SDAP-118 create a new ranking module

core/pom.xml

@@ -151,16 +151,10 @@
             <resource>
                 <directory>${basedir}/src/main/resources</directory>
                 <filtering>true</filtering>
-                <excludes>
-                    <exclude>${svmSgdModel.value}/**</exclude>
-                </excludes>
             </resource>
 
             <resource>
                 <directory>${project.build.directory}</directory>
-                <includes>
-                    <include>${svmSgdModel.value}.zip</include>
-                </includes>
             </resource>
 
             <resource>
@@ -206,26 +200,6 @@
                 <artifactId>maven-assembly-plugin</artifactId>
                 <version>2.6</version>
                 <executions>
-                    <execution>
-                        <id>zipSVMWithSGDModel</id>
-                        <phase>generate-resources</phase>
-                        <goals>
-                            <goal>single</goal>
-                        </goals>
-                        <configuration>
-                            <appendAssemblyId>false</appendAssemblyId>
-                            <tarLongFileMode>posix</tarLongFileMode>
-                            <finalName>${svmSgdModel.value}</finalName>
-                            <outputDirectory>${project.build.directory}
-                            </outputDirectory>
-                            <descriptors>
-                                <descriptor>
-                                    ${basedir}/src/main/assembly/zipSVMWithSGDModel.xml
-                                </descriptor>
-                            </descriptors>
-                        </configuration>
-                    </execution>
-
                     <execution>
                         <id>generateDistribution</id>
                         <phase>package</phase>

core/src/main/java/org/apache/sdap/mudrod/weblog/process/ClickStreamAnalyzer.java

@@ -18,7 +18,6 @@
 import org.apache.sdap.mudrod.driver.SparkDriver;
 import org.apache.sdap.mudrod.main.MudrodConstants;
 import org.apache.sdap.mudrod.semantics.SVDAnalyzer;
-import org.apache.sdap.mudrod.ssearch.ClickStreamImporter;
 import org.apache.sdap.mudrod.utils.LinkageTriple;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
@@ -60,10 +59,7 @@ public Object execute() {
             props.getProperty(MudrodConstants.CLICKSTREAM_SVD_PATH));
         List<LinkageTriple> tripleList = svd.calTermSimfromMatrix(props.getProperty(MudrodConstants.CLICKSTREAM_SVD_PATH));
         svd.saveToES(tripleList, props.getProperty(MudrodConstants.ES_INDEX_NAME), MudrodConstants.CLICK_STREAM_LINKAGE_TYPE);
-
-        // Store click stream in ES for the ranking use
-        ClickStreamImporter cs = new ClickStreamImporter(props, es, spark);
-        cs.importfromCSVtoES();
+
       }
     } catch (Exception e) {
       LOG.error("Encountered an error during execution of ClickStreamAnalyzer.", e);

core/src/main/java/org/apache/sdap/mudrod/weblog/structure/session/Session.java

@@ -203,7 +203,7 @@ public List<ClickStream> getClickStreamList(String indexName, String type, Strin
    * @return an instance of session tree structure
    * @throws UnsupportedEncodingException UnsupportedEncodingException
    */
-  private SessionTree getSessionTree(String indexName, String type, String sessionID) throws UnsupportedEncodingException {
+  public SessionTree getSessionTree(String indexName, String type, String sessionID) throws UnsupportedEncodingException {
 
     SearchResponse response = es.getClient().prepareSearch(indexName).setTypes(type).setQuery(QueryBuilders.termQuery("SessionID", sessionID)).setSize(100).addSort("Time", SortOrder.ASC)
         .execute().actionGet();
@@ -261,31 +261,4 @@ private JsonElement getRequests(String cleanuptype, String sessionID) throws Uns
     }
     return gson.toJsonTree(requestList);
   }
-
-  /**
-   * getClickStreamList: Extracted ranking training data from current session.
-   *
-   * @param indexName    an index from which to obtain ranked training data.
-   * @param cleanuptype: Session type name in Elasticsearch
-   * @param sessionID:   Session ID
-   * @return Click stram data list
-   * {@link ClickStream}
-   */
-  public List<RankingTrainData> getRankingTrainData(String indexName, String cleanuptype, String sessionID) {
-    SessionTree tree = null;
-    try {
-      tree = this.getSessionTree(indexName, cleanuptype, sessionID);
-    } catch (UnsupportedEncodingException e) {
-      LOG.error("Error whilst retreiving Session Tree: {}", e);
-    }
-
-    List<RankingTrainData> trainData = new ArrayList<>();
-    try {
-      trainData = tree.getRankingTrainData(indexName);
-    } catch (UnsupportedEncodingException e) {
-      LOG.error("Error whilst retreiving ranking training data: {}", e);
-    }
-
-    return trainData;
-  }
 }