#1 initial data typer

irodsext-data-profiler/src/test/java/org/irodsext/dataprofiler/IrodsextDataProfilerServiceTest.java

@@ -3,13 +3,14 @@
  */
 package org.irodsext.dataprofiler;
 
-import static org.junit.Assert.fail;
-
 import java.util.Properties;
 
+import org.irods.jargon.core.connection.IRODSAccount;
 import org.irods.jargon.core.connection.JargonProperties;
 import org.irods.jargon.core.connection.SettableJargonProperties;
+import org.irods.jargon.core.pub.IRODSAccessObjectFactory;
 import org.irods.jargon.core.pub.IRODSFileSystem;
+import org.irods.jargon.extensions.dataprofiler.DataProfilerSettings;
 import org.irods.jargon.testutils.IRODSTestSetupUtilities;
 import org.irods.jargon.testutils.TestingPropertiesHelper;
 import org.irods.jargon.testutils.filemanip.ScratchFileUtils;
@@ -69,8 +70,21 @@ public void afterEach() throws Exception {
 	 * {@link org.irodsext.dataprofiler.IrodsextDataProfilerService#retrieveDataProfile(java.lang.String, org.irods.jargon.extensions.dataprofiler.DataProfilerSettings)}.
 	 */
 	@Test
-	public void testRetrieveDataProfileStringDataProfilerSettings() {
-		fail("Not yet implemented");
+	public void testBasicDataProfileWithCollection() throws Exception {
+		IRODSAccount irodsAccount = testingPropertiesHelper.buildIRODSAccountFromTestProperties(testingProperties);
+		IRODSAccessObjectFactory accessObjectFactory = irodsFileSystem.getIRODSAccessObjectFactory();
+		String targetIrodsCollection = testingPropertiesHelper
+				.buildIRODSCollectionAbsolutePathFromTestProperties(testingProperties, IRODS_TEST_SUBDIR_PATH);
+
+		DataProfilerSettings dataProfilerSettings = new DataProfilerSettings();
+		dataProfilerSettings.setDetectMimeAndInfoType(false);
+		dataProfilerSettings.setRetrieveAcls(true);
+		dataProfilerSettings.setRetrieveMetadata(true);
+		dataProfilerSettings.setRetrieveReplicas(false);
+		dataProfilerSettings.setRetrieveShared(false);
+		dataProfilerSettings.setRetrieveStarred(false);
+		dataProfilerSettings.setRetrieveTickets(false);
+
 	}
 
 }

irodsext-data-typer/pom.xml

@@ -22,6 +22,10 @@
 			<artifactId>jargon-core</artifactId>
 			<version>${project.version}</version>
 		</dependency>
+		<dependency>
+			<groupId>org.apache.tika</groupId>
+			<artifactId>tika-core</artifactId>
+		</dependency>
 	</dependencies>
 	<description>Tools for managing data type/format recognition</description>
 	<build>

irodsext-data-typer/src/main/java/org/irodsext/datatyper/IrodsextDataTypeResolutionService.java

@@ -0,0 +1,124 @@
+/**
+ * 
+ */
+package org.irodsext.datatyper;
+
+import java.io.IOException;
+
+import org.apache.tika.detect.Detector;
+import org.apache.tika.metadata.Metadata;
+import org.apache.tika.mime.MediaType;
+import org.apache.tika.parser.AutoDetectParser;
+import org.irods.jargon.core.connection.IRODSAccount;
+import org.irods.jargon.core.exception.DataNotFoundException;
+import org.irods.jargon.core.exception.JargonException;
+import org.irods.jargon.core.pub.IRODSAccessObjectFactory;
+import org.irods.jargon.core.utils.LocalFileUtils;
+import org.irods.jargon.core.utils.MiscIRODSUtils;
+import org.irods.jargon.extensions.datatyper.DataType;
+import org.irods.jargon.extensions.datatyper.DataTypeResolutionService;
+import org.irods.jargon.extensions.datatyper.DataTyperSettings;
+import org.irods.jargon.extensions.datatyper.IrodsMimeTypes;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+/**
+ * Data type resolution service to determine MIME and info types of a file. Note
+ * that this is at first a very basic service that will need to evolve over
+ * time.
+ * 
+ * @author Mike Conway - NIEHS
+ *
+ */
+public class IrodsextDataTypeResolutionService extends DataTypeResolutionService {
+
+	public static final Logger log = LoggerFactory.getLogger(IrodsextDataTypeResolutionService.class);
+
+	public IrodsextDataTypeResolutionService(IRODSAccessObjectFactory irodsAccessObjectFactory,
+			IRODSAccount irodsAccount, DataTyperSettings dataTyperSettings) {
+		super(irodsAccessObjectFactory, irodsAccount, dataTyperSettings);
+	}
+
+	@Override
+	public DataType resolveDataType(String irodsAbsolutePath) throws DataNotFoundException, JargonException {
+		log.info("resolveDataType()");
+
+		if (irodsAbsolutePath == null || irodsAbsolutePath.isEmpty()) {
+			throw new IllegalArgumentException("null or empty irodsAbsolutePath");
+		}
+
+		log.info("irodsAbsolutePath:{}", irodsAbsolutePath);
+
+		if (this.getDefaultDataTyperSettings().isDetailedDetermination()) {
+			log.warn("detailedDetermination not yet implemented, will default to check of file path");
+		}
+
+		log.info("checking for known irods types - interim code...");
+
+		String mimeType = determimeMimeTypeOfIrodsObjects(irodsAbsolutePath);
+
+		log.info("use Tika to derive based on file extenstion");
+
+		if (mimeType == null) {
+			log.info("not a known irods type, try tika");
+			mimeType = determineMimeTypeViaTika(irodsAbsolutePath);
+		}
+
+		if (mimeType == null) {
+			log.info("no mime type found via tika");
+			mimeType = "";
+		}
+
+		DataType dataType = new DataType();
+		dataType.setMimeType(mimeType);
+		log.info("dataType:{}", dataType);
+		return dataType;
+
+	}
+
+	private String determineMimeTypeViaTika(String irodsAbsolutePath) throws JargonException {
+		AutoDetectParser parser = new AutoDetectParser();
+		Detector detector = parser.getDetector();
+		Metadata md = new Metadata();
+		String fileName = MiscIRODSUtils.getLastPathComponentForGivenAbsolutePath(irodsAbsolutePath);
+
+		md.add(Metadata.RESOURCE_NAME_KEY, fileName);
+		MediaType mediaType;
+		try {
+			mediaType = detector.detect(null, md);
+		} catch (IOException e) {
+			throw new JargonException("io exception determining file type by extension", e);
+		}
+		return mediaType.toString();
+	}
+
+	@Override
+	public DataType resolveDataType(String irodsAbsolutePath, DataTyperSettings dataTyperSettings)
+			throws DataNotFoundException, JargonException {
+		// TODO Auto-generated method stub
+		return null;
+	}
+
+	/**
+	 * front-load detection of special irods file types
+	 * 
+	 * @param dataObject
+	 * @return
+	 */
+	private String determimeMimeTypeOfIrodsObjects(final String irodsAbsolutePath) {
+
+		String extension = LocalFileUtils.getFileExtension(irodsAbsolutePath);
+		if (extension == null || extension.isEmpty()) {
+			return null;
+		}
+
+		if (extension.equals(".r")) {
+			log.info("irods rule detected in:{}", irodsAbsolutePath);
+			return IrodsMimeTypes.APPLICATION_IRODS_RULE;
+		} else {
+			return null;
+		}
+
+	}
+
+}

irodsext-data-typer/src/main/java/org/irodsext/datatyper/package-info.java

@@ -0,0 +1,8 @@
+
+/**
+ * Default data typer service to determine MIME and info types of a file
+ * 
+ * @author Mike Conway - NIEHS
+ *
+ */
+package org.irodsext.datatyper;

irodsext-data-typer/src/test/java/org/irodsext/datatyper/IrodsextDataTypeResolutionServiceTest.java

@@ -0,0 +1,30 @@
+package org.irodsext.datatyper;
+
+import org.irods.jargon.core.connection.IRODSAccount;
+import org.irods.jargon.core.pub.IRODSAccessObjectFactory;
+import org.irods.jargon.extensions.datatyper.DataType;
+import org.irods.jargon.extensions.datatyper.DataTypeResolutionService;
+import org.irods.jargon.extensions.datatyper.DataTyperSettings;
+import org.junit.Assert;
+import org.junit.Test;
+import org.mockito.Mockito;
+
+public class IrodsextDataTypeResolutionServiceTest {
+
+	@Test
+	public void testResolveDataTypeString() throws Exception {
+		String testName = "/a/path/file.txt";
+		IRODSAccount dummyAccount = Mockito.mock(IRODSAccount.class);
+		IRODSAccessObjectFactory irodsAccessObjectFactory = Mockito.mock(IRODSAccessObjectFactory.class);
+		DataTyperSettings dataTyperSettings = new DataTyperSettings();
+		dataTyperSettings.setDetailedDetermination(false);
+		dataTyperSettings.setPersistDataTypes(false);
+		DataTypeResolutionService dtrs = new IrodsextDataTypeResolutionService(irodsAccessObjectFactory, dummyAccount,
+				dataTyperSettings);
+		DataType actual = dtrs.resolveDataType(testName);
+		Assert.assertNotNull("no type returned", actual);
+		Assert.assertEquals("text/plain", actual.getMimeType());
+
+	}
+
+}