From 058ef3d39029da492477c3c59f1e405193ca3013 Mon Sep 17 00:00:00 2001
From: Maurice Laveaux <m.laveaux@student.tue.nl>
Date: Mon, 19 May 2014 11:11:22 +0200
Subject: Updated the brandchecker code and rules.

	* Changed the brandrules.txt ruleset to conform
	the new parsing of rules.
	* Brandchecker now doesn't count multiple equal words.
	* Updated the testcases, which don't work for others.
---
 src/analysis/BrandChecker.java | 91 +++++++++++++++++++++++++++++++-----------
 1 file changed, 68 insertions(+), 23 deletions(-)

(limited to 'src')

diff --git a/src/analysis/BrandChecker.java b/src/analysis/BrandChecker.java
index 740424c..5bde55b 100644
--- a/src/analysis/BrandChecker.java
+++ b/src/analysis/BrandChecker.java
@@ -19,6 +19,7 @@ import java.io.FileNotFoundException;
 import java.io.InputStream;
 import java.util.ArrayList;
 import java.util.Arrays;
+import java.util.HashMap;
 import java.util.HashSet;
 import java.util.List;
 import java.util.Locale;
@@ -57,6 +58,7 @@ public class BrandChecker {
      * @return The list of brands that are contained in this text or null.
      */
     public List<String> getBrands(String text) {
+        text = removePunct(text);
         String[] words = text.toLowerCase().split("\\s+");
 
         List<String> brands = new ArrayList();
@@ -70,6 +72,9 @@ public class BrandChecker {
         return brands;
     }
 
+    /**
+     * Reads the file and parses the rules, which are added to the ruleset.
+     */
     private void readFile(final String filename) throws FileNotFoundException {
         InputStream inFile = new FileInputStream(filename);
         Scanner readFile = new Scanner(inFile);
@@ -81,29 +86,52 @@ public class BrandChecker {
         }
     }
 
+    /**
+     * Parses the line and adds the BrandRule to the ruleset.
+     */
     private void parseRule(String line) {
         if (line.isEmpty()) {
             return;
         }
-
+        
         if (!line.contains("-")) {
-            // only positive search entries.
-            String[] sequence = line.split("\\s+");
-            String[] blacklist = {""};
-            ruleset.add(new BrandRule(line, sequence, blacklist));
+            System.err.println("illformatted rule: " + line + ", missing -");
         } else {
             String[] parts = line.split("-");
             // positive and negative.
             if (parts.length < 2) {
-                throw new IllegalArgumentException("Brand rule contained '-' but not two parts.");
+                System.err.println("illformatted rule: " + line + ", missing <name> - <positive>.");
+                return;
             }
 
-            String[] sequence = parts[0].trim().split("\\s+");
-            String[] blacklist = parts[1].trim().split("\\s+");
+            if (parts.length > 4) {
+                System.err.println("illformatted rule: " + line + ", forth part with - was given thus will be ignored.");
+            }
 
-            ruleset.add(new BrandRule(parts[0].trim(), sequence, blacklist));
+            // Read the <name> line.
+            String name = parts[0].trim();
+            
+            // Read the positive words.
+            String positive = parts[1].replaceAll(" ","");            
+            String[] sequence = positive.split(",");
+            
+            if (parts.length == 3) {
+                String negative = parts[2].replaceAll(" ", "");                
+                String[] blacklist = negative.split(",");
+                ruleset.add(new BrandRule(name, sequence, blacklist));
+            } else {
+                ruleset.add(new BrandRule(name, sequence, null));
+            }
         }
+    }
 
+    /**
+     * Removes punctuation and urls.
+     */
+    private String removePunct(String text) {
+        //text = text.replaceAll("https?://\\S*", "");
+        text = text.replaceAll(",", " ");
+        return text;
     }
 
     private class BrandRule {
@@ -111,8 +139,8 @@ public class BrandChecker {
         /**
          * The words that should be in the text.
          */
-        private final ArrayList<String> names;
-
+        private final HashMap<String, Boolean> names;
+        
         /**
          * A blacklist of words that are not interesting.
          */
@@ -131,8 +159,16 @@ public class BrandChecker {
          */
         public BrandRule(final String brandname, final String[] names, final String[] blacklist) {
             this.brand = brandname;
-            this.names = new ArrayList(Arrays.asList(names));
-            this.blacklist = new HashSet(Arrays.asList(blacklist));
+            this.names = new HashMap();
+            if (blacklist != null) {
+                this.blacklist = new HashSet(Arrays.asList(blacklist));
+            } else {
+                this.blacklist = null;
+            }
+            
+            for (String name : names) {
+                this.names.put(name, Boolean.FALSE);
+            }
         }
 
         /**
@@ -141,27 +177,36 @@ public class BrandChecker {
          * @param words A list of words in a line.
          */
         public boolean analyze(String[] words) {
+            reset();
+            
             int found = 0;
 
             for (String word : words) {
-                if (blacklist.contains(word)) {
-                    return false;
-                }
-                
-                if (names.contains(word)) {
-                    found++;
+                if (blacklist != null) {
+                    if (blacklist.contains(word)) {
+                        return false;
+                    }
                 }
-            }
 
-            if (found == names.size()) {
-                return true;
+                if (names.containsKey(word)) {
+                    if (names.get(word) == false) {
+                        found++;
+                        names.put(word, Boolean.TRUE);
+                    }
+                }
             }
 
-            return false;
+            return found == names.size();
         }
 
         public String getBrand() {
             return brand;
         }
+        
+        private void reset() {
+            for (String name : this.names.keySet()) {
+                this.names.put(name, Boolean.FALSE);
+            }
+        }
     }
 }
-- 
cgit v1.2.1