/**
* Parse the parameters of a connection into a CoreNLP properties file that can be passed into
* {@link StanfordCoreNLP}, and used in the I/O stages.
*
* @param httpExchange The http exchange; effectively, the request information.
* @return A {@link Properties} object corresponding to a combination of default and passed
* properties.
* @throws UnsupportedEncodingException Thrown if we could not decode the key/value pairs with
* UTF-8.
*/
private Properties getProperties(HttpExchange httpExchange)
throws UnsupportedEncodingException {
// Load the default properties
Properties props = new Properties();
defaultProps
.entrySet()
.stream()
.forEach(
entry -> props.setProperty(entry.getKey().toString(), entry.getValue().toString()));
// Try to get more properties from query string.
Map<String, String> urlParams = getURLParams(httpExchange.getRequestURI());
if (urlParams.containsKey("properties")) {
StringUtils.decodeMap(URLDecoder.decode(urlParams.get("properties"), "UTF-8"))
.entrySet()
.forEach(entry -> props.setProperty(entry.getKey(), entry.getValue()));
} else if (urlParams.containsKey("props")) {
StringUtils.decodeMap(URLDecoder.decode(urlParams.get("properties"), "UTF-8"))
.entrySet()
.forEach(entry -> props.setProperty(entry.getKey(), entry.getValue()));
}
// Make sure the properties compile
props.setProperty(
"annotators",
StanfordCoreNLP.ensurePrerequisiteAnnotators(
props.getProperty("annotators").split("[, \t]+")));
return props;
}
public StanfordCoreNLPServer(int port) throws IOException {
serverPort = port;
defaultProps = new Properties();
defaultProps.setProperty(
"annotators", "tokenize, ssplit, pos, lemma, ner, parse, depparse, natlog, openie, dcoref");
defaultProps.setProperty("inputFormat", "text");
defaultProps.setProperty("outputFormat", "json");
// Generate and write a shutdown key
String tmpDir = System.getProperty("java.io.tmpdir");
File tmpFile = new File(tmpDir + File.separator + "corenlp.shutdown");
tmpFile.deleteOnExit();
if (tmpFile.exists()) {
if (!tmpFile.delete()) {
throw new IllegalStateException("Could not delete shutdown key file");
}
}
this.shutdownKey = new BigInteger(130, new Random()).toString(32);
IOUtils.writeStringToFile(shutdownKey, tmpFile.getPath(), "utf-8");
// Set the static page handler
this.staticPageHandle = new FileHandler("edu/stanford/nlp/pipeline/demo/corenlp-brat.html");
}
/**
* Initializes store.
*
* @throws GridException If failed to initialize.
*/
private void init() throws GridException {
if (initGuard.compareAndSet(false, true)) {
if (log.isDebugEnabled()) log.debug("Initializing cache store.");
try {
if (sesFactory != null)
// Session factory has been provided - nothing to do.
return;
if (!F.isEmpty(hibernateCfgPath)) {
try {
URL url = new URL(hibernateCfgPath);
sesFactory = new Configuration().configure(url).buildSessionFactory();
if (log.isDebugEnabled()) log.debug("Configured session factory using URL: " + url);
// Session factory has been successfully initialized.
return;
} catch (MalformedURLException e) {
if (log.isDebugEnabled())
log.debug("Caught malformed URL exception: " + e.getMessage());
}
// Provided path is not a valid URL. File?
File cfgFile = new File(hibernateCfgPath);
if (cfgFile.exists()) {
sesFactory = new Configuration().configure(cfgFile).buildSessionFactory();
if (log.isDebugEnabled())
log.debug("Configured session factory using file: " + hibernateCfgPath);
// Session factory has been successfully initialized.
return;
}
// Provided path is not a file. Classpath resource?
sesFactory = new Configuration().configure(hibernateCfgPath).buildSessionFactory();
if (log.isDebugEnabled())
log.debug("Configured session factory using classpath resource: " + hibernateCfgPath);
} else {
if (hibernateProps == null) {
U.warn(
log, "No Hibernate configuration has been provided for store (will use default).");
hibernateProps = new Properties();
hibernateProps.setProperty("hibernate.connection.url", DFLT_CONN_URL);
hibernateProps.setProperty("hibernate.show_sql", DFLT_SHOW_SQL);
hibernateProps.setProperty("hibernate.hbm2ddl.auto", DFLT_HBM2DDL_AUTO);
}
Configuration cfg = new Configuration();
cfg.setProperties(hibernateProps);
assert resourceAvailable(MAPPING_RESOURCE);
cfg.addResource(MAPPING_RESOURCE);
sesFactory = cfg.buildSessionFactory();
if (log.isDebugEnabled())
log.debug("Configured session factory using properties: " + hibernateProps);
}
} catch (HibernateException e) {
throw new GridException("Failed to initialize store.", e);
} finally {
initLatch.countDown();
}
} else if (initLatch.getCount() > 0) U.await(initLatch);
if (sesFactory == null) throw new GridException("Cache store was not properly initialized.");
}
@SuppressWarnings("unchecked")
public void execute() {
dataImporter.store(DataImporter.STATUS_MSGS, statusMessages);
document = dataImporter.getConfig().document;
final AtomicLong startTime = new AtomicLong(System.currentTimeMillis());
statusMessages.put(
TIME_ELAPSED,
new Object() {
public String toString() {
return getTimeElapsedSince(startTime.get());
}
});
statusMessages.put(DataImporter.MSG.TOTAL_QUERIES_EXECUTED, importStatistics.queryCount);
statusMessages.put(DataImporter.MSG.TOTAL_ROWS_EXECUTED, importStatistics.rowsCount);
statusMessages.put(DataImporter.MSG.TOTAL_DOC_PROCESSED, importStatistics.docCount);
statusMessages.put(DataImporter.MSG.TOTAL_DOCS_SKIPPED, importStatistics.skipDocCount);
List<String> entities = requestParameters.entities;
// Trigger onImportStart
if (document.onImportStart != null) {
invokeEventListener(document.onImportStart);
}
AtomicBoolean fullCleanDone = new AtomicBoolean(false);
// we must not do a delete of *:* multiple times if there are multiple root entities to be run
Properties lastIndexTimeProps = new Properties();
lastIndexTimeProps.setProperty(
LAST_INDEX_KEY,
DataImporter.DATE_TIME_FORMAT.get().format(dataImporter.getIndexStartTime()));
for (DataConfig.Entity e : document.entities) {
if (entities != null && !entities.contains(e.name)) continue;
lastIndexTimeProps.setProperty(
e.name + "." + LAST_INDEX_KEY, DataImporter.DATE_TIME_FORMAT.get().format(new Date()));
root = e;
String delQuery = e.allAttributes.get("preImportDeleteQuery");
if (dataImporter.getStatus() == DataImporter.Status.RUNNING_DELTA_DUMP) {
cleanByQuery(delQuery, fullCleanDone);
doDelta();
delQuery = e.allAttributes.get("postImportDeleteQuery");
if (delQuery != null) {
fullCleanDone.set(false);
cleanByQuery(delQuery, fullCleanDone);
}
} else {
cleanByQuery(delQuery, fullCleanDone);
doFullDump();
delQuery = e.allAttributes.get("postImportDeleteQuery");
if (delQuery != null) {
fullCleanDone.set(false);
cleanByQuery(delQuery, fullCleanDone);
}
}
statusMessages.remove(DataImporter.MSG.TOTAL_DOC_PROCESSED);
}
if (stop.get()) {
// Dont commit if aborted using command=abort
statusMessages.put("Aborted", DataImporter.DATE_TIME_FORMAT.get().format(new Date()));
rollback();
} else {
// Do not commit unnecessarily if this is a delta-import and no documents were created or
// deleted
if (!requestParameters.clean) {
if (importStatistics.docCount.get() > 0 || importStatistics.deletedDocCount.get() > 0) {
finish(lastIndexTimeProps);
}
} else {
// Finished operation normally, commit now
finish(lastIndexTimeProps);
}
if (writer != null) {
writer.finish();
}
if (document.onImportEnd != null) {
invokeEventListener(document.onImportEnd);
}
}
statusMessages.remove(TIME_ELAPSED);
statusMessages.put(DataImporter.MSG.TOTAL_DOC_PROCESSED, "" + importStatistics.docCount.get());
if (importStatistics.failedDocCount.get() > 0)
statusMessages.put(
DataImporter.MSG.TOTAL_FAILED_DOCS, "" + importStatistics.failedDocCount.get());
statusMessages.put("Time taken ", getTimeElapsedSince(startTime.get()));
LOG.info("Time taken = " + getTimeElapsedSince(startTime.get()));
}
