Changeset 2559

Timestamp:

11/03/08 16:39:41 (2 months ago)

Author:

huangb

Message:

get stopWords file from through BundleContext?; read the file, construct String[] stopWords, and pass it to StandardNormalyzer? algorithm

Files:

• trunk/plugins/preprocessing/edu.iu.nwb.preprocessing.text.normalization/src/edu/iu/nwb/preprocessing/text/normalization/StandardNormalyzerFactory.java (modified) (5 diffs)

trunk/plugins/preprocessing/edu.iu.nwb.preprocessing.text.normalization/src/edu/iu/nwb/preprocessing/text/normalization/StandardNormalyzerFactory.java

@@ -1 +1 @@
-package edu.iu.nwb.preprocessing.text.normalization;
-
+import java.io.BufferedReader;
-import java.io.IOException;
+import java.io.InputStream;
+import java.io.InputStreamReader;
-import java.util.ArrayList;
-import java.util.Dictionary;
-import java.util.List;
+import java.net.URL;
+import java.net.URLConnection;
-
-import org.cishell.framework.CIShellContext;
@@ -13 +18 @@
-import org.cishell.reference.service.metatype.BasicAttributeDefinition;
-import org.cishell.reference.service.metatype.BasicObjectClassDefinition;
+import org.osgi.framework.BundleContext;
+import org.osgi.service.component.ComponentContext;
-import org.osgi.service.metatype.AttributeDefinition;
-import org.osgi.service.metatype.ObjectClassDefinition;
@@ -23 +30 @@
-
-    protected static final String PREFIX = "column_";
-
-
-
+
+
+
+
+
+
+
+
-    }
+    
+        public Algorithm createAlgorithm(Data[] data, Dictionary parameters, 
+                         CIShellContext context) {
+                URL  filePath = bContext.getBundle().getResource(stopWordsFilePath);
+        stopWords = getStopWords (filePath);
+                return new StandardNormalyzer(data, parameters, context, stopWords);
+        }
-
-        public ObjectClassDefinition mutateParameters(Data[] data,
@@ -33 +52 @@
-
-                ObjectClassDefinition oldDefinition = parameters;
-
-                
-
-                String[] columnNames = createKeyArray(t.getSchema());
-                
@@ -78 +94 @@
-                return (String[]) keys.toArray(new String[]{});
-        }
+        
+    private String[] getStopWords(URL filePathURL) {
+        
+        InputStream inStream = null;
+        BufferedReader input = null;
+        String line;
+        ArrayList list = new ArrayList();
+        String[]stopWords = null;
+
+        try {
+             URLConnection connection = filePathURL.openConnection();
+             connection.setDoInput(true);
+             inStream = connection.getInputStream();
+             input = new BufferedReader(new InputStreamReader(inStream, "UTF-8"));
+                                
+            while (null != (line = input.readLine())) {
+                 list.add(line);
+            }
+                stopWords = new String[list.size()];
+                for (int ii=0; ii<list.size(); ii++){
+                        stopWords[ii] = (String) list.get(ii);
+                        System.out.println(">>Debug: index = "+ii+", value = "+stopWords[ii]);
+                }
+        }
+        catch (Exception e) {
+                e.printStackTrace();
+        }
+        finally {
+                try {
+                        if (input != null) input.close();
+                if (inStream != null) inStream.close();
+            }
+            catch (IOException e) {
+                e.printStackTrace();
+            }
+        }
+        return stopWords;
+        
+    }   
+
+        
+
-}