add xml support to discoverlife taxonomic name indexing; related to #161

globalbioticinteractions · Apr 5, 2024 · 2433824 · 2433824
1 parent 41c5a7f
commit 2433824
Show file tree

Hide file tree

Showing 10 changed files with 93 additions and 98 deletions.
diff --git a/...org/eol/globi/taxon/DiscoverLifeUtil.java → ...ol/globi/taxon/DiscoverLifeUtilXHTML.java b/...org/eol/globi/taxon/DiscoverLifeUtil.java → ...ol/globi/taxon/DiscoverLifeUtilXHTML.java
@@ -34,7 +34,7 @@
 import java.util.regex.Pattern;
 import java.util.stream.Collectors;
 
-public class DiscoverLifeUtil {
+public class DiscoverLifeUtilXHTML {
 
     private static final List<String> PATH_STATIC = Arrays.asList("Animalia", "Arthropoda", "Insecta", "Hymenoptera");
     public static final String URL_ENDPOINT_DISCOVER_LIFE = "https://www.discoverlife.org";

diff --git a/...rg/eol/globi/taxon/DiscoverLifeUtil2.java → .../eol/globi/taxon/DiscoverLifeUtilXML.java b/...rg/eol/globi/taxon/DiscoverLifeUtil2.java → .../eol/globi/taxon/DiscoverLifeUtilXML.java
@@ -39,9 +39,9 @@
 import java.util.stream.Stream;
 
 import static org.eol.globi.service.TaxonUtil.generateTaxonPathNames;
-import static org.eol.globi.taxon.DiscoverLifeUtil.emitNameRelatedToFocalTaxon;
+import static org.eol.globi.taxon.DiscoverLifeUtilXHTML.emitNameRelatedToFocalTaxon;
 
-public class DiscoverLifeUtil2 {
+public class DiscoverLifeUtilXML {
 
     public static final List<String> RANKS = Arrays.asList("Family", "Subfamily", "Tribe", "Subtribe", "Genus", "Subgenus");
 
@@ -112,7 +112,7 @@ public void accept(String recordXml) {
 
                     Document doc = builder.parse(recordInputStream);
                     Map<String, String> nameMap = parseFocalTaxon(doc);
-                    DiscoverLifeUtil.emitNameRelation(listener, nameMap, TaxonUtil.mapToTaxon(nameMap));
+                    DiscoverLifeUtilXHTML.emitNameRelation(listener, nameMap, TaxonUtil.mapToTaxon(nameMap));
 
                     List<Taxon> relatedTaxa = parseRelatedNames(doc);
 
@@ -162,7 +162,7 @@ static Map<String, String> parseFocalTaxon(Document doc) throws XPathExpressionE
         nameMap.put(PropertyAndValueDictionary.RANK, taxonomicRank);
         String name = nameMap.get(PropertyAndValueDictionary.NAME);
         nameMap.put(taxonomicRank, name);
-        nameMap.put(PropertyAndValueDictionary.EXTERNAL_ID, DiscoverLifeUtil.URL_ENDPOINT_DISCOVER_LIFE_SEARCH + StringUtils.replace(name, " ", "+"));
+        nameMap.put(PropertyAndValueDictionary.EXTERNAL_ID, DiscoverLifeUtilXHTML.URL_ENDPOINT_DISCOVER_LIFE_SEARCH + StringUtils.replace(name, " ", "+"));
 
 
         NodeList attr = (NodeList) XmlUtil.applyXPath(setNode, "//attributes", XPathConstants.NODESET);
@@ -231,7 +231,7 @@ public static List<Taxon> parseRelatedNames(Document doc) throws XPathExpression
                         List<String> namesWithoutRemarks = Arrays.asList(nameParts).subList(0, nameParts.length > 2 ? nameParts.length - 1 : nameParts.length);
                         return StringUtils.join(namesWithoutRemarks, ",");
                     })
-                    .map(DiscoverLifeUtil2::parse)
+                    .map(DiscoverLifeUtilXML::parse)
                     .filter(Objects::nonNull);
         }
         return relatedNames.collect(Collectors.toList());

diff --git a/...axon/DiscoverLifeUtilIntegrationTest.java → ...DiscoverLifeUtilXHTMLIntegrationTest.java b/...axon/DiscoverLifeUtilIntegrationTest.java → ...DiscoverLifeUtilXHTMLIntegrationTest.java
@@ -1,34 +1,23 @@
 package org.eol.globi.taxon;
 
-import org.apache.commons.io.IOUtils;
-import org.eol.globi.domain.NameType;
 import org.eol.globi.domain.Taxon;
-import org.eol.globi.domain.Term;
 import org.eol.globi.service.ResourceService;
 import org.hamcrest.core.Is;
 import org.junit.Test;
-import org.w3c.dom.NodeList;
-import org.xml.sax.SAXException;
 
-import javax.xml.parsers.ParserConfigurationException;
-import javax.xml.xpath.XPathConstants;
-import javax.xml.xpath.XPathExpressionException;
 import java.io.IOException;
 import java.io.InputStream;
 import java.net.URI;
-import java.nio.charset.StandardCharsets;
-import java.util.ArrayList;
-import java.util.List;
 import java.util.concurrent.atomic.AtomicInteger;
 import java.util.concurrent.atomic.AtomicReference;
 
 import static org.hamcrest.MatcherAssert.assertThat;
 import static org.junit.Assert.assertNull;
 
-public class DiscoverLifeUtilIntegrationTest {
+public class DiscoverLifeUtilXHTMLIntegrationTest {
 
     private static final String DISCOVER_LIFE_URL
-            = DiscoverLifeUtil.URL_ENDPOINT_DISCOVER_LIFE +
+            = DiscoverLifeUtilXHTML.URL_ENDPOINT_DISCOVER_LIFE +
             "/mp/20q" +
             "?act=x_checklist" +
             "&guide=Apoidea_species" +
@@ -50,7 +39,7 @@ public void parseBees() throws IOException {
             }
         };
 
-        DiscoverLifeUtil.parse(DiscoverLifeUtil.getBeeNameTable(new ResourceService() {
+        DiscoverLifeUtilXHTML.parse(DiscoverLifeUtilXHTML.getBeeNameTable(new ResourceService() {
             @Override
             public InputStream retrieve(URI uri) throws IOException {
                 return DiscoverLifeTestUtil.getStreamOfBees(BEE_NAMES_CACHED);

diff --git a/...eol/globi/taxon/DiscoverLifeUtilTest.java → ...lobi/taxon/DiscoverLifeUtilXHTMLTest.java b/...eol/globi/taxon/DiscoverLifeUtilTest.java → ...lobi/taxon/DiscoverLifeUtilXHTMLTest.java
@@ -6,7 +6,6 @@
 import org.eol.globi.domain.NameType;
 import org.eol.globi.domain.Taxon;
 import org.eol.globi.domain.Term;
-import org.eol.globi.service.ResourceService;
 import org.hamcrest.core.Is;
 import org.junit.Test;
 import org.w3c.dom.NodeList;
@@ -17,17 +16,15 @@
 import javax.xml.xpath.XPathExpressionException;
 import java.io.IOException;
 import java.io.InputStream;
-import java.net.URI;
 import java.nio.charset.StandardCharsets;
 import java.util.ArrayList;
 import java.util.List;
-import java.util.concurrent.atomic.AtomicInteger;
 import java.util.concurrent.atomic.AtomicReference;
 
 import static org.hamcrest.MatcherAssert.assertThat;
 import static org.junit.Assert.assertNull;
 
-public class DiscoverLifeUtilTest {
+public class DiscoverLifeUtilXHTMLTest {
 
     @Test
     public void parseNameAncylandrenaAtoposoma() throws SAXException, ParserConfigurationException, XPathExpressionException, IOException {
@@ -61,7 +58,7 @@ public void parseNameAncylandrenaAtoposoma() throws SAXException, ParserConfigur
 
         List<Triple<Term, NameType, Taxon>> relatedTaxa = new ArrayList<>();
 
-        DiscoverLifeUtil.parseNames(null, nodes.item(0), new TermMatchListener() {
+        DiscoverLifeUtilXHTML.parseNames(null, nodes.item(0), new TermMatchListener() {
 
             @Override
             public void foundTaxonForTerm(Long requestId, Term providedTerm, NameType nameType, Taxon resolvedTaxon) {
@@ -208,7 +205,7 @@ public void parseNameRelationsAndrenaAccepta() throws SAXException, ParserConfig
 
         List<Triple<Term, NameType, Taxon>> relatedTaxa = new ArrayList<>();
 
-        DiscoverLifeUtil.parseNames(null, nodes.item(0), new TermMatchListener() {
+        DiscoverLifeUtilXHTML.parseNames(null, nodes.item(0), new TermMatchListener() {
 
             @Override
             public void foundTaxonForTerm(Long requestId, Term providedTerm, NameType nameType, Taxon resolvedTaxon) {
@@ -272,7 +269,7 @@ public void parseHomonymAllodapeClypeata() throws SAXException, ParserConfigurat
 
         List<Triple<Term, NameType, Taxon>> relatedTaxa = new ArrayList<>();
 
-        DiscoverLifeUtil.parseNames(null, nodes.item(0), new TermMatchListener() {
+        DiscoverLifeUtilXHTML.parseNames(null, nodes.item(0), new TermMatchListener() {
 
             @Override
             public void foundTaxonForTerm(Long requestId, Term providedTerm, NameType nameType, Taxon resolvedTaxon) {
@@ -339,7 +336,7 @@ public void parseNameRelationsDanglingVar() throws SAXException, ParserConfigura
 
         List<Triple<Term, NameType, Taxon>> relatedTaxa = new ArrayList<>();
 
-        DiscoverLifeUtil.parseNames(null, nodes.item(0), new TermMatchListener() {
+        DiscoverLifeUtilXHTML.parseNames(null, nodes.item(0), new TermMatchListener() {
 
             @Override
             public void foundTaxonForTerm(Long requestId, Term providedTerm, NameType nameType, Taxon resolvedTaxon) {
@@ -393,7 +390,7 @@ public void parseNameRelationsDanglingVarWithLeadingComma() throws SAXException,
 
         List<Triple<Term, NameType, Taxon>> relatedTaxa = new ArrayList<>();
 
-        DiscoverLifeUtil.parseNames(null, nodes.item(0), new TermMatchListener() {
+        DiscoverLifeUtilXHTML.parseNames(null, nodes.item(0), new TermMatchListener() {
 
             @Override
             public void foundTaxonForTerm(Long requestId, Term providedTerm, NameType nameType, Taxon resolvedTaxon) {
@@ -445,7 +442,7 @@ public void parseNameAcceptedHomonym() throws SAXException, ParserConfigurationE
 
         List<Triple<Term, NameType, Taxon>> relatedTaxa = new ArrayList<>();
 
-        DiscoverLifeUtil.parseNames(null, nodes.item(0), new TermMatchListener() {
+        DiscoverLifeUtilXHTML.parseNames(null, nodes.item(0), new TermMatchListener() {
 
             @Override
             public void foundTaxonForTerm(Long requestId, Term providedTerm, NameType nameType, Taxon resolvedTaxon) {
@@ -557,7 +554,7 @@ public void parseNameRelationsWithHomonym() throws SAXException, ParserConfigura
 
         List<Triple<Term, NameType, Taxon>> relatedTaxa = new ArrayList<>();
 
-        DiscoverLifeUtil.parseNames(null, nodes.item(0), new TermMatchListener() {
+        DiscoverLifeUtilXHTML.parseNames(null, nodes.item(0), new TermMatchListener() {
 
             @Override
             public void foundTaxonForTerm(Long requestId, Term providedTerm, NameType nameType, Taxon resolvedTaxon) {
@@ -614,7 +611,7 @@ public void parseNameRelationsWithSicSuffix() throws SAXException, ParserConfigu
 
         List<Triple<Term, NameType, Taxon>> relatedTaxa = new ArrayList<>();
 
-        DiscoverLifeUtil.parseNames(null, nodes.item(0), new TermMatchListener() {
+        DiscoverLifeUtilXHTML.parseNames(null, nodes.item(0), new TermMatchListener() {
 
             @Override
             public void foundTaxonForTerm(Long requestId, Term providedTerm, NameType nameType, Taxon resolvedTaxon) {
@@ -665,7 +662,7 @@ public void parseNameRelationsWithSicSuffix2() throws SAXException, ParserConfig
 
         List<Triple<Term, NameType, Taxon>> relatedTaxa = new ArrayList<>();
 
-        DiscoverLifeUtil.parseNames(null, nodes.item(0), new TermMatchListener() {
+        DiscoverLifeUtilXHTML.parseNames(null, nodes.item(0), new TermMatchListener() {
 
             @Override
             public void foundTaxonForTerm(Long requestId, Term providedTerm, NameType nameType, Taxon resolvedTaxon) {
@@ -702,7 +699,7 @@ public void parseNamePseudopanurgus_parvus() throws SAXException, ParserConfigur
 
         List<Triple<Term, NameType, Taxon>> relatedTaxa = new ArrayList<>();
 
-        DiscoverLifeUtil.parseNames(null, nodes.item(0), new TermMatchListener() {
+        DiscoverLifeUtilXHTML.parseNames(null, nodes.item(0), new TermMatchListener() {
 
             @Override
             public void foundTaxonForTerm(Long requestId, Term providedTerm, NameType nameType, Taxon resolvedTaxon) {
@@ -729,27 +726,27 @@ public void foundTaxonForTerm(Long requestId, Term providedTerm, NameType nameTy
 
     @Test
     public void guessRank() throws IOException {
-        assertThat(DiscoverLifeUtil.guessRankFromName("Bla bla"), Is.is("species"));
+        assertThat(DiscoverLifeUtilXHTML.guessRankFromName("Bla bla"), Is.is("species"));
     }
 
     @Test
     public void guessRankFamily() throws IOException {
-        assertThat(DiscoverLifeUtil.guessRankFromName("Bla (Bla) bla"), Is.is("species"));
+        assertThat(DiscoverLifeUtilXHTML.guessRankFromName("Bla (Bla) bla"), Is.is("species"));
     }
 
     @Test
     public void guessRankSubspecies() throws IOException {
-        assertThat(DiscoverLifeUtil.guessRankFromName("Bla bla bla"), Is.is("subspecies"));
+        assertThat(DiscoverLifeUtilXHTML.guessRankFromName("Bla bla bla"), Is.is("subspecies"));
     }
 
     @Test
     public void guessRankVariant() throws IOException {
-        assertThat(DiscoverLifeUtil.guessRankFromName("Bla bla var bla"), Is.is("variety"));
+        assertThat(DiscoverLifeUtilXHTML.guessRankFromName("Bla bla var bla"), Is.is("variety"));
     }
 
     @Test
     public void guessRankSubvariant() throws IOException {
-        assertThat(DiscoverLifeUtil.guessRankFromName("Bla bla bla var bla"), Is.is("subvariety"));
+        assertThat(DiscoverLifeUtilXHTML.guessRankFromName("Bla bla bla var bla"), Is.is("subvariety"));
     }
 
     @Test
@@ -770,7 +767,7 @@ public void foundTaxonForTerm(Long requestId, Term providedTerm, NameType nameTy
             }
         };
 
-        DiscoverLifeUtil.parseTaxonPage(is, termMatchListener);
+        DiscoverLifeUtilXHTML.parseTaxonPage(is, termMatchListener);
 
 
         assertThat(acceptedFound.get().getName(), Is.is("Agapostemon texanus"));

diff --git a/...xon/DiscoverLifeUtil2IntegrationTest.java → ...n/DiscoverLifeUtilXMLIntegrationTest.java b/...xon/DiscoverLifeUtil2IntegrationTest.java → ...n/DiscoverLifeUtilXMLIntegrationTest.java
@@ -3,7 +3,6 @@
 import org.eol.globi.domain.NameType;
 import org.eol.globi.domain.Taxon;
 import org.eol.globi.domain.Term;
-import org.eol.globi.service.TaxonUtil;
 import org.hamcrest.core.Is;
 import org.junit.Test;
 
@@ -14,22 +13,22 @@
 
 import static org.hamcrest.MatcherAssert.assertThat;
 
-public class DiscoverLifeUtil2IntegrationTest {
+public class DiscoverLifeUtilXMLIntegrationTest {
 
     public static final String BEES_XML_GZIP = "/org/globalbioticinteractions/nomer/match/discoverlife/bees.xml.gz";
 
     @Test
     public void compareLocalVersionToRemoteVersion() throws IOException {
         DiscoverLifeTestUtil.compareLocalVersionToRemoteVersion(
                 BEES_XML_GZIP,
-                DiscoverLifeUtil.URL_ENDPOINT_DISCOVER_LIFE + "/nh/id/20q/Apoidea_species.xml"
+                DiscoverLifeUtilXHTML.URL_ENDPOINT_DISCOVER_LIFE + "/nh/id/20q/Apoidea_species.xml"
         );
     }
 
     @Test
     public void parseNames() throws ParserConfigurationException, IOException {
         AtomicInteger counter = new AtomicInteger(0);
-        DiscoverLifeUtil2.parse(new GZIPInputStream(getClass().getResourceAsStream(BEES_XML_GZIP)), new TermMatchListener() {
+        DiscoverLifeUtilXML.parse(new GZIPInputStream(getClass().getResourceAsStream(BEES_XML_GZIP)), new TermMatchListener() {
             @Override
             public void foundTaxonForTerm(Long requestId, Term providedTerm, NameType nameType, Taxon resolvedTaxon) {
                 counter.incrementAndGet();