, $new_matcher_id, $current_matcher_id);
+    import_biblios_list($template, $import_batch_id, $offset, $results_per_page);
 }
 
 output_html_with_http_headers $input, $cookie, $template->output;
 
 exit 0;
 
+sub redo_matching {
+    my ($template, $import_batch_id, $new_matcher_id, $current_matcher_id) = @_;
+    my $rematch_failed = 0;
+    return if not defined $new_matcher_id and not defined $current_matcher_id;
+    return if $new_matcher_id == $current_matcher_id;
+    my $num_with_matches = 0;
+    if (defined $new_matcher_id and $new_matcher_id ne "") {
+        my $matcher = C4::Matcher->fetch($new_matcher_id);
+        if (defined $matcher) {
+            $num_with_matches = BatchFindBibDuplicates($import_batch_id, $matcher);
+            SetImportBatchMatcher($import_batch_id, $new_matcher_id);
+        } else {
+            $rematch_failed = 1;
+        }
+    } else {
+        $num_with_matches = BatchFindBibDuplicates($import_batch_id, undef);
+         SetImportBatchMatcher($import_batch_id, undef);
+    }
+    $template->param(rematch_failed => $rematch_failed);
+    $template->param(rematch_attempted => 1);
+    $template->param(num_with_matches => $num_with_matches);
+}
+
 sub import_batches_list {
     my ($template, $offset, $results_per_page) = @_;
     my $batches = GetImportBatchRangeDesc($offset, $results_per_page);
@@ -178,6 +206,28 @@ sub batch_info {
     if ($batch->{'import_status'} eq 'imported') {
         $template->param(can_revert => 1);
     }
+    if (defined $batch->{'matcher_id'}) {
+        my $matcher = C4::Matcher->fetch($batch->{'matcher_id'});
+        if (defined $matcher) {
+            $template->param('current_matcher_id' => $batch->{'matcher_id'});
+            $template->param('current_matcher_code' => $matcher->code());
+            $template->param('current_matcher_description' => $matcher->description());
+        }
+    }
+    add_matcher_list($batch->{'matcher_id'});
+}
+
+sub add_matcher_list {
+    my $current_matcher_id = shift;
+    my @matchers = C4::Matcher::GetMatcherList();
+    if (defined $current_matcher_id) {
+        for (my $i = 0; $i <= $#matchers; $i++) {
+            if ($matchers[$i]->{'matcher_id'} == $current_matcher_id) {
+                $matchers[$i]->{'selected'} = 1;
+            }
+        }
+    }
+    $template->param(available_matchers => \@matchers);
 }
 
 sub add_page_numbers {

diff --git a/tools/stage-marc-import.pl b/tools/stage-marc-import.pl
index 7f8063d..22dff90 100755 (executable)
--- a/tools/stage-marc-import.pl
+++ b/tools/stage-marc-import.pl
@@ -94,6 +94,7 @@ if ($uploadmarc && length($uploadmarc)>0) {
             $checked_matches = 1;
             $matcher_code = $matcher->code();
             $num_with_matches = BatchFindBibDuplicates($batch_id, $matcher);
+            SetImportBatchMatcher($batch_id, $matcher_id);
         } else {
             $matcher_failed = 1;
         }