Add FastInfosetExporter

examples/complete/template.yaml

@@ -7,6 +7,9 @@ Description: >
 
 # More info about Globals: https://github.com/awslabs/serverless-application-model/blob/master/docs/globals.rst
 Globals:
+  Api:
+    BinaryMediaTypes:
+      - application~1fastinfoset
   Function:
     Timeout: 120
 

src/main/java/co/zeroae/gate/App.java

@@ -1,10 +1,12 @@
 package co.zeroae.gate;
 
+import co.zeroae.gate.b64.Handler;
 import com.amazonaws.services.lambda.runtime.Context;
 import com.amazonaws.services.lambda.runtime.RequestHandler;
 import com.amazonaws.services.lambda.runtime.events.APIGatewayProxyRequestEvent;
 import com.amazonaws.services.lambda.runtime.events.APIGatewayProxyResponseEvent;
 
+import com.amazonaws.util.Base64;
 import com.amazonaws.xray.AWSXRay;
 import com.amazonaws.xray.entities.Subsegment;
 
@@ -70,11 +72,18 @@ public APIGatewayProxyResponseEvent handleRequest(APIGatewayProxyRequestEvent in
                 response.getHeaders().put("Content-Type", responseType);
             }
 
-            final String bodyType = input.getHeaders().get("Content-Type");
             final FeatureMap featureMap = Factory.newFeatureMap();
-            featureMap.put(Document.DOCUMENT_STRING_CONTENT_PARAMETER_NAME, input.getBody());
+            final String bodyType = input.getHeaders().get("Content-Type");
             if (bodyType != null)
                 featureMap.put(Document.DOCUMENT_MIME_TYPE_PARAMETER_NAME, bodyType);
+            if (input.getIsBase64Encoded() != null && input.getIsBase64Encoded())
+                featureMap.put(
+                        Document.DOCUMENT_URL_PARAMETER_NAME,
+                        new URL("b64", "localhost", 64, input.getBody(),
+                                new Handler()));
+            else
+                featureMap.put(Document.DOCUMENT_STRING_CONTENT_PARAMETER_NAME, input.getBody());
+
             final String inputDigest = AWSXRay.createSubsegment("Message Digest",
                     (subsegment) -> {
                         String rv = computeMessageDigest(featureMap);
@@ -108,19 +117,19 @@ public APIGatewayProxyResponseEvent handleRequest(APIGatewayProxyRequestEvent in
             AWSXRay.beginSubsegment("Gate Export");
             AWSXRay.getCurrentSubsegment().putMetadata("Content-Type", responseType);
             try {
-                return response.withBody(export(doc, exporter)).withStatusCode(200);
+                return export(exporter, doc, response).withStatusCode(200);
             } finally {
                 Factory.deleteResource(doc);
                 AWSXRay.endSubsegment();
             }
         } catch (GateException e) {
             logger.error(e);
-            AWSXRay.getCurrentSegment().addException(e);
+            AWSXRay.getCurrentSubsegmentOptional().ifPresent((segment -> segment.addException(e)));
             response.getHeaders().put("Content-Type", "text/plain");
             return response.withBody(e.getMessage()).withStatusCode(400);
         } catch (IOException e) {
             logger.error(e);
-            AWSXRay.getCurrentSegment().addException(e);
+            AWSXRay.getCurrentSubsegmentOptional().ifPresent((segment -> segment.addException(e)));
             response.getHeaders().put("Content-Type", "text/plain");
             return response.withBody(e.getMessage()).withStatusCode(406);
         }
@@ -165,26 +174,34 @@ private Document cacheComputeIfNull(String key, Utils.TextProcessor processor) t
     }
 
     private String computeMessageDigest(FeatureMap featureMap) {
-        final String sha256;
         try {
             final MessageDigest md = MessageDigest.getInstance("SHA-256");
-            final String bodyContent = (String)featureMap.get(Document.DOCUMENT_STRING_CONTENT_PARAMETER_NAME);
             final String mimeType = (String)featureMap.get(Document.DOCUMENT_MIME_TYPE_PARAMETER_NAME);
-            md.update(bodyContent.getBytes());
+            final String bodyContent = (String)featureMap.get(Document.DOCUMENT_STRING_CONTENT_PARAMETER_NAME);
+            final URL sourceUrl = (URL)featureMap.get(Document.DOCUMENT_URL_PARAMETER_NAME);
             if (mimeType != null)
                 md.update(mimeType.getBytes());
-            sha256 = Hex.encode(md.digest());
+            if (bodyContent != null)
+                md.update(bodyContent.getBytes());
+            if (sourceUrl != null)
+                md.update(sourceUrl.toString().getBytes());
+            return Hex.encode(md.digest());
         } catch (NoSuchAlgorithmException e) {
             throw new RuntimeException(e);
         }
-        return sha256;
     }
 
     /**
-     * @param doc an instance of gate.Document
      * @param exporter The document exporter
+     * @param doc an instance of gate.Document
+     * @param response The response where we put the exported Document as body
+     * @return the modified response
      */
-    private String export(Document doc, DocumentExporter exporter) throws IOException {
+    private APIGatewayProxyResponseEvent export(
+            DocumentExporter exporter,
+            Document doc,
+            APIGatewayProxyResponseEvent response
+    ) throws IOException {
         final FeatureMap exportOptions = Factory.newFeatureMap();
 
         // Take *all* annotation types.
@@ -202,7 +219,13 @@ private String export(Document doc, DocumentExporter exporter) throws IOExceptio
             AWSXRay.getCurrentSubsegment().addException(e);
             throw e;
         }
-        return baos.toString();
+        // If we add a second type, then we should create a "Set" at the Utils level and test against it.
+        if (exporter.getMimeType().equals("application/fastinfoset")) {
+            response.withIsBase64Encoded(true).setBody(Base64.encodeAsString(baos.toByteArray()));
+        } else {
+            response.setBody(baos.toString());
+        }
+        return response;
     }
 
     private static CorpusController loadApplication() {

src/main/java/co/zeroae/gate/Utils.java

@@ -81,9 +81,13 @@ static void loadDocumentFormats() {
     static Map<String, DocumentExporter> loadExporters() {
         final GateXMLExporter gateXMLExporter = new GateXMLExporter();
         final GATEJsonExporter gateJsonExporter = new GATEJsonExporter();
+        final FastInfosetExporter fastInfosetExporter = new FastInfosetExporter();
+
         final Map<String, DocumentExporter> rv = new HashMap<>();
         rv.put("application/gate+xml", gateXMLExporter);
         rv.put("application/gate+json", gateJsonExporter);
+        rv.put("application/json", gateJsonExporter);
+        rv.put("application/fastinfoset", fastInfosetExporter);
         return Collections.unmodifiableMap(rv);
     }
 }

src/main/java/co/zeroae/gate/b64/Handler.java

@@ -0,0 +1,35 @@
+package co.zeroae.gate.b64;
+
+import com.amazonaws.util.StringInputStream;
+import org.apache.commons.codec.binary.Base64InputStream;
+
+import java.io.IOException;
+import java.io.InputStream;
+import java.net.URL;
+import java.net.URLConnection;
+import java.net.URLStreamHandler;
+
+public class Handler extends URLStreamHandler {
+    private class Connection extends URLConnection {
+        /**
+         * Constructs a URL connection to the specified URL. A connection to
+         * the object referenced by the URL is not created.
+         *
+         * @param url the specified URL.
+         */
+        protected Connection(URL url) {
+            super(url);
+        }
+        @Override
+        public void connect() {
+        }
+        @Override
+        public InputStream getInputStream() throws IOException {
+            return new Base64InputStream(new StringInputStream(url.getPath()));
+        }
+    }
+    @Override
+    protected URLConnection openConnection(URL u) {
+        return new Connection(u);
+    }
+}

src/test/java/co/zeroae/gate/AppTest.java

@@ -2,6 +2,7 @@
 
 import com.amazonaws.services.lambda.runtime.events.APIGatewayProxyRequestEvent;
 import com.amazonaws.services.lambda.runtime.events.APIGatewayProxyResponseEvent;
+import com.amazonaws.util.Base64;
 import com.amazonaws.util.IOUtils;
 import com.amazonaws.xray.AWSXRay;
 import com.fasterxml.jackson.core.JsonFactory;
@@ -81,18 +82,32 @@ public void testMissingContentType() {
 
     @Test
     public void testGateJSONResponse() throws Exception {
-        input_headers.put("Accept", "application/gate+json");
+        for (String responseType: new String[]{"application/gate+json", "application/json"}) {
+            input_headers.put("Accept", responseType);
+
+            final APIGatewayProxyResponseEvent result = app.handleRequest(input, context);
+            assertEquals(200, result.getStatusCode().intValue());
+
+            // Ensure we get back application/gate+json back
+            assertEquals(responseType, result.getHeaders().get("Content-Type"));
+            final JsonFactory factory = new JsonFactory();
+            final JsonParser parser = factory.createParser(result.getBody());
+            while (!parser.isClosed()) {
+                parser.nextToken();
+            }
+        }
+    }
+
+    @Test
+    public void testFastInfosetResponse() {
+        input_headers.put("Accept", "application/fastinfoset");
 
         final APIGatewayProxyResponseEvent result = app.handleRequest(input, context);
         assertEquals(200, result.getStatusCode().intValue());
 
-        // Ensure we get back application/gate+json back
-        assertEquals("application/gate+json", result.getHeaders().get("Content-Type"));
-        final JsonFactory factory = new JsonFactory();
-        final JsonParser parser = factory.createParser(result.getBody());
-        while (!parser.isClosed()) {
-            parser.nextToken();
-        }
+        // Ensure we get back application/fastinfoset back
+        assertEquals("application/fastinfoset", result.getHeaders().get("Content-Type"));
+        assertTrue(result.getIsBase64Encoded());
     }
 
     @Test
@@ -156,6 +171,22 @@ public void testMediaWikiInput() throws IOException {
         assertFalse(result.getBody().contains("{{Short Description|"));
     }
 
+    @Test
+    public void testFastInfosetInput() throws IOException {
+        final byte[] body = IOUtils.toByteArray(getClass().getResourceAsStream("example.finf"));
+        input.withIsBase64Encoded(true)
+                .withBody(Base64.encodeAsString(body))
+                .getHeaders().put("Content-Type", "application/fastinfoset");
+
+        final APIGatewayProxyResponseEvent result = app.handleRequest(input, context);
+        assertEquals(200, result.getStatusCode().intValue());
+        assertEquals("MISS", result.getHeaders().get("x-zae-gate-cache"));
+
+        final APIGatewayProxyResponseEvent cachedResult = app.handleRequest(input, context);
+        assertEquals(200, cachedResult.getStatusCode().intValue());
+        assertEquals("HIT", cachedResult.getHeaders().get("x-zae-gate-cache"));
+    }
+
     @Test
     public void testCacheCoherenceWithContentType() {
         final int cacheBust= new Random().nextInt();