package com.hughes.android.dictionary.engine;
-import java.io.File;
-import java.io.FileNotFoundException;
-import java.io.IOException;
-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;
import java.util.Map;
import junit.framework.TestCase;
+import com.hughes.android.dictionary.parser.enwiktionary.EnWiktionaryLangs;
+
public class DictionaryBuilderMain extends TestCase {
- static final String INPUTS = "../DictionaryData/inputs/";
- static final String STOPLISTS = "../DictionaryData/inputs/stoplists/";
- static final String OUTPUTS = "../DictionaryData/outputs/";
+ static final String INPUTS = "data/inputs/";
+ static final String STOPLISTS = "data/inputs/stoplists/";
+ static final String OUTPUTS = "data/outputs/";
public static void main(final String[] args) throws Exception {
- final Map<String,String> isoToWikiName = new LinkedHashMap<String, String>(Language.isoCodeToWikiName);
+ // Builds all the dictionaries it can, outputs list to a text file.
+
+ final Map<String,String> isoToWikiName = new LinkedHashMap<String, String>(EnWiktionaryLangs.isoCodeToWikiName);
isoToWikiName.remove("EN");
isoToWikiName.remove("DE");
final Map<String,String> isoToDedication = new LinkedHashMap<String, String>();
isoToDedication.put("AF", "Afrikaans dictionary dedicated to Heiko and Mariƫtte Horn.");
- isoToDedication.put("HR", "Croation dictionary dedicated to Ines Viskic and Miro Kresonja.");
+ isoToDedication.put("HR", "Croatian dictionary dedicated to Ines Viskic and Miro Kresonja.");
isoToDedication.put("NL", "Dutch dictionary dedicated to Mike LeBeau.");
// German handled in file.
isoToDedication.put("EL", "Greek dictionary dedicated to Noah Egge.");
isoToStoplist.put("FR", "fr.txt");
final Map<String,String> isoToRegex = new LinkedHashMap<String, String>();
- isoToRegex.put("ZH", "Chinese|Mandarin|Cantonese");
+ // HACK: The missing "e" prevents a full match, causing "Cantonese" to be appended to the entries.
+ isoToRegex.put("ZH", "Chinese|Mandarin|Cantones");
// isoToWikiName.keySet().retainAll(Arrays.asList("UK", "HR", "FI"));
- boolean go = true;
- isoToWikiName.clear();
+ //isoToWikiName.clear();
+ boolean go = true;
for (final String foreignIso : isoToWikiName.keySet()) {
- if (foreignIso.equals("SV")) {
+ if (foreignIso.equals("blah")) {
go = true;
}
if (!go) {
continue;
}
- final String dictFile = String.format(OUTPUTS + "/EN-%s_enwiktionary.quickdic", foreignIso);
+ final String dictFile = String.format("%s/EN-%s_enwiktionary.quickdic", OUTPUTS, foreignIso);
System.out.println("building dictFile: " + dictFile);
if (!isoToStoplist.containsKey(foreignIso)) {
String.format("--lang2=%s", foreignIso),
String.format("--lang1Stoplist=%s", STOPLISTS + isoToStoplist.get("EN")),
String.format("--lang2Stoplist=%s", STOPLISTS + isoToStoplist.get(foreignIso)),
- String.format("--dictInfo=(EN)Wikitionary-based EN-%s dictionary. %s", foreignIso, isoToDedication.get(foreignIso)),
+ String.format("--dictInfo=(EN)Wikitionary-based EN-%s dictionary.\n\n%s", foreignIso, isoToDedication.get(foreignIso)),
"--input2=" + INPUTS + "enWikiSplit/" + foreignIso + ".data",
"--input2Name=enwiktionary." + foreignIso,
});
- // Print the entries for diffing.
- printToText(dictFile);
-
} // foreignIso
- final String dictFile = OUTPUTS + "DE-EN_chemnitz_enwiktionary.quickdic";
+ final String dictFile = String.format("%s/DE-EN_chemnitz_enwiktionary.quickdic", OUTPUTS);
DictionaryBuilder.main(new String[] {
"--dictOut=" + dictFile,
"--lang1=DE",
"--input3LangCodePattern=de",
"--input3EnIndex=2",
});
- printToText(dictFile);
}
-
- static void printToText(final String dictFile) throws IOException {
- final RandomAccessFile raf = new RandomAccessFile(new File(dictFile), "r");
- final Dictionary dict = new Dictionary(raf);
- final PrintWriter textOut = new PrintWriter(new File(dictFile + ".text"));
- final List<PairEntry> sorted = new ArrayList<PairEntry>(dict.pairEntries);
- Collections.sort(sorted);
- for (final PairEntry pairEntry : sorted) {
- textOut.println(pairEntry.getRawText(false));
- }
- textOut.close();
- raf.close();
- }
-
+
}