From f67f85dd9dca7bb3f0683ccbe33a7b03bcbdd289 Mon Sep 17 00:00:00 2001
From: Dan Scott <dscott@laurentian.ca>
Date: Thu, 10 Nov 2011 01:05:40 -0500
Subject: [PATCH] Index leading articles with apostrophes correctly
MIME-Version: 1.0
Content-Type: text/plain; charset=utf8
Content-Transfer-Encoding: 8bit

Titles with non-filing characters used to get separated by spaces; not a
problem in English with leading articles like "The" and "A" which are
followed by spaces anyway, but a significant problem for French where
"L'année" is not supposed to be turned into "L' année".

To combat this, we define a titleNonfiling variant in the MODS32
stylesheets that ignores the second indicator and does not wrap leading
articles in the nonSort element, avoiding the injection of unwanted
spaces.

Signed-off-by: Dan Scott <dscott@laurentian.ca>
---
 Open-ILS/src/sql/Pg/950.data.seed-values.sql |  2 +-
 Open-ILS/src/sql/Pg/953.data.MODS32-xsl.sql  | 36 ++++++++++++++++++++
 Open-ILS/xsl/MARC21slim2MODS32.xsl           | 36 ++++++++++++++++++++
 3 files changed, 73 insertions(+), 1 deletion(-)

diff --git a/Open-ILS/src/sql/Pg/950.data.seed-values.sql b/Open-ILS/src/sql/Pg/950.data.seed-values.sql
index 686471e585..e032835fb5 100644
--- a/Open-ILS/src/sql/Pg/950.data.seed-values.sql
+++ b/Open-ILS/src/sql/Pg/950.data.seed-values.sql
@@ -47,7 +47,7 @@ INSERT INTO config.metabib_field ( id, field_class, name, label, format, xpath )
 INSERT INTO config.metabib_field ( id, field_class, name, label, format, xpath ) VALUES 
     (5, 'title', 'uniform', oils_i18n_gettext(5, 'Uniform Title', 'cmf', 'label'), 'mods32', $$//mods32:mods/mods32:titleInfo[mods32:title and (@type='uniform')]$$ );
 INSERT INTO config.metabib_field ( id, field_class, name, label, format, xpath ) VALUES 
-    (6, 'title', 'proper', oils_i18n_gettext(6, 'Title Proper', 'cmf', 'label'), 'mods32', $$//mods32:mods/mods32:titleInfo[mods32:title and not (@type)]$$ );
+    (6, 'title', 'proper', oils_i18n_gettext(6, 'Title Proper', 'cmf', 'label'), 'mods32', $$//mods32:mods/mods32:titleNonfiling[mods32:title and not (@type)]$$ );
 
 INSERT INTO config.metabib_field ( id, field_class, name, label, format, xpath, facet_xpath, facet_field ) VALUES 
     (7, 'author', 'corporate', oils_i18n_gettext(7, 'Corporate Author', 'cmf', 'label'), 'mods32', $$//mods32:mods/mods32:name[@type='corporate' and mods32:role/mods32:roleTerm[text()='creator']]$$, $$//*[local-name()='namePart']$$, TRUE ); -- /* to fool vim */;
diff --git a/Open-ILS/src/sql/Pg/953.data.MODS32-xsl.sql b/Open-ILS/src/sql/Pg/953.data.MODS32-xsl.sql
index b0f3e4eaac..e775b66b4c 100644
--- a/Open-ILS/src/sql/Pg/953.data.MODS32-xsl.sql
+++ b/Open-ILS/src/sql/Pg/953.data.MODS32-xsl.sql
@@ -130,6 +130,42 @@ Added Log Comment
 				</xsl:if>
 				<xsl:call-template name="part"></xsl:call-template>
 			</titleInfo>
+			<!-- A form of title that ignores non-filing characters; useful
+				 for not converting "L'Oreal" into "L' Oreal" at index time -->
+			<titleNonfiling>
+				<xsl:variable name="title">
+					<xsl:choose>
+						<xsl:when test="marc:subfield[@code='b']">
+							<xsl:call-template name="specialSubfieldSelect">
+								<xsl:with-param name="axis">b</xsl:with-param>
+								<xsl:with-param name="beforeCodes">afgk</xsl:with-param>
+							</xsl:call-template>
+						</xsl:when>
+						<xsl:otherwise>
+							<xsl:call-template name="subfieldSelect">
+								<xsl:with-param name="codes">abfgk</xsl:with-param>
+							</xsl:call-template>
+						</xsl:otherwise>
+					</xsl:choose>
+				</xsl:variable>
+				<title>
+					<xsl:value-of select="$title"/>
+				</title>
+				<xsl:if test="marc:subfield[@code='b']">
+					<subTitle>
+						<xsl:call-template name="chopPunctuation">
+							<xsl:with-param name="chopString">
+								<xsl:call-template name="specialSubfieldSelect">
+									<xsl:with-param name="axis">b</xsl:with-param>
+									<xsl:with-param name="anyCodes">b</xsl:with-param>
+									<xsl:with-param name="afterCodes">afgk</xsl:with-param>
+								</xsl:call-template>
+							</xsl:with-param>
+						</xsl:call-template>
+					</subTitle>
+				</xsl:if>
+				<xsl:call-template name="part"></xsl:call-template>
+			</titleNonfiling>
 		</xsl:for-each>
 		<xsl:for-each select="marc:datafield[@tag='210']">
 			<titleInfo type="abbreviated">
diff --git a/Open-ILS/xsl/MARC21slim2MODS32.xsl b/Open-ILS/xsl/MARC21slim2MODS32.xsl
index 98cea9780c..303c378359 100644
--- a/Open-ILS/xsl/MARC21slim2MODS32.xsl
+++ b/Open-ILS/xsl/MARC21slim2MODS32.xsl
@@ -130,6 +130,42 @@ Added Log Comment
 				</xsl:if>
 				<xsl:call-template name="part"></xsl:call-template>
 			</titleInfo>
+			<!-- A form of title that ignores non-filing characters; useful
+				 for not converting "L'Oreal" into "L' Oreal" at index time -->
+			<titleNonfiling>
+				<xsl:variable name="title">
+					<xsl:choose>
+						<xsl:when test="marc:subfield[@code='b']">
+							<xsl:call-template name="specialSubfieldSelect">
+								<xsl:with-param name="axis">b</xsl:with-param>
+								<xsl:with-param name="beforeCodes">afgk</xsl:with-param>
+							</xsl:call-template>
+						</xsl:when>
+						<xsl:otherwise>
+							<xsl:call-template name="subfieldSelect">
+								<xsl:with-param name="codes">abfgk</xsl:with-param>
+							</xsl:call-template>
+						</xsl:otherwise>
+					</xsl:choose>
+				</xsl:variable>
+				<title>
+					<xsl:value-of select="$title"/>
+				</title>
+				<xsl:if test="marc:subfield[@code='b']">
+					<subTitle>
+						<xsl:call-template name="chopPunctuation">
+							<xsl:with-param name="chopString">
+								<xsl:call-template name="specialSubfieldSelect">
+									<xsl:with-param name="axis">b</xsl:with-param>
+									<xsl:with-param name="anyCodes">b</xsl:with-param>
+									<xsl:with-param name="afterCodes">afgk</xsl:with-param>
+								</xsl:call-template>
+							</xsl:with-param>
+						</xsl:call-template>
+					</subTitle>
+				</xsl:if>
+				<xsl:call-template name="part"></xsl:call-template>
+			</titleNonfiling>
 		</xsl:for-each>
 		<xsl:for-each select="marc:datafield[@tag='210']">
 			<titleInfo type="abbreviated">
-- 
2.43.2