hazza/DSpace
1
0
Fork 0
mirror of https://github.com/DSpace/DSpace.git synced 2025-10-07 01:54:22 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #2698 from toniprieto/DS-4034-strip-diacritics-master

Browse Source 
DS-4034 - The "first few letters" search doesn't work with diacritics (master port)
This commit is contained in:
Tim Donohue
2021-09-07 12:52:07 -05:00
committed by GitHub
parent 209f25e17c 8fb214c061
commit 4f9e9bce7a
4 changed files with 157 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
dspace-api/src/main/java/org/dspace/browse/BrowseEngine.java
View File

@@ -11,7 +11,6 @@ import java.sql.SQLException;
import java.util.ArrayList;
import java.util.List;
import org.apache.commons.lang3.StringUtils;
import org.apache.logging.log4j.Logger;
import org.dspace.content.Collection;
import org.dspace.content.Community;
@@ -381,7 +380,7 @@ public class BrowseEngine {
// this is the distinct table constrained to either community or collection
dao.setTable(browseIndex.getDistinctTableName());
dao.setStartsWith(StringUtils.lowerCase(scope.getStartsWith()));
dao.setStartsWith(normalizeJumpToValue(scope.getStartsWith()));
// remind the DAO that this is a distinct value browse, so it knows what sort
// of query to build
dao.setDistinct(true);

2
dspace-api/src/main/java/org/dspace/sort/OrderFormatAuthor.java
View File

@@ -9,6 +9,7 @@ package org.dspace.sort;
import org.dspace.text.filter.DecomposeDiactritics;
import org.dspace.text.filter.LowerCaseAndTrim;
import org.dspace.text.filter.StripDiacritics;
import org.dspace.text.filter.TextFilter;
/**
@@ -19,6 +20,7 @@ import org.dspace.text.filter.TextFilter;
public class OrderFormatAuthor extends AbstractTextFilterOFD {
{
filters = new TextFilter[] {new DecomposeDiactritics(),
new StripDiacritics(),
new LowerCaseAndTrim()};
}
}

2
dspace-api/src/main/java/org/dspace/sort/OrderFormatText.java
View File

@@ -9,6 +9,7 @@ package org.dspace.sort;
import org.dspace.text.filter.DecomposeDiactritics;
import org.dspace.text.filter.LowerCaseAndTrim;
import org.dspace.text.filter.StripDiacritics;
import org.dspace.text.filter.TextFilter;
/**
@@ -19,6 +20,7 @@ import org.dspace.text.filter.TextFilter;
public class OrderFormatText extends AbstractTextFilterOFD {
{
filters = new TextFilter[] {new DecomposeDiactritics(),
new StripDiacritics(),
new LowerCaseAndTrim()};
}
}

152
dspace-server-webapp/src/test/java/org/dspace/app/rest/BrowsesResourceControllerIT.java
View File

@@ -827,6 +827,158 @@ public class BrowsesResourceControllerIT extends AbstractControllerIntegrationTe
};
@Test
public void testBrowseByEntriesStartsWithAndDiacritics() throws Exception {
context.turnOffAuthorisationSystem();
//** GIVEN **
//1. A community-collection structure with one parent community with sub-community and two collections.
parentCommunity = CommunityBuilder.createCommunity(context)
.withName("Parent Community")
.build();
Community child1 = CommunityBuilder.createSubCommunity(context, parentCommunity)
.withName("Sub Community")
.build();
Collection col1 = CollectionBuilder.createCollection(context, child1).withName("Collection 1").build();
Collection col2 = CollectionBuilder.createCollection(context, child1).withName("Collection 2").build();
//2. 4 public items that are readable by Anonymous
Item item1 = ItemBuilder.createItem(context, col1)
.withTitle("Item1")
.withAuthor("Álvarez, Nombre")
.withIssueDate("1912-06-23")
.withSubject("Teléfono")
.build();
Item item2 = ItemBuilder.createItem(context, col1)
.withTitle("Item2")
.withAuthor("Ögren, Name")
.withIssueDate("1982-06-25")
.withSubject("Televisor")
.build();
Item item3 = ItemBuilder.createItem(context, col2)
.withTitle("Item3")
.withAuthor("Azuaga, Nombre")
.withIssueDate("1990")
.withSubject("Telecomunicaciones")
.build();
Item item4 = ItemBuilder.createItem(context, col2)
.withTitle("Item4")
.withAuthor("Alonso, Nombre")
.withAuthor("Ortiz, Nombre")
.withIssueDate("1995-05-23")
.withSubject("Guion")
.build();
// ---- BROWSES BY ENTRIES ----
//** WHEN **
//An anonymous user browses the entries in the Browse by Author endpoint
//with startsWith set to A
getClient().perform(get("/api/discover/browses/author/entries?startsWith=A")
.param("size", "4"))
//** THEN **
//The status has to be 200 OK
.andExpect(status().isOk())
//We expect the content type to be "application/hal+json;charset=UTF-8"
.andExpect(content().contentType(contentType))
//We expect 3 elements
.andExpect(jsonPath("$.page.totalElements", is(3)))
//As entry browsing works as a filter, we expect to be on page 0
.andExpect(jsonPath("$.page.number", is(0)))
//Verify that the index filters to the "Alonso, Nombre", "Álvarez, Nombre" and "Azuaga, Nombre"
// and diacritics are ignored in sorting
.andExpect(jsonPath("$._embedded.entries",
contains(BrowseEntryResourceMatcher.matchBrowseEntry("Alonso, Nombre", 1),
BrowseEntryResourceMatcher.matchBrowseEntry("Álvarez, Nombre", 1),
BrowseEntryResourceMatcher.matchBrowseEntry("Azuaga, Nombre", 1)
)))
//Verify startsWith parameter is included in the links
.andExpect(jsonPath("$._links.self.href", containsString("?startsWith=A")));
//** WHEN **
//An anonymous user browses the entries in the Browse by Author endpoint
//with startsWith set to Ú (accented)
getClient().perform(get("/api/discover/browses/author/entries?startsWith=Ó"))
//** THEN **
//The status has to be 200 OK
.andExpect(status().isOk())
//We expect the content type to be "application/hal+json;charset=UTF-8"
.andExpect(content().contentType(contentType))
//We expect 2 elements
.andExpect(jsonPath("$.page.totalElements", is(2)))
//As entry browsing works as a filter, we expect to be on page 0
.andExpect(jsonPath("$.page.number", is(0)))
//Verify that the index filters to the "Ögren, Name"" and "Ortiz, Nombre"
.andExpect(jsonPath("$._embedded.entries",
contains(BrowseEntryResourceMatcher.matchBrowseEntry("Ögren, Name", 1),
BrowseEntryResourceMatcher.matchBrowseEntry("Ortiz, Nombre", 1)
)))
//Verify that the startsWith paramater is included in the links
.andExpect(jsonPath("$._links.self.href", containsString("?startsWith=Ó")));
//** WHEN **
//An anonymous user browses the entries in the Browse by Subject endpoint
//with startsWith set to Cana
getClient().perform(get("/api/discover/browses/subject/entries?startsWith=Tele"))
//** THEN **
//The status has to be 200 OK
.andExpect(status().isOk())
//We expect the content type to be "application/hal+json;charset=UTF-8"
.andExpect(content().contentType(contentType))
//We expect 3 elements
.andExpect(jsonPath("$.page.totalElements", is(3)))
//As entry browsing works as a filter, we expect to be on page 0
.andExpect(jsonPath("$.page.number", is(0)))
//Verify that the index filters to the "Telecomunicaciones', "Teléfono" and "Televisor" and
// it is sorted ignoring diacritics
.andExpect(jsonPath("$._embedded.entries",
contains(BrowseEntryResourceMatcher.matchBrowseEntry("Telecomunicaciones", 1),
BrowseEntryResourceMatcher.matchBrowseEntry("Teléfono", 1),
BrowseEntryResourceMatcher.matchBrowseEntry("Televisor", 1)
)))
//Verify that the startsWith paramater is included in the links
.andExpect(jsonPath("$._links.self.href", containsString("?startsWith=Tele")));
//** WHEN **
//An anonymous user browses the entries in the Browse by Subject endpoint
//with startsWith set to Guión
getClient().perform(get("/api/discover/browses/subject/entries?startsWith=Guión"))
//** THEN **
//The status has to be 200 OK
.andExpect(status().isOk())
//We expect the content type to be "application/hal+json;charset=UTF-8"
.andExpect(content().contentType(contentType))
//We expect only the entry "Guion" to be present
.andExpect(jsonPath("$.page.totalElements", is(1)))
//As entry browsing works as a filter, we expect to be on page 0
.andExpect(jsonPath("$.page.number", is(0)))
//Verify that the index filters to the "Guion"
.andExpect(jsonPath("$._embedded.entries",
contains(BrowseEntryResourceMatcher.matchBrowseEntry("Guion", 1)
)))
//Verify that the startsWith paramater is included in the links
.andExpect(jsonPath("$._links.self.href", containsString("?startsWith=Guión")));
};
@Test
public void testBrowseByItemsStartsWith() throws Exception {
context.turnOffAuthorisationSystem();