Extractor Modules / Taverna Extractor / Diff of /src/main/java/net/timbusproject/extractors/modules/tavernaextractor/provenance/ProvenanceAccessClient.java

Diff of /src/main/java/net/timbusproject/extractors/modules/tavernaextractor/provenance/ProvenanceAccessClient.java [346c31] .. [c78deb]

Switch to unified view


...
import java.io.UnsupportedEncodingException;
import java.net.MalformedURLException;
import java.net.URL;
import java.nio.file.Files;
import java.nio.file.Path;
import java.nio.file.Paths;
import java.sql.SQLException;
import java.util.ArrayList;
import java.util.Collections;
import java.util.HashMap;
import java.util.List;
...
                    break; // no further treatment!
                }

                Identified identified = pc.getReferenceService().resolveIdentifier(ref, null, ic);
                if (identified instanceof ReferenceSet) {

                    LOGGER.info("try to fetch reference set for binding "+binding.getPortName());

                    ReferenceSet referenceSet = (ReferenceSet) identified;
                    Set<ExternalReferenceSPI> externalReferences = referenceSet.getExternalReferences();

                    for (ExternalReferenceSPI externalReference : externalReferences) {
                        if (externalReference instanceof ValueCarryingExternalReference<?>) {
                            ValueCarryingExternalReference<?> vcer = (ValueCarryingExternalReference<?>) externalReference;


                            // string ? filepath or URL
                            if (String.class.isAssignableFrom(vcer.getValueType())) {

                                String possibleFilePath = (String) vcer.getValue();

                                // valid file ?
                                File file = isFile(possibleFilePath);
                                if (file != null && file.length() > 0) {

                                    LOGGER.debug("\t Found file (" + file.getPath() + ") for port= "
                                            + binding.getPortName());

                                    fileName += "[" + binding.getProcessorName() + "][" + binding.getPortName() + "]-";
                                    File tmpFile = File.createTempFile(fileName, ".file", tempDir.toFile());

                                    try {
                                        FileUtils.copyFile(file, tmpFile);
                                    } catch (IOException e) {
                                        LOGGER.error("Error during fetching file. " + e.getMessage());
                                    }
                                    fileName = "";
                                }

                                // valid URL ?
                                URL url = isURL(possibleFilePath);
...

                                    LOGGER.debug("\t Found valid URL (" + possibleFilePath + ") for port = "
                                            + binding.getPortName());

                                    // download file and copy it to temp directory
                                    fileName += "[" + binding.getProcessorName() + "][" + binding.getPortName() + "]-";
                                    File content = File.createTempFile(fileName, ".url", tempDir.toFile());
                                    FileUtils.copyURLToFile(url, content);

                                    LOGGER.debug("\t Created temporary file (" + content.getAbsolutePath() + ")");
                                    fileName = "";
...

                                LOGGER.debug("\t Found byte[] for port= " + binding.getPortName());

                                fileName += "[" + binding.getPortName() + "]-";
                                byte[] content = (byte[]) vcer.getValue();
                                File tmpFile = File.createTempFile(fileName, ".binary", tempDir.toFile());
                                FileUtils.writeByteArrayToFile(tmpFile, content);

                                LOGGER.debug("\t Create " + fileName + " file for port= " + binding.getPortName());

                                // reset file name
...
     * @param possibleFilePath
     * @return file if exists
     */
    private File isFile(String possibleFilePath) {

        LOGGER.debug("call isFile(" + possibleFilePath + ")");

        LOGGER.debug("\t Detect if file path (" + possibleFilePath + ") point to a file. ");
        try {
            return Paths.get(possibleFilePath).toFile();
        } catch (Exception e) {
            LOGGER.debug("\t File can not be detected. " + e.getMessage());
            return null;
        }

//        String regexPath = "([a-zA-Z]:)?(\\\\[a-zA-Z0-9_.-]+)+\\\\?";
//
//        if (possibleFilePath != null && possibleFilePath.length() > 5 && possibleFilePath.contains(File.separator)
//                && Pattern.matches(regexPath, possibleFilePath)) {
//
//            LOGGER.debug("\t Detect if file path (" + possibleFilePath + ") point to a file. ");
//            try {
//                // can we create a file of the possibleFilePath
//                File file = new File(possibleFilePath);
//                if (!file.isDirectory()) {
//                    return file;
//                }
//
//            } catch (Exception e) {
//                LOGGER.debug("\t File can not be detected. " + e.getMessage());
//                return null;
//            }
//        }

    }

    private URL isURL(String possibleURL) {

        LOGGER.debug("call isURL(" + possibleURL + ")");

        String regexURL = "\\b(https?|ftp|file)://[-a-zA-Z0-9+&@#/%?=~_|!:,.;]*[-a-zA-Z0-9+&@#/%=~_|]";

        if (possibleURL != null && possibleURL.length() > 5 && Pattern.matches(regexURL, possibleURL)) {

            LOGGER.debug("\t Detect if (" + possibleURL + ") is a valid URL.");

            try {
                return new URL(possibleURL);
            } catch (MalformedURLException e) {
                LOGGER.debug("\t Not a valid URL. "+e.getMessage());
            }
        }
        return null;
    }
}

	a/src/main/java/net/timbusproject/extractors/modules/tavernaextractor/provenance/ProvenanceAccessClient.java		b/src/main/java/net/timbusproject/extractors/modules/tavernaextractor/provenance/ProvenanceAccessClient.java
	...		...
9	import java.io.UnsupportedEncodingException;	9	import java.io.UnsupportedEncodingException;
10	import java.net.MalformedURLException;	10	import java.net.MalformedURLException;
11	import java.net.URL;	11	import java.net.URL;
12	import java.nio.file.Files;	12	import java.nio.file.Files;
13	import java.nio.file.Path;	13	import java.nio.file.Path;
		14	import java.nio.file.Paths;
14	import java.sql.SQLException;	15	import java.sql.SQLException;
15	import java.util.ArrayList;	16	import java.util.ArrayList;
16	import java.util.Collections;	17	import java.util.Collections;
17	import java.util.HashMap;	18	import java.util.HashMap;
18	import java.util.List;	19	import java.util.List;
	...		...
188	break; // no further treatment!	189	break; // no further treatment!
189	}	190	}
190		191
191	Identified identified = pc.getReferenceService().resolveIdentifier(ref, null, ic);	192	Identified identified = pc.getReferenceService().resolveIdentifier(ref, null, ic);
192	if (identified instanceof ReferenceSet) {	193	if (identified instanceof ReferenceSet) {
		194
		195	LOGGER.info("try to fetch reference set for binding "+binding.getPortName());
		196
193	ReferenceSet referenceSet = (ReferenceSet) identified;	197	ReferenceSet referenceSet = (ReferenceSet) identified;
194	Set<ExternalReferenceSPI> externalReferences = referenceSet.getExternalReferences();	198	Set<ExternalReferenceSPI> externalReferences = referenceSet.getExternalReferences();
195		199
196	for (ExternalReferenceSPI externalReference : externalReferences) {	200	for (ExternalReferenceSPI externalReference : externalReferences) {
197	if (externalReference instanceof ValueCarryingExternalReference<?>) {	201	if (externalReference instanceof ValueCarryingExternalReference<?>) {
198	ValueCarryingExternalReference<?> vcer = (ValueCarryingExternalReference<?>) externalReference;	202	ValueCarryingExternalReference<?> vcer = (ValueCarryingExternalReference<?>) externalReference;
199		203
		204
200	// string ? filepath or URL	205	// string ? filepath or URL
201	if (String.class.isAssignableFrom(vcer.getValueType())) {	206	if (String.class.isAssignableFrom(vcer.getValueType())) {
202		207
203	String possibleFilePath = (String) vcer.getValue();	208	String possibleFilePath = (String) vcer.getValue();
204		209
205	// valid file ?	210	// valid file ?
206	File file = isFile(possibleFilePath);	211	File file = isFile(possibleFilePath);
207	if (file != null) {	212	if (file != null && file.length() > 0) {
208		213
209	LOGGER.debug("\t Found file (" + file.getPath() + ") for port= "	214	LOGGER.debug("\t Found file (" + file.getPath() + ") for port= "
210	+ binding.getPortName());	215	+ binding.getPortName());
211		216
212	fileName += "[" + binding.getPortName() + "]-";	217	fileName += "[" + binding.getProcessorName() + "][" + binding.getPortName() + "]-";
213	File tmpFile = File.createTempFile(fileName.toString(), ".file", tempDir.toFile());	218	File tmpFile = File.createTempFile(fileName, ".file", tempDir.toFile());
214		219
		220	try {
215	FileUtils.copyFile(file, tmpFile);	221	FileUtils.copyFile(file, tmpFile);
		222	} catch (IOException e) {
		223	LOGGER.error("Error during fetching file. " + e.getMessage());
		224	}
216	fileName = "";	225	fileName = "";
217	}	226	}
218		227
219	// valid URL ?	228	// valid URL ?
220	URL url = isURL(possibleFilePath);	229	URL url = isURL(possibleFilePath);
	...		...
222		231
223	LOGGER.debug("\t Found valid URL (" + possibleFilePath + ") for port = "	232	LOGGER.debug("\t Found valid URL (" + possibleFilePath + ") for port = "
224	+ binding.getPortName());	233	+ binding.getPortName());
225		234
226	// download file and copy it to temp directory	235	// download file and copy it to temp directory
227	fileName += "[" + binding.getPortName() + "]-";	236	fileName += "[" + binding.getProcessorName() + "][" + binding.getPortName() + "]-";
228	File content = File.createTempFile(fileName, ".url", tempDir.toFile());	237	File content = File.createTempFile(fileName, ".url", tempDir.toFile());
229	FileUtils.copyURLToFile(url, content);	238	FileUtils.copyURLToFile(url, content);
230		239
231	LOGGER.debug("\t Created temporary file (" + content.getAbsolutePath() + ")");	240	LOGGER.debug("\t Created temporary file (" + content.getAbsolutePath() + ")");
232	fileName = "";	241	fileName = "";
	...		...
238		247
239	LOGGER.debug("\t Found byte[] for port= " + binding.getPortName());	248	LOGGER.debug("\t Found byte[] for port= " + binding.getPortName());
240		249
241	fileName += "[" + binding.getPortName() + "]-";	250	fileName += "[" + binding.getPortName() + "]-";
242	byte[] content = (byte[]) vcer.getValue();	251	byte[] content = (byte[]) vcer.getValue();
243	File tmpFile = File.createTempFile(fileName.toString(), ".binary", tempDir.toFile());	252	File tmpFile = File.createTempFile(fileName, ".binary", tempDir.toFile());
244	FileUtils.writeByteArrayToFile(tmpFile, content);	253	FileUtils.writeByteArrayToFile(tmpFile, content);
245		254
246	LOGGER.debug("\t Create " + fileName + " file for port= " + binding.getPortName());	255	LOGGER.debug("\t Create " + fileName + " file for port= " + binding.getPortName());
247		256
248	// reset file name	257	// reset file name
	...		...
403	* @param possibleFilePath	412	* @param possibleFilePath
404	* @return file if exists	413	* @return file if exists
405	*/	414	*/
406	private File isFile(String possibleFilePath) {	415	private File isFile(String possibleFilePath) {
407		416
		417	LOGGER.debug("call isFile(" + possibleFilePath + ")");
		418
		419	LOGGER.debug("\t Detect if file path (" + possibleFilePath + ") point to a file. ");
		420	try {
		421	return Paths.get(possibleFilePath).toFile();
		422	} catch (Exception e) {
		423	LOGGER.debug("\t File can not be detected. " + e.getMessage());
		424	return null;
		425	}
		426
408	String regexPath = "([a-zA-Z]:)?(\\\\[a-zA-Z0-9_.-]+)+\\\\?";	427	// String regexPath = "([a-zA-Z]:)?(\\\\[a-zA-Z0-9_.-]+)+\\\\?";
409		428	//
410	if (possibleFilePath != null && possibleFilePath.length() > 5 && possibleFilePath.contains(File.separator)	429	// if (possibleFilePath != null && possibleFilePath.length() > 5 && possibleFilePath.contains(File.separator)
411	&& Pattern.matches(regexPath, possibleFilePath)) {	430	// && Pattern.matches(regexPath, possibleFilePath)) {
412		431	//
413	LOGGER.debug("\t Detect if file path (" + possibleFilePath + ") point to a file. ");	432	// LOGGER.debug("\t Detect if file path (" + possibleFilePath + ") point to a file. ");
414	try {	433	// try {
415	// can we create a file of the possibleFilePath	434	// // can we create a file of the possibleFilePath
416	File file = new File(possibleFilePath);	435	// File file = new File(possibleFilePath);
417	if (!file.isDirectory()) {	436	// if (!file.isDirectory()) {
418	return file;	437	// return file;
419	}	438	// }
420		439	//
421	} catch (Exception e) {	440	// } catch (Exception e) {
422	LOGGER.debug("\t File can not be detected. " + e.getMessage());	441	// LOGGER.debug("\t File can not be detected. " + e.getMessage());
423	return null;	442	// return null;
424	}	443	// }
425	}	444	// }
426	return null;
427	}	445	}
428		446
429	private URL isURL(String possibleURL) {	447	private URL isURL(String possibleURL) {
		448
		449	LOGGER.debug("call isURL(" + possibleURL + ")");
430		450
431	String regexURL = "\\b(https?\|ftp\|file)://[-a-zA-Z0-9+&@#/%?=~_\|!:,.;]*[-a-zA-Z0-9+&@#/%=~_\|]";	451	String regexURL = "\\b(https?\|ftp\|file)://[-a-zA-Z0-9+&@#/%?=~_\|!:,.;]*[-a-zA-Z0-9+&@#/%=~_\|]";
432		452
433	if (possibleURL != null && possibleURL.length() > 5 && Pattern.matches(regexURL, possibleURL)) {	453	if (possibleURL != null && possibleURL.length() > 5 && Pattern.matches(regexURL, possibleURL)) {
434		454
435	LOGGER.debug("\t Detect if (" + possibleURL + ") is a valid URL.");	455	LOGGER.debug("\t Detect if (" + possibleURL + ") is a valid URL.");
436		456
437	try {	457	try {
438	return new URL(possibleURL);	458	return new URL(possibleURL);
439	} catch (MalformedURLException e) {	459	} catch (MalformedURLException e) {
440	LOGGER.debug("\t Not a valid URL.");	460	LOGGER.debug("\t Not a valid URL. "+e.getMessage());
441	}	461	}
442	}	462	}
443	return null;	463	return null;
444	}	464	}
445	}	465	}