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);
}
/**
* 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;
}
@Override
public void handle(HttpExchange httpExchange) throws IOException {
// Set common response headers
httpExchange.getResponseHeaders().add("Access-Control-Allow-Origin", "*");
Future<String> json =
corenlpExecutor.submit(
() -> {
try {
// Get the document
Properties props =
new Properties() {
{
setProperty("annotators", "tokenize,ssplit,pos,lemma,ner,depparse");
}
};
Annotation doc = getDocument(props, httpExchange);
if (!doc.containsKey(CoreAnnotations.SentencesAnnotation.class)) {
StanfordCoreNLP pipeline = mkStanfordCoreNLP(props);
pipeline.annotate(doc);
}
// Construct the matcher
Map<String, String> params = getURLParams(httpExchange.getRequestURI());
// (get the pattern)
if (!params.containsKey("pattern")) {
respondError("Missing required parameter 'pattern'", httpExchange);
return "";
}
String pattern = params.get("pattern");
// (get whether to filter / find)
String filterStr = params.getOrDefault("filter", "false");
final boolean filter =
filterStr.trim().isEmpty()
|| "true".equalsIgnoreCase(filterStr.toLowerCase());
// (create the matcher)
final SemgrexPattern regex = SemgrexPattern.compile(pattern);
// Run TokensRegex
return JSONOutputter.JSONWriter.objectToJSON(
(docWriter) -> {
if (filter) {
// Case: just filter sentences
docWriter.set(
"sentences",
doc.get(CoreAnnotations.SentencesAnnotation.class)
.stream()
.map(
sentence ->
regex
.matcher(
sentence.get(
SemanticGraphCoreAnnotations
.CollapsedCCProcessedDependenciesAnnotation
.class))
.matches())
.collect(Collectors.toList()));
} else {
// Case: find matches
docWriter.set(
"sentences",
doc.get(CoreAnnotations.SentencesAnnotation.class)
.stream()
.map(
sentence ->
(Consumer<JSONOutputter.Writer>)
(JSONOutputter.Writer sentWriter) -> {
SemgrexMatcher matcher =
regex.matcher(
sentence.get(
SemanticGraphCoreAnnotations
.CollapsedCCProcessedDependenciesAnnotation
.class));
int i = 0;
while (matcher.find()) {
sentWriter.set(
Integer.toString(i),
(Consumer<JSONOutputter.Writer>)
(JSONOutputter.Writer matchWriter) -> {
IndexedWord match = matcher.getMatch();
matchWriter.set("text", match.word());
matchWriter.set(
"begin", match.index() - 1);
matchWriter.set("end", match.index());
for (String capture :
matcher.getNodeNames()) {
matchWriter.set(
"$" + capture,
(Consumer<JSONOutputter.Writer>)
groupWriter -> {
IndexedWord node =
matcher.getNode(
capture);
groupWriter.set(
"text", node.word());
groupWriter.set(
"begin",
node.index() - 1);
groupWriter.set(
"end", node.index());
});
}
});
i += 1;
}
sentWriter.set("length", i);
}));
}
});
} catch (Exception e) {
e.printStackTrace();
try {
respondError(e.getClass().getName() + ": " + e.getMessage(), httpExchange);
} catch (IOException ignored) {
}
}
return "";
});
// Send response
byte[] response = new byte[0];
try {
response = json.get(5, TimeUnit.SECONDS).getBytes();
} catch (InterruptedException | ExecutionException | TimeoutException e) {
respondError("Timeout when executing Semgrex query", httpExchange);
}
if (response.length > 0) {
httpExchange.getResponseHeaders().add("Content-Type", "text/json");
httpExchange.getResponseHeaders().add("Content-Length", Integer.toString(response.length));
httpExchange.sendResponseHeaders(HTTP_OK, response.length);
httpExchange.getResponseBody().write(response);
httpExchange.close();
}
}
@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);
}
}