Attachment 84283 Details for Bug 21958 – Bug 21958: Fix bibliographic record field comparison with authority

[patch] Bug 21958: Fix bibliographic record field comparison with authority

Bug-21958-Fix-bibliographic-record-field-compariso.patch (text/plain), 18.08 KB, created by Ere Maijala on 2019-01-22 14:50:38 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: Ere Maijala

Created: 2019-01-22 14:50:38 UTC

Size: 18.08 KB

patch

obsolete

>From 399669385b475e113ca12ce84c3c47652c40aad0 Mon Sep 17 00:00:00 2001
>From: Ere Maijala <ere.maijala@helsinki.fi>
>Date: Tue, 22 Jan 2019 14:38:32 +0200
>Subject: [PATCH] Bug 21958: Fix bibliographic record field comparison with
> authority
>
>This makes the comparison between bibliographic field and authority field more robust and per subfield. This makes the comparison not consider the same e.g. the following fields:
>
>$a Test User
>$a Test $b User
>
>The actual issue cannot be as easily reproduced with the patches for bug 21826 applied, but here's a test plan anyway:
>
>1. Make sure tests pass (especially t/db_dependent/AuthoritiesMarc*)
>2. Make sure authority linking still works properly
>3. Make sure authority and biblio frameworks allow subfield i
>4. Make sure that even if you add subfield i to 700 in biblio, authority link is kept the same
>5. Make sure that even if you add subfield i to the authority record, the authority link is kept the same
>---
> C4/AuthoritiesMarc.pm                         | 121 +++++++++---------
> C4/Biblio.pm                                  |   5 +-
> C4/Heading.pm                                 |   2 +
> C4/Heading/MARC21.pm                          | 120 +++++++++++++++++
> .../mandatory/authorities_normal_marc21.sql   |   5 +-
> t/db_dependent/AuthoritiesMarc.t              |  28 +++-
> t/db_dependent/AuthoritiesMarc/MARC21.t       |  91 +++++++++++++
> 7 files changed, 308 insertions(+), 64 deletions(-)
> create mode 100644 t/db_dependent/AuthoritiesMarc/MARC21.t
>
>diff --git a/C4/AuthoritiesMarc.pm b/C4/AuthoritiesMarc.pm
>index 10c67fc9ff..6e3b3db344 100644
>--- a/C4/AuthoritiesMarc.pm
>+++ b/C4/AuthoritiesMarc.pm
>@@ -1,6 +1,7 @@
> package C4::AuthoritiesMarc;
> 
> # Copyright 2000-2002 Katipo Communications
>+# Copyright 2018 The National Library of Finland, University of Helsinki
> #
> # This file is part of Koha.
> #
>@@ -359,10 +360,12 @@ sub GuessAuthTypeCode {
>         '110'=>{authtypecode=>'CORPO_NAME'},
>         '111'=>{authtypecode=>'MEETI_NAME'},
>         '130'=>{authtypecode=>'UNIF_TITLE'},
>+        '147'=>{authtypecode=>'NAME_EVENT'},
>         '148'=>{authtypecode=>'CHRON_TERM'},
>         '150'=>{authtypecode=>'TOPIC_TERM'},
>         '151'=>{authtypecode=>'GEOGR_NAME'},
>         '155'=>{authtypecode=>'GENRE/FORM'},
>+        '162'=>{authtypecode=>'MED_PERFRM'},
>         '180'=>{authtypecode=>'GEN_SUBDIV'},
>         '181'=>{authtypecode=>'GEO_SUBDIV'},
>         '182'=>{authtypecode=>'CHRON_SUBD'},
>@@ -938,6 +941,8 @@ sub BuildSummary {
> # construct MARC21 summary
> # FIXME - looping over 1XX is questionable
> # since MARC21 authority should have only one 1XX
>+        use C4::Heading::MARC21;
>+        my $handler = C4::Heading::MARC21->new();
>         my $subfields_to_report;
>         foreach my $field ($record->field('1..')) {
>             my $tag = $field->tag();
>@@ -945,31 +950,9 @@ sub BuildSummary {
> # FIXME - 152 is not a good tag to use
> # in MARC21 -- purely local tags really ought to be
> # 9XX
>-            if ($tag eq '100') {
>-                $subfields_to_report = 'abcdefghjklmnopqrstvxyz';
>-            } elsif ($tag eq '110') {
>-                $subfields_to_report = 'abcdefghklmnoprstvxyz';
>-            } elsif ($tag eq '111') {
>-                $subfields_to_report = 'acdefghklnpqstvxyz';
>-            } elsif ($tag eq '130') {
>-                $subfields_to_report = 'adfghklmnoprstvxyz';
>-            } elsif ($tag eq '148') {
>-                $subfields_to_report = 'abvxyz';
>-            } elsif ($tag eq '150') {
>-                $subfields_to_report = 'abvxyz';
>-            } elsif ($tag eq '151') {
>-                $subfields_to_report = 'avxyz';
>-            } elsif ($tag eq '155') {
>-                $subfields_to_report = 'abvxyz';
>-            } elsif ($tag eq '180') {
>-                $subfields_to_report = 'vxyz';
>-            } elsif ($tag eq '181') {
>-                $subfields_to_report = 'vxyz';
>-            } elsif ($tag eq '182') {
>-                $subfields_to_report = 'vxyz';
>-            } elsif ($tag eq '185') {
>-                $subfields_to_report = 'vxyz';
>-            }
>+
>+            $subfields_to_report = $handler->get_auth_heading_subfields_to_report($tag);
>+
>             if ($subfields_to_report) {
>                 push @authorized, {
>                     heading => $field->as_string($subfields_to_report),
>@@ -1093,44 +1076,47 @@ sub GetAuthorizedHeading {
>             return $field->as_string('abcdefghijlmnopqrstuvwxyz');
>         }
>     } else {
>+        use C4::Heading::MARC21;
>+        my $handler = C4::Heading::MARC21->new();
>+
>         foreach my $field ($record->field('1..')) {
>-            my $tag = $field->tag();
>-            next if "152" eq $tag;
>-# FIXME - 152 is not a good tag to use
>-# in MARC21 -- purely local tags really ought to be
>-# 9XX
>-            if ($tag eq '100') {
>-                return $field->as_string('abcdefghjklmnopqrstvxyz68');
>-            } elsif ($tag eq '110') {
>-                return $field->as_string('abcdefghklmnoprstvxyz68');
>-            } elsif ($tag eq '111') {
>-                return $field->as_string('acdefghklnpqstvxyz68');
>-            } elsif ($tag eq '130') {
>-                return $field->as_string('adfghklmnoprstvxyz68');
>-            } elsif ($tag eq '148') {
>-                return $field->as_string('abvxyz68');
>-            } elsif ($tag eq '150') {
>-                return $field->as_string('abvxyz68');
>-            } elsif ($tag eq '151') {
>-                return $field->as_string('avxyz68');
>-            } elsif ($tag eq '155') {
>-                return $field->as_string('abvxyz68');
>-            } elsif ($tag eq '180') {
>-                return $field->as_string('vxyz68');
>-            } elsif ($tag eq '181') {
>-                return $field->as_string('vxyz68');
>-            } elsif ($tag eq '182') {
>-                return $field->as_string('vxyz68');
>-            } elsif ($tag eq '185') {
>-                return $field->as_string('vxyz68');
>-            } else {
>-                return $field->as_string();
>-            }
>+            my $subfields = $handler->get_valid_bib_heading_subfields($field->tag());
>+            return $field->as_string($subfields) if ($subfields);
>         }
>     }
>     return;
> }
> 
>+=head2 CompareFieldWithAuthority
>+
>+  $match = &CompareFieldWithAuthority({ field => $field, authid => $authid })
>+
>+Takes a MARC::Field from a bibliographic record and an authid, and returns true if they match.
>+
>+=cut
>+
>+sub CompareFieldWithAuthority {
>+    my $args = shift;
>+
>+    my $record = GetAuthority($args->{authid});
>+    return unless (ref $record eq 'MARC::Record');
>+    if (C4::Context->preference('marcflavour') eq 'UNIMARC') {
>+        # UNIMARC has same subfields for bibs and authorities
>+        foreach my $field ($record->field('2..')) {
>+            return compare_fields($field, $args->{field}, 'abcdefghijlmnopqrstuvwxyz');
>+        }
>+    } else {
>+        use C4::Heading::MARC21;
>+        my $handler = C4::Heading::MARC21->new();
>+
>+        foreach my $field ($record->field('1..')) {
>+            my $subfields = $handler->get_valid_bib_heading_subfields($field->tag());
>+            return compare_fields($field, $args->{field}, $subfields) if ($subfields);
>+        }
>+    }
>+    return 0;
>+}
>+
> =head2 BuildAuthHierarchies
> 
>   $text= &BuildAuthHierarchies( $authid, $force)
>@@ -1592,6 +1578,26 @@ sub get_auth_type_location {
>     }
> }
> 
>+=head2 compare_fields
>+
>+  my match = compare_fields($field1, $field2, 'abcde');
>+
>+Compares the listed subfields of both fields and return true if they all match
>+
>+=cut
>+
>+sub compare_fields {
>+    my ($field1, $field2, $subfields) = @_;
>+
>+    foreach my $subfield (split(//, $subfields)) {
>+        my $subfield1 = $field1->subfield($subfield) // '';
>+        my $subfield2 = $field2->subfield($subfield) // '';
>+        return 0 unless $subfield1 eq $subfield2;
>+    }
>+    return 1;
>+}
>+
>+
> END { }       # module clean-up code here (global destructor)
> 
> 1;
>@@ -1602,6 +1608,7 @@ __END__
> Koha Development Team <http://koha-community.org/>
> 
> Paul POULAIN paul.poulain@free.fr
>+Ere Maijala ere.maijala@helsinki.fi
> 
> =cut
> 
>diff --git a/C4/Biblio.pm b/C4/Biblio.pm
>index 0249983331..fd79566211 100644
>--- a/C4/Biblio.pm
>+++ b/C4/Biblio.pm
>@@ -628,10 +628,7 @@ sub _check_valid_auth_link {
> 
>     require C4::AuthoritiesMarc;
> 
>-    my $authorized_heading =
>-      C4::AuthoritiesMarc::GetAuthorizedHeading( { 'authid' => $authid } ) || '';
>-
>-   return ($field->as_string('abcdefghijklmnopqrstuvwxyz') eq $authorized_heading);
>+    return C4::AuthoritiesMarc::CompareFieldWithAuthority( { 'field' => $field, 'authid' => $authid } );
> }
> 
> =head2 GetRecordValue
>diff --git a/C4/Heading.pm b/C4/Heading.pm
>index 5dfd0fadb0..c4296a9393 100644
>--- a/C4/Heading.pm
>+++ b/C4/Heading.pm
>@@ -174,6 +174,8 @@ sub preferred_authorities {
> 
>     if (C4::Heading::valid_bib_heading_subfield('100', 'e', '')) ...
> 
>+Check if the given subfield is valid for the given field.
>+
> =cut
> 
> sub valid_bib_heading_subfield {
>diff --git a/C4/Heading/MARC21.pm b/C4/Heading/MARC21.pm
>index fc4d576a45..2dfcbf799e 100644
>--- a/C4/Heading/MARC21.pm
>+++ b/C4/Heading/MARC21.pm
>@@ -68,6 +68,52 @@ my $bib_heading_fields = {
>         subfields  => 'adfghklmnoprst',
>         main_entry => 1
>     },
>+    '147' => {
>+        auth_type => 'NAME_EVENT',
>+        subfields => 'acdgvxyz68',
>+        main_entry => 1
>+    },
>+    '148' => {
>+        auth_type => 'CHRON_TERM',
>+        subfields => 'abvxyz68',
>+        main_entry => 1
>+    },
>+    '150' => {
>+        auth_type => 'TOPIC_TERM',
>+        subfields => 'abvxyz68',
>+        main_entry => 1
>+    },
>+    '151' => {
>+        auth_type => 'GEOGR_NAME',
>+        subfields => 'avxyz68',
>+        main_entry => 1
>+    },
>+    '155' => {
>+        auth_type => 'GENRE/FORM',
>+        subfields => 'abvxyz68',
>+        main_entry => 1
>+    },
>+    '162' => {
>+        auth_type => 'MED_PERFRM',
>+        subfields => 'a68',
>+        main_entry => 1
>+    },
>+    '180' => {
>+        auth_type => 'TOPIC_TERM',
>+        subfields => 'vxyz68'
>+    },
>+    '181' => {
>+        auth_type => 'GEOGR_NAME',
>+        subfields => 'vxyz68'
>+    },
>+    '182' => {
>+        auth_type => 'CHRON_TERM',
>+        subfields => 'vxyz68'
>+    },
>+    '185' => {
>+        auth_type => 'GENRE/FORM',
>+        subfields => 'vxyz68'
>+    },
>     '440' => { auth_type => 'UNIF_TITLE', subfields => 'anp', series => 1 },
>     '600' => {
>         auth_type => 'PERSO_NAME',
>@@ -119,6 +165,56 @@ my $bib_heading_fields = {
>       { auth_type => 'UNIF_TITLE', subfields => 'adfghklmnoprst', series => 1 },
> };
> 
>+
>+=head2 auth_heading_fields
>+
>+=cut
>+
>+my $auth_heading_fields = {
>+    '100' => {
>+        subfields => 'abcdefghjklmnopqrstvxyz68',
>+    },
>+    '110' => {
>+        subfields => 'abcdefghklmnoprstvxyz68',
>+    },
>+    '111' => {
>+        subfields => 'acdefghklnpqstvxyz68',
>+    },
>+    '130' => {
>+        subfields => 'adfghklmnoprstvxyz68',
>+    },
>+    '147' => {
>+        subfields => 'acdgvxyz68',
>+    },
>+    '148' => {
>+        subfields => 'abvxyz68',
>+    },
>+    '150' => {
>+        subfields => 'abvxyz68',
>+    },
>+    '151' => {
>+        subfields => 'avxyz68',
>+    },
>+    '155' => {
>+        subfields => 'abvxyz68',
>+    },
>+    '162' => {
>+        subfields => 'a68',
>+    },
>+    '180' => {
>+        subfields => 'vxyz68',
>+    },
>+    '181' => {
>+        subfields => 'vxyz68',
>+    },
>+    '182' => {
>+        subfields => 'vxyz68',
>+    },
>+    '185' => {
>+        subfields => 'vxyz68',
>+    },
>+};
>+
> =head2 subdivisions
> 
> =cut
>@@ -176,6 +272,30 @@ sub valid_bib_heading_subfield {
>     return 0;
> }
> 
>+=head2 get_valid_bib_heading_subfields
>+
>+=cut
>+
>+sub get_valid_bib_heading_subfields {
>+    my $self          = shift;
>+    my $tag           = shift;
>+
>+    return $bib_heading_fields->{$tag}->{subfields} // undef;
>+}
>+
>+=head2 get_auth_heading_subfields_to_report
>+
>+=cut
>+
>+sub get_auth_heading_subfields_to_report {
>+    my $self          = shift;
>+    my $tag           = shift;
>+
>+    my $subfields = $auth_heading_fields->{$tag}->{subfields} // '';
>+    $subfields =~ s/[68]//;
>+    return $subfields;
>+}
>+
> =head2 parse_heading
> 
> =cut
>diff --git a/installer/data/mysql/en/marcflavour/marc21/mandatory/authorities_normal_marc21.sql b/installer/data/mysql/en/marcflavour/marc21/mandatory/authorities_normal_marc21.sql
>index 901561904d..cfb5e64091 100644
>--- a/installer/data/mysql/en/marcflavour/marc21/mandatory/authorities_normal_marc21.sql
>+++ b/installer/data/mysql/en/marcflavour/marc21/mandatory/authorities_normal_marc21.sql
>@@ -28,11 +28,12 @@ INSERT INTO auth_types (authtypecode, authtypetext, auth_tag_to_report, summary)
> 		('CORPO_NAME', 'Corporate Name', '110', 'Corporate Names'),
> 		('MEETI_NAME', 'Meeting Name', '111', 'Meeting Name'),
> 		('UNIF_TITLE', 'Uniform Title', '130', 'Uniform Title'),
>+		('NAME_EVENT', 'Named Event', '147', 'Named Event'),
> 		('CHRON_TERM', 'Chronological Term', '148', 'Chronological Term'),
> 		('TOPIC_TERM', 'Topical Term', '150', 'Topical Term'),
> 		('GEOGR_NAME', 'Geographic Name', '151', 'Geographic Name'),
>-		('GENRE/FORM', 'Genre/Form Term', '155', 'Genre/Form Term');
>-
>+		('GENRE/FORM', 'Genre/Form Term', '155', 'Genre/Form Term'),
>+	    ('MED_PERFRM', 'Medium of Performance', '162', 'Medium of Performance');
> 
> -- ******************************************************
> -- KOHA UNSUPPORTED STANDARD MARC 21 AUTHORITY TYPES.
>diff --git a/t/db_dependent/AuthoritiesMarc.t b/t/db_dependent/AuthoritiesMarc.t
>index a24d917f50..97f118ccf4 100755
>--- a/t/db_dependent/AuthoritiesMarc.t
>+++ b/t/db_dependent/AuthoritiesMarc.t
>@@ -5,9 +5,10 @@
> 
> use Modern::Perl;
> 
>-use Test::More tests => 9;
>+use Test::More tests => 10;
> use Test::MockModule;
> use Test::Warn;
>+use MARC::Field;
> use MARC::Record;
> 
> use t::lib::Mocks;
>@@ -54,6 +55,11 @@ $module->mock('GetAuthority', sub {
>             [ '151', ' ', ' ', a => 'New York (City)' ],
>             [ '551', ' ', ' ', a => 'New York (State)', w => 'g' ]
>             );
>+    } elsif ($authid eq '5') {
>+        $record->add_fields(
>+            [ '001', '5' ],
>+            [ '100', ' ', ' ', a => 'Lastname, Firstname', b => 'b', c => 'c', i => 'i' ]
>+            );
>     } else {
>         undef $record;
>     }
>@@ -214,4 +220,24 @@ subtest 'AddAuthority should respect AUTO_INCREMENT (BZ 18104)' => sub {
>     is( $record->field('001')->data, $id3, 'Check updated 001' );
> };
> 
>+subtest 'CompareFieldWithAuthority tests' => sub {
>+    plan tests => 3;
>+
>+    t::lib::Mocks::mock_preference('marcflavour', 'MARC21');
>+
>+    $builder->build({ source => 'AuthType', value => { authtypecode => 'PERSO_NAME' }});
>+
>+    my $field = MARC::Field->new('100', 0, 0, a => 'Lastname, Firstname', b => 'b', c => 'c');
>+
>+    ok(C4::AuthoritiesMarc::CompareFieldWithAuthority({'field' => $field, 'authid' => 5}), 'Authority matches');
>+
>+    $field->add_subfields(i => 'X');
>+
>+    ok(C4::AuthoritiesMarc::CompareFieldWithAuthority({'field' => $field, 'authid' => 5}), 'Compare ignores unlisted subfields');
>+
>+    $field->add_subfields(d => 'd');
>+
>+    ok(!C4::AuthoritiesMarc::CompareFieldWithAuthority({'field' => $field, 'authid' => 5}), 'Authority does not match');
>+};
>+
> $schema->storage->txn_rollback;
>diff --git a/t/db_dependent/AuthoritiesMarc/MARC21.t b/t/db_dependent/AuthoritiesMarc/MARC21.t
>new file mode 100644
>index 0000000000..77047e4bbc
>--- /dev/null
>+++ b/t/db_dependent/AuthoritiesMarc/MARC21.t
>@@ -0,0 +1,91 @@
>+#!/usr/bin/perl
>+#
>+
>+use strict;
>+use warnings;
>+
>+use Test::MockModule;
>+use Test::More tests => 2;
>+use MARC::Field;
>+use MARC::Record;
>+
>+use t::lib::Mocks;
>+use t::lib::TestBuilder;
>+
>+BEGIN {
>+    use_ok('C4::AuthoritiesMarc::MARC21');
>+}
>+
>+my $schema  = Koha::Database->new->schema;
>+$schema->storage->txn_begin;
>+my $dbh = C4::Context->dbh;
>+my $builder = t::lib::TestBuilder->new;
>+
>+t::lib::Mocks::mock_preference('marcflavour', 'MARC21');
>+
>+subtest 'CompareFieldWithAuthority tests' => sub {
>+    plan tests => 3;
>+
>+    # We are now going to be testing the authorities hierarchy code, and
>+    # therefore need to pretend that we have consistent data in our database
>+    my $module = new Test::MockModule('C4::AuthoritiesMarc');
>+    $module->mock('GetHeaderAuthority', sub {
>+        return {'authtrees' => ''};
>+    });
>+    $module->mock('AddAuthorityTrees', sub {
>+        return;
>+    });
>+    $module->mock('GetAuthority', sub {
>+        my ($authid) = @_;
>+        my $record = MARC::Record->new();
>+        if ($authid eq '1') {
>+            $record->add_fields(
>+                [ '001', '1' ],
>+                [ '151', ' ', ' ', a => 'United States' ]
>+            );
>+        } elsif ($authid eq '2') {
>+            $record->add_fields(
>+                [ '001', '2' ],
>+                [ '151', ' ', ' ', a => 'New York (State)' ],
>+                [ '551', ' ', ' ', a => 'United States', w => 'g', 9 => '1' ]
>+            );
>+        } elsif ($authid eq '3') {
>+            $record->add_fields(
>+                [ '001', '3' ],
>+                [ '151', ' ', ' ', a => 'New York (City)' ],
>+                [ '551', ' ', ' ', a => 'New York (State)', w => 'g', 9 => '2' ]
>+            );
>+        } elsif ($authid eq '4') {
>+            $record->add_fields(
>+                [ '001', '4' ],
>+                [ '151', ' ', ' ', a => 'New York (City)' ],
>+                [ '551', ' ', ' ', a => 'New York (State)', w => 'g' ]
>+            );
>+        } elsif ($authid eq '5') {
>+            $record->add_fields(
>+                [ '001', '5' ],
>+                [ '100', ' ', ' ', a => 'Lastname, Firstname', b => 'b', c => 'c', i => 'i' ]
>+            );
>+        } else {
>+            undef $record;
>+        }
>+        return $record;
>+    });
>+
>+    $dbh->do('DELETE FROM auth_types');
>+    $builder->build({ source => 'AuthType', value => { authtypecode => 'PERSO_NAME' }});
>+
>+    my $field = MARC::Field->new('100', 0, 0, a => 'Lastname, Firstname', b => 'b', c => 'c');
>+
>+    ok(C4::AuthoritiesMarc::CompareFieldWithAuthority({'field' => $field, 'authid' => 5}), 'Authority matches');
>+
>+    $field->add_subfields(i => 'X');
>+
>+    ok(C4::AuthoritiesMarc::CompareFieldWithAuthority({'field' => $field, 'authid' => 5}), 'Compare ignores unlisted subfields');
>+
>+    $field->add_subfields(d => 'd');
>+
>+    ok(!C4::AuthoritiesMarc::CompareFieldWithAuthority({'field' => $field, 'authid' => 5}), 'Authority does not match');
>+};
>+
>+$schema->storage->txn_rollback;
>-- 
>2.17.1

Actions: View | Diff | Splinter Review

Attachments on bug 21958: 84283 | 91150 | 92611 | 99528 | 108182 | 114303 | 118795 | 118796 | 119217