From: Bill Erickson Date: Thu, 28 Jul 2011 16:34:28 +0000 (-0400) Subject: Stamped upgrade for user/dbs/oclcnumbers X-Git-Url: https://old-git.evergreen-ils.org/?a=commitdiff_plain;h=ab28c7b8def4b015b92bf954dec5bbb5dd8143d9;p=evergreen%2Fpines.git Stamped upgrade for user/dbs/oclcnumbers Signed-off-by: Bill Erickson --- diff --git a/Open-ILS/src/sql/Pg/002.schema.config.sql b/Open-ILS/src/sql/Pg/002.schema.config.sql index c1849b9315..2ab29f9203 100644 --- a/Open-ILS/src/sql/Pg/002.schema.config.sql +++ b/Open-ILS/src/sql/Pg/002.schema.config.sql @@ -86,7 +86,7 @@ CREATE TRIGGER no_overlapping_deps BEFORE INSERT OR UPDATE ON config.db_patch_dependencies FOR EACH ROW EXECUTE PROCEDURE evergreen.array_overlap_check ('deprecates'); -INSERT INTO config.upgrade_log (version, applied_to) VALUES ('0586', :eg_version); -- senator/berick +INSERT INTO config.upgrade_log (version, applied_to) VALUES ('0587', :eg_version); -- dbs/berick CREATE TABLE config.bib_source ( id SERIAL PRIMARY KEY, diff --git a/Open-ILS/src/sql/Pg/upgrade/0587.maintain_control_numbers_with_OCLC_awareness.sql b/Open-ILS/src/sql/Pg/upgrade/0587.maintain_control_numbers_with_OCLC_awareness.sql new file mode 100644 index 0000000000..a583250a5e --- /dev/null +++ b/Open-ILS/src/sql/Pg/upgrade/0587.maintain_control_numbers_with_OCLC_awareness.sql @@ -0,0 +1,145 @@ +BEGIN; + +SELECT evergreen.upgrade_deps_block_check('0587', :eg_version); + +CREATE OR REPLACE FUNCTION maintain_control_numbers() RETURNS TRIGGER AS $func$ +use strict; +use MARC::Record; +use MARC::File::XML (BinaryEncoding => 'UTF-8'); +use MARC::Charset; +use Encode; +use Unicode::Normalize; + +MARC::Charset->assume_unicode(1); + +my $record = MARC::Record->new_from_xml($_TD->{new}{marc}); +my $schema = $_TD->{table_schema}; +my $rec_id = $_TD->{new}{id}; + +# Short-circuit if maintaining control numbers per MARC21 spec is not enabled +my $enable = spi_exec_query("SELECT enabled FROM config.global_flag WHERE name = 'cat.maintain_control_numbers'"); +if (!($enable->{processed}) or $enable->{rows}[0]->{enabled} eq 'f') { + return; +} + +# Get the control number identifier from an OU setting based on $_TD->{new}{owner} +my $ou_cni = 'EVRGRN'; + +my $owner; +if ($schema eq 'serial') { + $owner = $_TD->{new}{owning_lib}; +} else { + # are.owner and bre.owner can be null, so fall back to the consortial setting + $owner = $_TD->{new}{owner} || 1; +} + +my $ous_rv = spi_exec_query("SELECT value FROM actor.org_unit_ancestor_setting('cat.marc_control_number_identifier', $owner)"); +if ($ous_rv->{processed}) { + $ou_cni = $ous_rv->{rows}[0]->{value}; + $ou_cni =~ s/"//g; # Stupid VIM syntax highlighting" +} else { + # Fall back to the shortname of the OU if there was no OU setting + $ous_rv = spi_exec_query("SELECT shortname FROM actor.org_unit WHERE id = $owner"); + if ($ous_rv->{processed}) { + $ou_cni = $ous_rv->{rows}[0]->{shortname}; + } +} + +my ($create, $munge) = (0, 0); + +my @scns = $record->field('035'); + +foreach my $id_field ('001', '003') { + my $spec_value; + my @controls = $record->field($id_field); + + if ($id_field eq '001') { + $spec_value = $rec_id; + } else { + $spec_value = $ou_cni; + } + + # Create the 001/003 if none exist + if (scalar(@controls) == 1) { + # Only one field; check to see if we need to munge it + unless (grep $_->data() eq $spec_value, @controls) { + $munge = 1; + } + } else { + # Delete the other fields, as with more than 1 001/003 we do not know which 003/001 to match + foreach my $control (@controls) { + $record->delete_field($control); + } + $record->insert_fields_ordered(MARC::Field->new($id_field, $spec_value)); + $create = 1; + } +} + +my $cn = $record->field('001')->data(); +# Special handling of OCLC numbers, often found in records that lack 003 +if ($cn =~ /^oc[nm]/) { + $cn =~ s/^oc[nm]0*(\d+)/$1/; + $record->field('003')->data('OCoLC'); + $create = 0; +} + +# Now, if we need to munge the 001, we will first push the existing 001/003 +# into the 035; but if the record did not have one (and one only) 001 and 003 +# to begin with, skip this process +if ($munge and not $create) { + + my $scn = "(" . $record->field('003')->data() . ")" . $cn; + + # Do not create duplicate 035 fields + unless (grep $_->subfield('a') eq $scn, @scns) { + $record->insert_fields_ordered(MARC::Field->new('035', '', '', 'a' => $scn)); + } +} + +# Set the 001/003 and update the MARC +if ($create or $munge) { + $record->field('001')->data($rec_id); + $record->field('003')->data($ou_cni); + + my $xml = $record->as_xml_record(); + $xml =~ s/\n//sgo; + $xml =~ s/^<\?xml.+\?\s*>//go; + $xml =~ s/>\s+entityize() + # to avoid having to set PERL5LIB for PostgreSQL as well + + # If we are going to convert non-ASCII characters to XML entities, + # we had better be dealing with a UTF8 string to begin with + $xml = decode_utf8($xml); + + $xml = NFC($xml); + + # Convert raw ampersands to entities + $xml =~ s/&(?!\S+;)/&/gso; + + # Convert Unicode characters to entities + $xml =~ s/([\x{0080}-\x{fffd}])/sprintf('&#x%X;',ord($1))/sgoe; + + $xml =~ s/[\x00-\x1f]//go; + $_TD->{new}{marc} = $xml; + + return "MODIFY"; +} + +return; +$func$ LANGUAGE PLPERLU; + +INSERT INTO config.metabib_field ( id, field_class, name, label, format, xpath) VALUES + (29, 'identifier', 'scn', oils_i18n_gettext(28, 'System Control Number', 'cmf', 'label'), 'marcxml', $$//marc:datafield[@tag='035']/marc:subfield[@code="a"]$$); +INSERT INTO config.metabib_field ( id, field_class, name, label, format, xpath) VALUES + (30, 'identifier', 'lccn', oils_i18n_gettext(28, 'LC Control Number', 'cmf', 'label'), 'marcxml', $$//marc:datafield[@tag='010']/marc:subfield[@code="a" or @code='z']$$); + +-- Far from perfect, but much faster than reingesting every record +INSERT INTO metabib.identifier_field_entry(source, field, value) + SELECT record, 29, value FROM metabib.full_rec WHERE tag = '035' AND subfield = 'a'; +INSERT INTO metabib.identifier_field_entry(source, field, value) + SELECT record, 30, value FROM metabib.full_rec WHERE tag = '010' AND subfield IN ('a', 'z'); + +COMMIT; diff --git a/Open-ILS/src/sql/Pg/upgrade/xxxx.maintain_control_numbers_with_OCLC_awareness.sql b/Open-ILS/src/sql/Pg/upgrade/xxxx.maintain_control_numbers_with_OCLC_awareness.sql deleted file mode 100644 index 784744de72..0000000000 --- a/Open-ILS/src/sql/Pg/upgrade/xxxx.maintain_control_numbers_with_OCLC_awareness.sql +++ /dev/null @@ -1,143 +0,0 @@ -BEGIN; - -CREATE OR REPLACE FUNCTION maintain_control_numbers() RETURNS TRIGGER AS $func$ -use strict; -use MARC::Record; -use MARC::File::XML (BinaryEncoding => 'UTF-8'); -use MARC::Charset; -use Encode; -use Unicode::Normalize; - -MARC::Charset->assume_unicode(1); - -my $record = MARC::Record->new_from_xml($_TD->{new}{marc}); -my $schema = $_TD->{table_schema}; -my $rec_id = $_TD->{new}{id}; - -# Short-circuit if maintaining control numbers per MARC21 spec is not enabled -my $enable = spi_exec_query("SELECT enabled FROM config.global_flag WHERE name = 'cat.maintain_control_numbers'"); -if (!($enable->{processed}) or $enable->{rows}[0]->{enabled} eq 'f') { - return; -} - -# Get the control number identifier from an OU setting based on $_TD->{new}{owner} -my $ou_cni = 'EVRGRN'; - -my $owner; -if ($schema eq 'serial') { - $owner = $_TD->{new}{owning_lib}; -} else { - # are.owner and bre.owner can be null, so fall back to the consortial setting - $owner = $_TD->{new}{owner} || 1; -} - -my $ous_rv = spi_exec_query("SELECT value FROM actor.org_unit_ancestor_setting('cat.marc_control_number_identifier', $owner)"); -if ($ous_rv->{processed}) { - $ou_cni = $ous_rv->{rows}[0]->{value}; - $ou_cni =~ s/"//g; # Stupid VIM syntax highlighting" -} else { - # Fall back to the shortname of the OU if there was no OU setting - $ous_rv = spi_exec_query("SELECT shortname FROM actor.org_unit WHERE id = $owner"); - if ($ous_rv->{processed}) { - $ou_cni = $ous_rv->{rows}[0]->{shortname}; - } -} - -my ($create, $munge) = (0, 0); - -my @scns = $record->field('035'); - -foreach my $id_field ('001', '003') { - my $spec_value; - my @controls = $record->field($id_field); - - if ($id_field eq '001') { - $spec_value = $rec_id; - } else { - $spec_value = $ou_cni; - } - - # Create the 001/003 if none exist - if (scalar(@controls) == 1) { - # Only one field; check to see if we need to munge it - unless (grep $_->data() eq $spec_value, @controls) { - $munge = 1; - } - } else { - # Delete the other fields, as with more than 1 001/003 we do not know which 003/001 to match - foreach my $control (@controls) { - $record->delete_field($control); - } - $record->insert_fields_ordered(MARC::Field->new($id_field, $spec_value)); - $create = 1; - } -} - -my $cn = $record->field('001')->data(); -# Special handling of OCLC numbers, often found in records that lack 003 -if ($cn =~ /^oc[nm]/) { - $cn =~ s/^oc[nm]0*(\d+)/$1/; - $record->field('003')->data('OCoLC'); - $create = 0; -} - -# Now, if we need to munge the 001, we will first push the existing 001/003 -# into the 035; but if the record did not have one (and one only) 001 and 003 -# to begin with, skip this process -if ($munge and not $create) { - - my $scn = "(" . $record->field('003')->data() . ")" . $cn; - - # Do not create duplicate 035 fields - unless (grep $_->subfield('a') eq $scn, @scns) { - $record->insert_fields_ordered(MARC::Field->new('035', '', '', 'a' => $scn)); - } -} - -# Set the 001/003 and update the MARC -if ($create or $munge) { - $record->field('001')->data($rec_id); - $record->field('003')->data($ou_cni); - - my $xml = $record->as_xml_record(); - $xml =~ s/\n//sgo; - $xml =~ s/^<\?xml.+\?\s*>//go; - $xml =~ s/>\s+entityize() - # to avoid having to set PERL5LIB for PostgreSQL as well - - # If we are going to convert non-ASCII characters to XML entities, - # we had better be dealing with a UTF8 string to begin with - $xml = decode_utf8($xml); - - $xml = NFC($xml); - - # Convert raw ampersands to entities - $xml =~ s/&(?!\S+;)/&/gso; - - # Convert Unicode characters to entities - $xml =~ s/([\x{0080}-\x{fffd}])/sprintf('&#x%X;',ord($1))/sgoe; - - $xml =~ s/[\x00-\x1f]//go; - $_TD->{new}{marc} = $xml; - - return "MODIFY"; -} - -return; -$func$ LANGUAGE PLPERLU; - -INSERT INTO config.metabib_field ( id, field_class, name, label, format, xpath) VALUES - (29, 'identifier', 'scn', oils_i18n_gettext(28, 'System Control Number', 'cmf', 'label'), 'marcxml', $$//marc:datafield[@tag='035']/marc:subfield[@code="a"]$$); -INSERT INTO config.metabib_field ( id, field_class, name, label, format, xpath) VALUES - (30, 'identifier', 'lccn', oils_i18n_gettext(28, 'LC Control Number', 'cmf', 'label'), 'marcxml', $$//marc:datafield[@tag='010']/marc:subfield[@code="a" or @code='z']$$); - --- Far from perfect, but much faster than reingesting every record -INSERT INTO metabib.identifier_field_entry(source, field, value) - SELECT record, 29, value FROM metabib.full_rec WHERE tag = '035' AND subfield = 'a'; -INSERT INTO metabib.identifier_field_entry(source, field, value) - SELECT record, 30, value FROM metabib.full_rec WHERE tag = '010' AND subfield IN ('a', 'z'); - -COMMIT;