Changed the implementation of step 3 in checking for resolved hyperlinks

nitishgupta · Mar 26, 2018 · 1975fe7 · 1975fe7
1 parent f5df139
commit 1975fe7
Show file tree

Hide file tree

Showing 2 changed files with 31 additions and 0 deletions.
diff --git a/src/main/java/edu/illinois/cs/cogcomp/wikiparser/unifiedParsing/resolveHyperlinks.java b/src/main/java/edu/illinois/cs/cogcomp/wikiparser/unifiedParsing/resolveHyperlinks.java
@@ -15,6 +15,8 @@
  */
 public class resolveHyperlinks {
     private String file = "/shared/preprocessed/wikiparser/unifiedParserOutput/RedirectTitle2ResolvedTitle.txt";
+    private static String unresolvedOutput = "/shared/preprocessed/wikiparser/unifiedParserOutput/unresolvedHyperlinks.txt";
+    private static Set<String> unresolvedTitles = new HashSet<String>();
     private Set<String> resolvedTitles;
     private Map<String, String> redirectToResolved;
 
@@ -45,6 +47,22 @@ private void parseMap(){
         }
     }
 
+    public static void writeUnresolvedTitles(){
+      File unresolvedFile = new File(unresolvedOutput);
+      try{
+          FileWriter fw = new FileWriter(unresolvedFile.getAbsoluteFile());
+          BufferedWriter bw = new BufferedWriter(fw);
+          for(String title : unresolvedTitles){
+              bw.write(title + "\n");
+          }
+          bw.close();
+      }
+      catch (IOException e){
+          e.printStackTrace();
+          System.exit(-1);
+      }
+    }
+
     public Map<List<Integer>, String> resolve(Map<List<Integer>, String> hyperlinks){
         Set<List<Integer>> keys = hyperlinks.keySet();
         for(List<Integer> i : keys){
@@ -66,6 +84,13 @@ public Map<List<Integer>, String> resolve(Map<List<Integer>, String> hyperlinks)
                 } else{
                     // Capitalizes first letter
                     String cap_title = title.substring(0, 1).toUpperCase() + title.substring(1);
+                    if(resolvedTitles.contains(cap_title)){
+                        hyperlinks.replace(i, cap_title);
+                        continue;
+                    } else if(redirectToResolved.containsKey(title)){
+                        hyperlinks.replace(i, redirectToResolved.get(cap_title));
+                        continue;
+                    }
                     // Capitalizes all letters directly succeeding '_'
                     for(int c = 1; c < cap_title.length(); c++){
                         if(cap_title.charAt(c) == '_'){
@@ -79,6 +104,8 @@ public Map<List<Integer>, String> resolve(Map<List<Integer>, String> hyperlinks)
                         hyperlinks.replace(i, cap_title);
                     } else if(redirectToResolved.containsKey(title)){
                         hyperlinks.replace(i, redirectToResolved.get(cap_title));
+                    } else{
+                      unresolvedTitles.add(title);
                     }
                 }
             }

diff --git a/src/main/java/edu/illinois/cs/cogcomp/wikiparser/wikiparse/WikiExtractParser.java b/src/main/java/edu/illinois/cs/cogcomp/wikiparser/wikiparse/WikiExtractParser.java
@@ -68,6 +68,10 @@ public void extractWiki(){
         }
         logger.log.info("Total Files: " + Integer.toString(totalFiles));
         System.out.println("[#] Total Files: " + totalFiles);
+        while(parser.getActiveCount() > 0){
+          // wait
+        }
+        resolveHyperlinks.writeUnresolvedTitles();
     }
 
     public static void main(String [] args){