Agustin ZORZANO et Helder BETIOL

oracle-of-bacon-backend/src/main/java/com/serli/oracle/of/bacon/api/APIEndPoint.java

@@ -67,11 +67,12 @@ public String getConnectionsToKevinBacon(String actorName) {
 
     @Get("suggest?q=:searchQuery")
     public List<String> getActorSuggestion(String searchQuery) throws IOException {
-        return Arrays.asList("Niro, Chel",
-                "Senanayake, Niro",
-                "Niro, Juan Carlos",
-                "de la Rua, Niro",
-                "Niro, Simão");
+//        return Arrays.asList("Niro, Chel",
+//                "Senanayake, Niro",
+//                "Niro, Juan Carlos",
+//                "de la Rua, Niro",
+//                "Niro, Simão");
+        return elasticSearchRepository.getActorsSuggests(searchQuery);
     }
 
     @Get("last-searches")

oracle-of-bacon-backend/src/main/java/com/serli/oracle/of/bacon/repository/ElasticSearchRepository.java

@@ -1,13 +1,29 @@
 package com.serli.oracle.of.bacon.repository;
 
 import java.io.IOException;
+import java.util.ArrayList;
+import java.util.LinkedList;
 import java.util.List;
+import java.util.stream.Collectors;
 
 import org.apache.http.HttpHost;
+import org.elasticsearch.action.search.SearchRequest;
+import org.elasticsearch.action.search.SearchResponse;
+import org.elasticsearch.action.search.SearchType;
+import org.elasticsearch.client.RequestOptions;
 import org.elasticsearch.client.RestClient;
 import org.elasticsearch.client.RestHighLevelClient;
+import org.elasticsearch.index.query.QueryBuilders;
+import org.elasticsearch.search.builder.SearchSourceBuilder;
+import org.elasticsearch.search.suggest.Suggest;
+import org.elasticsearch.search.suggest.SuggestBuilder;
+import org.elasticsearch.search.suggest.SuggestBuilders;
+import org.elasticsearch.search.suggest.SuggestionBuilder;
+import org.elasticsearch.search.suggest.completion.CompletionSuggestion;
+import org.elasticsearch.search.suggest.completion.CompletionSuggestionBuilder;
 
 public class ElasticSearchRepository {
+    public static final String INDEX = "actors";
 
     private final RestHighLevelClient client;
 
@@ -26,6 +42,46 @@ public static RestHighLevelClient createClient() {
 
     public List<String> getActorsSuggests(String searchQuery) throws IOException {
         // TODO
-        return null;
+        /*
+            GET actors/_search
+            {
+              "suggest": {
+                "mySuggest": {
+                  "text": "niro",
+                  "completion": {
+                    "field": "suggest",
+                    "skip_duplicates": true
+                  }
+                }
+              }
+            }
+        */
+
+        final List<String> suggestions = new ArrayList<String>();
+
+        SearchRequest searchRequest = new SearchRequest(ElasticSearchRepository.INDEX);
+        SuggestBuilder suggestBuilder = new SuggestBuilder();
+        SearchSourceBuilder searchBuilder = new SearchSourceBuilder();
+
+
+        SuggestionBuilder suggestionBuilder = SuggestBuilders.completionSuggestion("suggest")
+                .size(5).prefix(searchQuery).skipDuplicates(true);
+        suggestBuilder.addSuggestion("suggestion", suggestionBuilder);
+        searchBuilder.suggest(suggestBuilder);
+
+        searchRequest.source(searchBuilder);
+        SearchResponse searchResponse = client.search(searchRequest, RequestOptions.DEFAULT);
+
+        CompletionSuggestion completionSuggestion = searchResponse.getSuggest().getSuggestion("suggestion");
+        List<CompletionSuggestion.Entry> entries = completionSuggestion.getEntries();
+
+        entries.forEach(entry -> {
+            entry.getOptions().forEach(option -> {
+                suggestions.add(option.getHit().getSourceAsMap().get("name").toString());
+            });
+        });
+
+        return suggestions;
+
     }
 }

oracle-of-bacon-backend/src/main/java/com/serli/oracle/of/bacon/repository/Neo4JRepository.java

@@ -21,6 +21,24 @@ public List<Map<String, GraphItem>> getConnectionsToKevinBacon(String actorName)
         Session session = driver.session();
 
         // TODO
+        // Result result = session.writeTransaction( new TransactionWork<String>()
+        //     {
+        //         @Override
+        //         public String execute( Transaction tx )
+        //         {
+        //             Result result = tx.run( "MATCH p= " +
+        //                                              "(bacon:Person {name:\"Bacon, Kevin (I)\"})-[*]-(actor:Person {name:\""+
+        //                                              actorName+"\"}) " + "RETURN p");
+        //             return result;
+        //         }
+        //     } );
+        // System.out.println( result );
+        /* Le plus court chemin :
+        MATCH p=shortestPath(
+          (bacon:Person {name:"Kevin Bacon"})-[*]-(actor:Person {name:"Al Pacino"})
+        )
+        RETURN p
+        */
         return null;
     }
 

script/insert.js

@@ -3,21 +3,42 @@ const fs = require("fs");
 const { Client } = require("@elastic/elasticsearch");
 
 const client = new Client({ node: "http://localhost:9200" });
+const INDEX_NAME = 'actors';
 
 async function insert() {
-  // TODO créer l'index (et plus pour la ré-exécution ?)
+  // créer l'index (et delete pour la ré-exécution)
+  await client.indices.delete({
+    index: INDEX_NAME,
+    ignore_unavailable: true
+  });
+
+  await client.indices.create({
+    index: INDEX_NAME,
+    body : {
+      "mappings": {
+        "properties": {
+          "name": { "type": "text" },
+          "suggest": { "type": "completion" },
+        }
+      }
+    }
+  });
 
   let actors = [];
   let first = true;
-  fs.createReadStream("./imdb-data/actors.csv")
+  fs.createReadStream("/home/agustin/Desktop/IMT/big_data/actors.csv")
     .pipe(csv())
     // Pour chaque ligne on créé un document JSON pour l'acteur correspondant
-    .on("data", async ({ name }) => {
-      // TODO ajouter les acteurs au tableau
+    .on("data", async ({ name }) => { // on a change la premiere ligne du csv : name:ID -> name
+      // ajouter les acteurs dans la liste
+      actors.push(name);
     })
     // A la fin on créé l'ensemble des acteurs dans ElasticSearch
-    .on("end", () => {
-      // TODO insérer dans elastic (les fonctions ci-dessous peuvent vous aider)
+    .on("end", async () => {
+      // insérer dans elastic
+      recBulk(client, createBulkInsertQueries(actors, 100000)).then(res => {
+        console.log(res);
+      });
     });
 }
 
@@ -35,18 +56,28 @@ function recBulk(client, bulks) {
 }
 
 function createBulkInsertQueries(names, length) {
-  // TODO
+  let i;
+  let slices = []
+  let bulks = []
+  for(i = 0; i <= names.length / length; i++) {
+    slices.push(names.slice(i, i + length));
+  }
+  slices.forEach(slice => {
+    bulks.push(createBulkInsertQuery(slice))
+  });
+  return bulks;
 }
 
 // Fonction utilitaire permettant de formatter les données pour l'insertion "bulk" dans elastic
 function createBulkInsertQuery(names) {
   const body = names.reduce((acc, name) => {
+
     const suggest = [name];
     const parts = name.split(",");
     parts.forEach((part) => suggest.push(part.trim()));
     parts.reverse().forEach((part) => suggest.push(part.trim()));
 
-    acc.push({ index: { _index: "actor", _type: "_doc" } });
+    acc.push({ index: { _index: "actors", _type: "_doc" } });
     acc.push({ name, suggest });
     return acc;
   }, []);