Set up PUT route for indexing single records (RPB-223)

app/controllers/nwbib/Application.java

@@ -3,11 +3,14 @@
 package controllers.nwbib;
 
 import java.io.File;
+import java.io.FileNotFoundException;
 import java.io.IOException;
 import java.io.InputStream;
 import java.io.UnsupportedEncodingException;
 import java.net.URL;
 import java.net.URLEncoder;
+import java.nio.file.Files;
+import java.nio.file.Paths;
 import java.text.Collator;
 import java.util.ArrayList;
 import java.util.Arrays;
@@ -28,6 +31,7 @@
 import java.util.stream.Stream;
 import java.util.stream.StreamSupport;
 
+import org.antlr.runtime.RecognitionException;
 import org.apache.commons.lang3.tuple.Pair;
 import org.elasticsearch.common.base.Charsets;
 import org.elasticsearch.common.geo.GeoPoint;
@@ -54,6 +58,7 @@
 import play.mvc.Result;
 import play.mvc.Results;
 import play.twirl.api.HtmlFormat;
+import rpb.ETL;
 import views.html.browse_classification;
 import views.html.browse_register;
 import views.html.classification;
@@ -943,4 +948,21 @@ private static List<String> starredIds() {
 		return new ArrayList<>(Arrays.asList(currentlyStarred().split(" ")).stream()
 				.filter(s -> !s.trim().isEmpty()).collect(Collectors.toList()));
 	}
+
+	public static Promise<Result> put(String id, String secret) throws FileNotFoundException, RecognitionException, IOException {
+		File input = new File("conf/output/test-output-strapi.json");
+		File output = new File("conf/output/test-output-0.json");
+		Files.write(Paths.get(input.getAbsolutePath()), request().body().asJson().toString().getBytes(Charsets.UTF_8));
+		ETL.main(new String[] {"conf/rpb-test-titel-to-lobid.flux"});
+		String result = Files.readAllLines(Paths.get(output.getAbsolutePath())).stream().collect(Collectors.joining("\n"));
+		boolean authorized = !secret.trim().isEmpty() && secret.equals(CONFIG.getString("secret"));
+		if (authorized) {
+			String url = "http://weywot3:9200/resources-rpb-test/resource/"
+					+ URLEncoder.encode("https://lobid.org/resources/" + id, "UTF-8");
+			WSRequest request = WS.url(url).setHeader("Content-Type", "application/json");
+			return request.put(result).map(response -> status(response.getStatus(), response.getBody()));
+		} else {
+			return Promise.pure(unauthorized());
+		}
+	}
 }

app/controllers/nwbib/Lobid.java

@@ -324,7 +324,7 @@ public static String resourceLabel(String id) {
 		Callable<String> getLabel = () -> {
 			// e.g. take TT000086525 from http://lobid.org/resources/TT000086525#!
 			String simpleId =
-					id.replaceAll("https?://[^/]+/resources?/(?:search\\?q=rpbId:)?(.+?)[^A-Z0-9]*$", "$1");
+					id.replaceAll("https?://[^/]+/(?:resources?/)?(?:search\\?q=rpbId:)?(.+?)[^A-Z0-9]*$", "$1");
 			JsonNode json = getResource(simpleId).findValue("title");
 			String label =
 					json == null ? "" : HtmlEscapers.htmlEscaper().escape(json.asText());

app/views/TableRow.java

@@ -38,6 +38,7 @@ public String process(JsonNode doc, String property, String param,
 			return filtered.isEmpty() ? ""
 					: String.format("<tr><td>%s</td><td>%s</td></tr>", label,
 							filtered.stream()
+									.flatMap(s -> Arrays.asList(s.split("; ")).stream())
 									.map(val -> label(doc, property, param, val, keys))
 									.collect(Collectors.joining(
 											property.equals("subjectChain") ? " <br/> " : " | ")));
@@ -48,7 +49,9 @@ private String label(JsonNode doc, String property, String param,
 			String value = property.equals("subjectChain")
 					? val.replaceAll("\\([\\d,]+\\)$", "").trim() : val;
 			if (!labels.isPresent()) {
-				return refAndLabel(property, value, labels)[0];
+				String[] refAndLabel = refAndLabel(property, value, labels);
+				return value.startsWith("http") ? String.format("<a title='%s' href='%s'>%s</a>",
+						refAndLabel[0], refAndLabel[0], refAndLabel[1]) : refAndLabel[0];
 			}
 			String term = value;
 			if (param.equals("q")) {
@@ -210,11 +213,9 @@ private static String lifeDates(JsonNode node) {
 	String[] refAndLabel(String property, String value,
 			Optional<List<String>> labels) {
 		if ((property.equals("containedIn") || property.equals("hasPart")
-				|| property.equals("isPartOf") || property.equals("hasSuperordinate"))
-				&& value.contains("lobid.org")) {
-			return new String[] {
-					value.replaceAll("https://lobid.org/resources?/", "http://rpb.lobid.org/"),
-					Lobid.resourceLabel(value) };
+				|| property.equals("isPartOf") || property.equals("hasSuperordinate")
+				|| property.equals("bibliographicCitation")) && value.contains("lobid.org")) {
+			return new String[] { value, Lobid.resourceLabel(value) };
 		}
 		String label =
 				labels.isPresent() && labels.get().size() > 0 ? labels.get().get(0)

app/views/tags/result_doc.scala.html

@@ -137,6 +137,7 @@
 }
 
 @subjects(subjects: Seq[JsValue]) = {
+  @if(!subjects.isEmpty) {
   <tr>
     <td>Schlagwörter</td>
     <td>
@@ -157,6 +158,7 @@
       } else { | }
     }</td>
   </tr>
+  }
 }
 
 @sortedPublications(seq: Seq[JsValue]) = @{

build.sbt

@@ -45,6 +45,8 @@ lazy val root = (project in file(".")).enablePlugins(PlayJava)
 
 javacOptions ++= Seq("-source", "1.8", "-target", "1.8")
 
+javaOptions += "-Dnet.sf.ehcache.disabled=true"
+
 import com.typesafe.sbteclipse.core.EclipsePlugin.EclipseKeys
 
 EclipseKeys.projectFlavor := EclipseProjectFlavor.Java // Java project. Don't expect Scala IDE

conf/nwbib.conf

@@ -4,6 +4,7 @@ item.api="http://lobid.org/items"
 hbz01.api="http://lobid.org/hbz01"
 orgs.api="http://lobid.org/organisations"
 nwbib.filter="inCollection.id:\"http://lobid.org/resources/HT013494180#!\""
+secret=""
 
 # Embedded Elasticsearch index for classification data
 index {

conf/nwbib.routes

@@ -33,4 +33,5 @@ GET    /robots.txt            controllers.Assets.at(path="/public",file="robots.
 GET    /cgi-bin/wwwalleg/:name.pl   controllers.nwbib.Application.showPl(name, db ?= "rnam", index: Int ?= 1, zeilen: Int ?= 1, s1)
 GET    /sw/:rpbId             controllers.nwbib.Application.showSw(rpbId)
 GET    /o:id                  controllers.nwbib.Application.searchSpatial(id, from:Int?=0, size:Int?=25, format?="html")
-GET    /:id                   controllers.nwbib.Application.show(id, format ?= "")
+GET    /:id                   controllers.nwbib.Application.show(id, format ?= "")
+PUT    /:id                   controllers.nwbib.Application.put(id, secret ?= "")

conf/rpb-titel-to-lobid.fix

@@ -28,6 +28,9 @@ copy_field("rpbId", "_id")
 # ------- Set "type" -------
 
 set_array("type[]", "BibliographicResource")
+unless exists("type")
+  set_field("type[].$append", "Article")
+end
 if all_equal("type", "u")
   set_field("type[].$append", "Article")
 end
@@ -40,7 +43,10 @@ end
 if any_equal("type", "Band")
   set_field("type[].$append", "Book")
   call_macro("move_here", field: "isPartOf[]")
-  copy_field("isPartOf[].value", "superordinateLabel")
+  copy_field("isPartOf[].label", "superordinateLabel")
+  unless exists("superordinateLabel")
+    copy_field("isPartOf[].value", "superordinateLabel")
+  end
   set_array("isPartOf[]")
   set_array("isPartOf[].$append.type[]", "IsPartOfRelation")
   set_array("isPartOf[].$last.hasSuperordinate[]")
@@ -120,9 +126,14 @@ if exists("isPartOf[]")
     set_array("temp")
     set_array("temp.$append.type[]", "IsPartOfRelation")
     set_array("temp.$last.hasSuperordinate[]")
-    split_field("isPartOf[].1.value", " ; ")
-    copy_field("isPartOf[].1.value.1", "temp.$last.hasSuperordinate[].$append.label")
-    copy_field("isPartOf[].1.value.2", "temp.$last.numbering")
+    if exists("isPartOf[].1.label")
+      copy_field("isPartOf[].1.value", "temp.$last.hasSuperordinate[].$append.id")
+      copy_field("isPartOf[].1.label", "temp.$last.hasSuperordinate[].$last.label")
+    else
+      split_field("isPartOf[].1.value", " ; ")
+      copy_field("isPartOf[].1.value.1", "temp.$last.hasSuperordinate[].$append.label")
+      copy_field("isPartOf[].1.value.2", "temp.$last.numbering")
+    end
     move_field("temp", "isPartOf[]")
   end
 end
@@ -204,10 +215,12 @@ end
 do
  list (path:"temp_subject.*.componentList[]", "var": "$i")
   move_field("$i.value", "$i.id")
-  copy_field("$i.id", "$i.label")
-  replace_all("$i.label","^http.+[/#](.+)$","$1")
-  replace_all("$i.label","^(\\d+)-(\\d+)$","$1n$2")
-  lookup("$i.label", "./RPB-Export_HBZ_SW.tsv","sep_char":"\t")
+  unless exists("$i.label")
+    copy_field("$i.id", "$i.label")
+    replace_all("$i.label","^http.+[/#](.+)$","$1")
+    replace_all("$i.label","^(\\d+)-(\\d+)$","$1n$2")
+    lookup("$i.label", "./RPB-Export_HBZ_SW.tsv","sep_char":"\t")
+  end
   if all_match("$i.id","^http://rpb.lobid.org/sw/.*$")
     add_field("$i.source.id", "http://rpb.lobid.org/sw")
     add_field("$i.source.label", "RPB-Sachsystematik")
@@ -283,11 +296,14 @@ do put_macro("contributions")
   set_array("_temp")
   copy_field("$[from][].*.value", "_temp.$append.agent.id")
   do list_as(f: "$[from][]", c: "_temp")
-    copy_field("c.agent.id", "c.agent.label")
+    copy_field("f.label", "c.agent.label")
+    unless exists("c.agent.label")
+      copy_field("c.agent.id", "c.agent.label")
+      replace_all("c.agent.label","^http.+[/#](.+)$","$1")
+      replace_all("c.agent.label","^(\\d+)-(\\d+)$","$1n$2")
+      lookup("c.agent.label", "./RPB-Export_HBZ_SW.tsv","sep_char":"\t")
+    end
     set_array("c.agent.type[]", "Person")
-    replace_all("c.agent.label","^http.+[/#](.+)$","$1")
-    replace_all("c.agent.label","^(\\d+)-(\\d+)$","$1n$2")
-    lookup("c.agent.label", "./RPB-Export_HBZ_SW.tsv","sep_char":"\t")
     copy_field("f.type", "c.role.id")
     copy_field("f.type", "c.role.label")
     lookup("c.role.id", "role-to-uri")