public static void main(String[] args) throws IOException {
PrintWriter out;
if (args.length > 1) {
out = new PrintWriter(args[1]);
} else {
out = new PrintWriter(System.out);
}
PrintWriter xmlOut = null;
if (args.length > 2) {
xmlOut = new PrintWriter(args[2]);
}
Properties props = new Properties();
props.put("annotators", "tokenize, ssplit, pos, lemma, ner,parse");
StanfordCoreNLP pipeline = new StanfordCoreNLP(props);
Annotation annotation;
if (args.length > 0) {
annotation = new Annotation(IOUtils.slurpFileNoExceptions(args[0]));
} else {
annotation =
new Annotation(
"Kosgi Santosh sent an email to Stanford University. He didn't get a reply.");
}
pipeline.annotate(annotation);
pipeline.prettyPrint(annotation, out);
}
private static Properties getProperties(
String name, String mapping, boolean ignoreCase, String validPosRegex) {
String prefix = (name != null && !name.isEmpty()) ? name + '.' : "";
Properties props = new Properties();
props.setProperty(prefix + "mapping", mapping);
props.setProperty(prefix + "ignorecase", String.valueOf(ignoreCase));
if (validPosRegex != null) {
props.setProperty(prefix + "validpospattern", validPosRegex);
}
return props;
}
/**
* Reads the POST contents of the request and parses it into an Annotation object, ready to be
* annotated. This method can also read a serialized document, if the input format is set to be
* serialized.
*
* @param props The properties we are annotating with. This is where the input format is retrieved
* from.
* @param httpExchange The exchange we are reading POST data from.
* @return An Annotation representing the read document.
* @throws IOException Thrown if we cannot read the POST data.
* @throws ClassNotFoundException Thrown if we cannot load the serializer.
*/
private Annotation getDocument(Properties props, HttpExchange httpExchange)
throws IOException, ClassNotFoundException {
String inputFormat = props.getProperty("inputFormat", "text");
switch (inputFormat) {
case "text":
return new Annotation(
IOUtils.slurpReader(new InputStreamReader(httpExchange.getRequestBody())));
case "serialized":
String inputSerializerName =
props.getProperty("inputSerializer", ProtobufAnnotationSerializer.class.getName());
AnnotationSerializer serializer = MetaClass.create(inputSerializerName).createInstance();
Pair<Annotation, InputStream> pair = serializer.read(httpExchange.getRequestBody());
return pair.first;
default:
throw new IOException("Could not parse input format: " + inputFormat);
}
}
/**
* 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 TokensRegexNERAnnotator(String name, Properties properties) {
String prefix = (name != null && !name.isEmpty()) ? name + '.' : "";
String backgroundSymbol =
properties.getProperty(prefix + "backgroundSymbol", DEFAULT_BACKGROUND_SYMBOL);
String[] backgroundSymbols = backgroundSymbol.split("\\s*,\\s*");
String mappingFiles =
properties.getProperty(prefix + "mapping", DefaultPaths.DEFAULT_REGEXNER_RULES);
String[] mappings = mappingFiles.split("\\s*[,;]\\s*");
String validPosRegex = properties.getProperty(prefix + "validpospattern");
this.posMatchType =
PosMatchType.valueOf(
properties.getProperty(prefix + "posmatchtype", DEFAULT_POS_MATCH_TYPE.name()));
String noDefaultOverwriteLabelsProp =
properties.getProperty(prefix + "noDefaultOverwriteLabels");
this.noDefaultOverwriteLabels =
(noDefaultOverwriteLabelsProp != null)
? Collections.unmodifiableSet(
CollectionUtils.asSet(noDefaultOverwriteLabelsProp.split("\\s*,\\s*")))
: Collections.unmodifiableSet(new HashSet<>());
this.ignoreCase = PropertiesUtils.getBool(properties, prefix + "ignorecase", false);
this.verbose = PropertiesUtils.getBool(properties, prefix + "verbose", false);
if (validPosRegex != null && !validPosRegex.isEmpty()) {
validPosPattern = Pattern.compile(validPosRegex);
} else {
validPosPattern = null;
}
entries =
Collections.unmodifiableList(
readEntries(name, noDefaultOverwriteLabels, ignoreCase, verbose, mappings));
IdentityHashMap<SequencePattern<CoreMap>, Entry> patternToEntry = new IdentityHashMap<>();
multiPatternMatcher = createPatternMatcher(patternToEntry);
this.patternToEntry = Collections.unmodifiableMap(patternToEntry);
Set<String> myLabels = Generics.newHashSet();
// Can always override background or none.
Collections.addAll(myLabels, backgroundSymbols);
myLabels.add(null);
// Always overwrite labels
for (Entry entry : entries) myLabels.add(entry.type);
this.myLabels = Collections.unmodifiableSet(myLabels);
}
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");
}
/**
* Get the response data type to send to the client, based off of the output format requested
* from CoreNLP.
*
* @param props The properties being used by CoreNLP.
* @param of The output format being output by CoreNLP.
* @return An identifier for the type of the HTTP response (e.g., 'text/json').
*/
public String getContentType(Properties props, StanfordCoreNLP.OutputFormat of) {
switch (of) {
case JSON:
return "text/json";
case TEXT:
case CONLL:
return "text/plain";
case XML:
return "text/xml";
case SERIALIZED:
String outputSerializerName = props.getProperty("outputSerializer");
if (outputSerializerName != null
&& outputSerializerName.equals(ProtobufAnnotationSerializer.class.getName())) {
return "application/x-protobuf";
}
default:
return "application/octet-stream";
}
}
@Override
public void handle(HttpExchange httpExchange) throws IOException {
// Set common response headers
httpExchange.getResponseHeaders().add("Access-Control-Allow-Origin", "*");
// Get sentence.
Properties props;
Annotation ann;
StanfordCoreNLP.OutputFormat of;
log("[" + httpExchange.getRemoteAddress() + "] Received message");
try {
props = getProperties(httpExchange);
ann = getDocument(props, httpExchange);
of =
StanfordCoreNLP.OutputFormat.valueOf(
props.getProperty("outputFormat", "json").toUpperCase());
// Handle direct browser connections (i.e., not a POST request).
if (ann.get(CoreAnnotations.TextAnnotation.class).length() == 0) {
log("[" + httpExchange.getRemoteAddress() + "] Interactive connection");
staticPageHandle.handle(httpExchange);
return;
}
log("[" + httpExchange.getRemoteAddress() + "] API call");
} catch (Exception e) {
// Return error message.
e.printStackTrace();
String response = e.getMessage();
httpExchange.getResponseHeaders().add("Content-Type", "text/plain");
httpExchange.sendResponseHeaders(HTTP_BAD_INPUT, response.length());
httpExchange.getResponseBody().write(response.getBytes());
httpExchange.close();
return;
}
try {
// Annotate
StanfordCoreNLP pipeline = mkStanfordCoreNLP(props);
Future<Annotation> completedAnnotationFuture =
corenlpExecutor.submit(
() -> {
pipeline.annotate(ann);
return ann;
});
Annotation completedAnnotation = completedAnnotationFuture.get(5, TimeUnit.SECONDS);
// Get output
ByteArrayOutputStream os = new ByteArrayOutputStream();
StanfordCoreNLP.createOutputter(props, AnnotationOutputter.getOptions(pipeline))
.accept(completedAnnotation, os);
os.close();
byte[] response = os.toByteArray();
httpExchange.getResponseHeaders().add("Content-Type", getContentType(props, of));
httpExchange.getResponseHeaders().add("Content-Length", Integer.toString(response.length));
httpExchange.sendResponseHeaders(HTTP_OK, response.length);
httpExchange.getResponseBody().write(response);
httpExchange.close();
} catch (TimeoutException e) {
respondError("CoreNLP request timed out", httpExchange);
} catch (Exception e) {
// Return error message.
respondError(e.getClass().getName() + ": " + e.getMessage(), httpExchange);
}
}