X-Git-Url: http://gitweb.fperrin.net/?a=blobdiff_plain;f=src%2Fcom%2Fhughes%2Fandroid%2Fdictionary%2Fengine%2FDictionaryBuilderMain.java;h=db5caa864562fc76322372fd2d5a7945090aa76a;hb=b3dd51971927861bdaeab3ea1569006c07653873;hp=36564eacce3ef918459e6c906849d9833c9af7a6;hpb=a7ae2524281869de5aa756ae35524b21bab3e08a;p=DictionaryPC.git

diff --git a/src/com/hughes/android/dictionary/engine/DictionaryBuilderMain.java b/src/com/hughes/android/dictionary/engine/DictionaryBuilderMain.java
index 36564ea..db5caa8 100644
--- a/src/com/hughes/android/dictionary/engine/DictionaryBuilderMain.java
+++ b/src/com/hughes/android/dictionary/engine/DictionaryBuilderMain.java
@@ -18,6 +18,7 @@ import java.io.File;
 import java.io.PrintWriter;
 import java.io.RandomAccessFile;
 import java.util.ArrayList;
+import java.util.Arrays;
 import java.util.Collections;
 import java.util.LinkedHashMap;
 import java.util.List;
@@ -44,7 +45,6 @@ public class DictionaryBuilderMain extends TestCase {
     // German handled in file.
     isoToDedication.put("EL", "Greek dictionary dedicated to Noah Egge.");
     isoToDedication.put("IT", "Italian dictionary dedicated to Carolina Tropini, my favorite stardust in the whole universe!  Ti amo!");
-    isoToDedication.put("JA", "Japanese dictionary dedicated to Akane Watanabe.");
     isoToDedication.put("KO", "Korean dictionary dedicated to Ande Elwood--fall fashion und Fernsehturms!");
     isoToDedication.put("PT", "Portuguese dictionary dedicated to Carlos Melo, one Tough Mudder.");
     isoToDedication.put("RO", "Romanian dictionary dedicated to Radu Teodorescu.");
@@ -61,12 +61,13 @@ public class DictionaryBuilderMain extends TestCase {
     isoToStoplist.put("FR", "fr.txt");
 
     final Map<String,String>  isoToRegex = new LinkedHashMap<String, String>();
-    isoToRegex.put("ZH", ".*Chinese.*|.*Mandarin.*|.*Cantonese.*");
+    isoToRegex.put("ZH", "Chinese|Mandarin|Cantonese");
     
-    boolean go = false; 
+//    isoToWikiName.keySet().retainAll(Arrays.asList("UK", "HR", "FI"));
+    boolean go = true; 
     isoToWikiName.clear();
     for (final String foreignIso : isoToWikiName.keySet()) {
-      if (foreignIso.equals("GA")) {
+      if (foreignIso.equals("SV")) {
         go = true;
       }
       if (!go) {
@@ -83,7 +84,7 @@ public class DictionaryBuilderMain extends TestCase {
           isoToDedication.put(foreignIso, "");
         }
         if (!isoToRegex.containsKey(foreignIso)) {
-          isoToRegex.put(foreignIso, ".*" + isoToWikiName.get(foreignIso) + ".*");
+          isoToRegex.put(foreignIso, isoToWikiName.get(foreignIso));
         }
   
         DictionaryBuilder.main(new String[] {
@@ -99,14 +100,14 @@ public class DictionaryBuilderMain extends TestCase {
             "--input2Format=enwiktionary",
             "--input2LangPattern=" + isoToRegex.get(foreignIso),
             "--input2LangCodePattern=" + foreignIso.toLowerCase(),
-            "--input2EnIndex=2",
+            "--input2EnIndex=1",
 
             "--input3=" + INPUTS + "enWikiSplit/EN.data",
             "--input3Name=enwiktionary.english",
             "--input3Format=enwiktionary",
             "--input3LangPattern=" + isoToRegex.get(foreignIso),
             "--input3LangCodePattern=" + foreignIso.toLowerCase(),
-            "--input3EnIndex=2",
+            "--input3EnIndex=1",
 
         });
         
@@ -125,27 +126,29 @@ public class DictionaryBuilderMain extends TestCase {
     }  // foreignIso
 
     DictionaryBuilder.main(new String[] {
-        "--dictOut=" + OUTPUTS + "DE-EN_chemnitz_enwiktionary",
+        "--dictOut=" + OUTPUTS + "DE-EN_chemnitz_enwiktionary.quickdic",
         "--lang1=DE",
         "--lang2=EN",
+        String.format("--lang1Stoplist=%s", STOPLISTS + "de.txt"),
+        String.format("--lang2Stoplist=%s", STOPLISTS + "en.txt"),
         "--dictInfo=@" + INPUTS + "de-en_chemnitz_enwiktionary.info",
 
-        "--input1=" + INPUTS + "de-en_chemnitz.txt",
-        "--input1Name=chemnitz",
-        "--input1Charset=UTF8",
-        "--input1Format=chemnitz",
+        "--input4=" + INPUTS + "de-en_chemnitz.txt",
+        "--input4Name=chemnitz",
+        "--input4Charset=UTF8",
+        "--input4Format=chemnitz",
         
         "--input2=" + INPUTS + "enWikiSplit/DE.data",
         "--input2Name=enwiktionary.DE",
         "--input2Format=enwiktionary",
-        "--input2LangPattern=" + isoToRegex.get("DE"),
+        "--input2LangPattern=German",
         "--input2LangCodePattern=de",
         "--input2EnIndex=2",
 
         "--input3=" + INPUTS + "enWikiSplit/EN.data",
         "--input3Name=enwiktionary.english",
         "--input3Format=enwiktionary",
-        "--input3LangPattern=" + isoToRegex.get("DE"),
+        "--input3LangPattern=German",
         "--input3LangCodePattern=de",
         "--input3EnIndex=2",