move exporters into separate module

eol-globi-data-export/pom.xml

@@ -0,0 +1,28 @@
+<project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
+         xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
+    <modelVersion>4.0.0</modelVersion>
+
+    <parent>
+        <artifactId>eol-globi-parent</artifactId>
+        <groupId>org.eol</groupId>
+        <version>0.6-SNAPSHOT</version>
+        <relativePath>../eol-globi-parent/pom.xml</relativePath>
+    </parent>
+
+    <artifactId>eol-globi-data-export</artifactId>
+    <name>EOL's Global Biotic Interaction Data Exporters</name>
+
+    <dependencies>
+        <dependency>
+            <groupId>${project.groupId}</groupId>
+            <artifactId>eol-globi-lib</artifactId>
+            <version>${project.version}</version>
+        </dependency>
+        <dependency>
+            <groupId>org.mapdb</groupId>
+            <artifactId>mapdb</artifactId>
+            <version>${mapdb.version}</version>
+        </dependency>
+    </dependencies>
+
+</project>

eol-globi-data-tool/src/main/java/org/eol/globi/export/ExporterTaxaRollup.java → eol-globi-data-export/src/main/java/org/eol/globi/export/ExporterTaxaRollup.java

@@ -1,5 +1,6 @@
 package org.eol.globi.export;
 
+import org.apache.commons.lang.StringUtils;
 import org.eol.globi.data.CharsetConstant;
 import org.eol.globi.domain.PropertyAndValueDictionary;
 import org.eol.globi.domain.Study;
@@ -18,10 +19,6 @@
 import java.util.List;
 import java.util.Map;
 
-import static org.apache.commons.lang.StringUtils.join;
-import static org.apache.commons.lang.StringUtils.split;
-import static org.apache.commons.lang.StringUtils.trim;
-
 /**
  * Exports an expanded taxon list. Rolls up to taxonomic rank "order".
  * <p/>
@@ -38,22 +35,22 @@ public class ExporterTaxaRollup extends ExporterTaxa {
     public static List<Map<String, Object>> expandTaxonResult(Map<String, Object> result) {
         List<Map<String, Object>> expandedResults = new ArrayList<Map<String, Object>>();
 
-        String[] pathNames = split((String) result.get("pathNames"), CharsetConstant.SEPARATOR_CHAR);
-        String[] pathIds = split((String) result.get("pathIds"), "|");
-        String[] path = split((String) result.get("path"), "|");
+        String[] pathNames = StringUtils.split((String) result.get("pathNames"), CharsetConstant.SEPARATOR_CHAR);
+        String[] pathIds = StringUtils.split((String) result.get("pathIds"), "|");
+        String[] path = StringUtils.split((String) result.get("path"), "|");
         if (pathNames != null && pathIds != null && path != null
                 && pathNames.length == pathIds.length && pathNames.length == path.length) {
             for (int i = 0; i < pathNames.length; i++) {
-                String rank = trim(pathNames[i]);
+                String rank = StringUtils.trim(pathNames[i]);
                 if (EXPANDABLE_RANKS.contains(rank)) {
                     Map<String, Object> taxon = new HashMap<String, Object>();
-                    taxon.put("scientificName", trim(path[i]));
-                    taxon.put("taxonId", trim(pathIds[i]));
+                    taxon.put("scientificName", StringUtils.trim(path[i]));
+                    taxon.put("taxonId", StringUtils.trim(pathIds[i]));
                     taxon.put("rank", rank);
                     String[] partialPathNames = Arrays.copyOfRange(pathNames, 0, i + 1);
                     String[] partialPath = Arrays.copyOfRange(path, 0, i + 1);
-                    taxon.put("path", trim(join(partialPath, CharsetConstant.SEPARATOR_CHAR)));
-                    taxon.put("pathNames", trim(join(partialPathNames, CharsetConstant.SEPARATOR_CHAR)));
+                    taxon.put("path", StringUtils.trim(StringUtils.join(partialPath, CharsetConstant.SEPARATOR_CHAR)));
+                    taxon.put("pathNames", StringUtils.trim(StringUtils.join(partialPathNames, CharsetConstant.SEPARATOR_CHAR)));
                     expandedResults.add(taxon);
                 }
             }
@@ -89,7 +86,7 @@ private void exportRollupTaxa(Writer writer, GraphDatabaseService graphDatabase)
 
             Map<String, String> row = new HashMap<String, String>();
             for (Map<String, Object> taxon : taxonMap.values()) {
-                resultsToRow(row, taxon);
+                ExporterTaxa.resultsToRow(row, taxon);
                 writeProperties(writer, row);
                 row.clear();
             }

eol-globi-data-tool/src/main/java/org/eol/globi/export/GraphExporter.java → eol-globi-data-export/src/main/java/org/eol/globi/export/GraphExporter.java

@@ -7,9 +7,9 @@
 import org.apache.commons.io.FileUtils;
 import org.apache.commons.logging.Log;
 import org.apache.commons.logging.LogFactory;
-import org.eol.globi.data.NodeFactoryImpl;
 import org.eol.globi.data.StudyImporterException;
 import org.eol.globi.domain.Study;
+import org.eol.globi.util.NodeUtil;
 import org.neo4j.graphdb.GraphDatabaseService;
 
 import java.io.BufferedOutputStream;
@@ -33,10 +33,9 @@ public void export(GraphDatabaseService graphService, String baseDir) throws Stu
         } catch (IOException e) {
             throw new StudyImporterException("failed to create output dir [" + baseDir + "]", e);
         }
-        List<Study> studies = NodeFactoryImpl.findAllStudies(graphService);
+        List<Study> studies = NodeUtil.findAllStudies(graphService);
         exportDataOntology(studies, baseDir);
         exportUnmatchedTaxa(studies, baseDir);
-        //exportGoMexSI(studies, baseDir);
         exportDarwinCoreAggregatedByStudy(baseDir, studies);
         exportDarwinCoreAll(baseDir, studies);
     }
@@ -55,21 +54,6 @@ private void exportUnmatchedTaxa(List<Study> studies, String baseDir) throws Stu
         }
     }
 
-    // Provide simple data representation
-    private void exportGoMexSI(List<Study> studies, String baseDir) throws StudyImporterException {
-        try {
-            FileUtils.forceMkdir(new File(baseDir));
-            OutputStreamWriter writer = openStream(baseDir + "GoMexSIInteractionsTaxa.csv");
-            for (Study importedStudy : studies) {
-                boolean includeHeader = studies.indexOf(importedStudy) == 0;
-                new ExporterGoMexSI().exportStudy(importedStudy, writer, includeHeader);
-            }
-            closeStream(baseDir + "GoMexSIInteractionsTaxa.csv", writer);
-        } catch (IOException e) {
-            throw new StudyImporterException("failed to export GoMexSI", e);
-        }
-    }
-
     private void exportDarwinCoreAggregatedByStudy(String baseDir, List<Study> studies) throws StudyImporterException {
         exportDarwinCoreArchive(studies,
                 baseDir + "aggregatedByStudy/", new HashMap<String, DarwinCoreExporter>() {

eol-globi-data-sources/pom.xml

@@ -30,16 +30,6 @@
             <version>${project.version}</version>
             <type>pom</type>
         </dependency>
-        <dependency>
-            <groupId>org.apache.jena</groupId>
-            <artifactId>jena-tdb</artifactId>
-            <version>1.0.2</version>
-        </dependency>
-        <dependency>
-            <groupId>commons-logging</groupId>
-            <artifactId>commons-logging</artifactId>
-            <version>1.0.4</version>
-        </dependency>
         <dependency>
             <groupId>org.codehaus.swizzle</groupId>
             <artifactId>swizzle-stream</artifactId>

eol-globi-data-tool/pom.xml

@@ -13,21 +13,21 @@
     <name>EOL's Global Biotic Interaction Data Import/Export Tool</name>
 
     <dependencies>
-        <dependency>
+         <dependency>
             <groupId>${project.groupId}</groupId>
-            <artifactId>eol-globi-lib</artifactId>
+            <artifactId>eol-globi-enrich</artifactId>
             <version>${project.version}</version>
         </dependency>
 
         <dependency>
             <groupId>${project.groupId}</groupId>
-            <artifactId>eol-globi-enrich</artifactId>
+            <artifactId>eol-globi-data-sources</artifactId>
             <version>${project.version}</version>
         </dependency>
 
         <dependency>
             <groupId>${project.groupId}</groupId>
-            <artifactId>eol-globi-data-sources</artifactId>
+            <artifactId>eol-globi-data-export</artifactId>
             <version>${project.version}</version>
         </dependency>
 
@@ -58,12 +58,6 @@
             <type>test-jar</type>
             <scope>test</scope>
         </dependency>
-
-        <dependency>
-            <groupId>commons-collections</groupId>
-            <artifactId>commons-collections</artifactId>
-            <version>3.2.1</version>
-        </dependency>
     </dependencies>
 
 

eol-globi-data-tool/src/main/java/org/eol/globi/data/NodeFactoryImpl.java

@@ -83,16 +83,6 @@ public NodeFactoryImpl(GraphDatabaseService graphDb, TaxonIndex taxonIndex) {
 
     }
 
-    public static List<Study> findAllStudies(GraphDatabaseService graphService) {
-        List<Study> studies = new ArrayList<Study>();
-        Index<Node> studyIndex = graphService.index().forNodes("studies");
-        IndexHits<Node> hits = studyIndex.query("title", "*");
-        for (Node hit : hits) {
-            studies.add(new Study(hit));
-        }
-        return studies;
-    }
-
     public GraphDatabaseService getGraphDb() {
         return graphDb;
     }

eol-globi-data-tool/src/test/java/org/eol/globi/data/StudyImporterForBarnesTest.java

@@ -5,6 +5,7 @@
 import org.eol.globi.domain.Specimen;
 import org.eol.globi.domain.Study;
 import org.eol.globi.domain.TaxonNode;
+import org.eol.globi.util.NodeUtil;
 import org.junit.Test;
 import org.neo4j.graphdb.Direction;
 import org.neo4j.graphdb.Node;
@@ -94,7 +95,7 @@ public boolean shouldImportRecord(Long recordNumber) {
         });
         studyImporterForBarnes.importStudy();
 
-        List<Study> studies = NodeFactoryImpl.findAllStudies(getGraphDb());
+        List<Study> studies = NodeUtil.findAllStudies(getGraphDb());
         assertTrue(studies.size() > 0);
         for (Study study : studies) {
             assertThat(study.getTitle(), is(notNullValue()));

eol-globi-data-tool/src/test/java/org/eol/globi/data/StudyImporterForBellTest.java

@@ -1,6 +1,7 @@
 package org.eol.globi.data;
 
 import org.eol.globi.domain.Study;
+import org.eol.globi.util.NodeUtil;
 import org.junit.Test;
 import org.neo4j.cypher.javacompat.ExecutionEngine;
 import org.neo4j.cypher.javacompat.ExecutionResult;
@@ -18,7 +19,7 @@ public class StudyImporterForBellTest extends GraphDBTestCase {
     public void importAll() throws StudyImporterException, NodeFactoryException {
         StudyImporter importer = new StudyImporterForBell(new ParserFactoryImpl(), nodeFactory);
         importer.importStudy();
-        List<Study> allStudies = NodeFactoryImpl.findAllStudies(getGraphDb());
+        List<Study> allStudies = NodeUtil.findAllStudies(getGraphDb());
         assertThat(allStudies.size(), is(6));
         ExecutionResult execute = new ExecutionEngine(getGraphDb()).execute("START taxon = node:taxons('*:*') RETURN taxon.name");
         String actual = execute.dumpToString();

eol-globi-data-tool/src/test/java/org/eol/globi/data/StudyImporterForBroseTest.java

@@ -7,6 +7,7 @@
 import org.eol.globi.domain.Specimen;
 import org.eol.globi.domain.Study;
 import org.eol.globi.domain.TaxonNode;
+import org.eol.globi.util.NodeUtil;
 import org.junit.Test;
 import org.neo4j.graphdb.Direction;
 import org.neo4j.graphdb.Node;
@@ -81,7 +82,7 @@ public void importAll() throws StudyImporterException {
         StudyImporterForBrose studyImporterForBrose = new StudyImporterForBrose(new ParserFactoryImpl(), nodeFactory);
         studyImporterForBrose.importStudy();
 
-        List<Study> studies = NodeFactoryImpl.findAllStudies(getGraphDb());
+        List<Study> studies = NodeUtil.findAllStudies(getGraphDb());
         assertThat(studies.size(), is(20));
         for (Study study : studies) {
             assertThat(study.getTitle(), is(notNullValue()));

eol-globi-data-tool/src/test/java/org/eol/globi/data/StudyImporterForByrnesTest.java

@@ -1,6 +1,7 @@
 package org.eol.globi.data;
 
 import org.eol.globi.domain.Study;
+import org.eol.globi.util.NodeUtil;
 import org.junit.Test;
 import org.neo4j.cypher.javacompat.ExecutionEngine;
 import org.neo4j.cypher.javacompat.ExecutionResult;
@@ -28,7 +29,7 @@ public void importAll() throws StudyImporterException, NodeFactoryException {
 
         List<String> citationList = new ArrayList<String>();
         Set<String> citations = new HashSet<String>();
-        List<Study> studies = NodeFactoryImpl.findAllStudies(getGraphDb());
+        List<Study> studies = NodeUtil.findAllStudies(getGraphDb());
         assertTrue(studies.size() > 0);
         for (Study study : studies) {
             assertThat(study.getTitle(), is(notNullValue()));