src/com/hughes/android/dictionary/parser/DictFileParser.java

   1 // Copyright 2011 Google Inc. All Rights Reserved.
   2 //
   3 // Licensed under the Apache License, Version 2.0 (the "License");
   4 // you may not use this file except in compliance with the License.
   5 // You may obtain a copy of the License at
   6 //
   7 //     http://www.apache.org/licenses/LICENSE-2.0
   8 //
   9 // Unless required by applicable law or agreed to in writing, software
  10 // distributed under the License is distributed on an "AS IS" BASIS,
  11 // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  12 // See the License for the specific language governing permissions and
  13 // limitations under the License.
  14
  15 package com.hughes.android.dictionary.parser;
  16
  17 import java.io.BufferedReader;
  18 import java.io.File;
  19 import java.io.FileInputStream;
  20 import java.io.IOException;
  21 import java.io.InputStreamReader;
  22 import java.nio.charset.Charset;
  23 import java.util.Arrays;
  24 import java.util.Collections;
  25 import java.util.LinkedHashSet;
  26 import java.util.List;
  27 import java.util.Set;
  28 import java.util.logging.Logger;
  29 import java.util.regex.Matcher;
  30 import java.util.regex.Pattern;
  31
  32 import com.hughes.android.dictionary.engine.DictionaryBuilder;
  33 import com.hughes.android.dictionary.engine.EntrySource;
  34 import com.hughes.android.dictionary.engine.IndexedEntry;
  35 import com.hughes.android.dictionary.engine.EntryTypeName;
  36 import com.hughes.android.dictionary.engine.IndexBuilder;
  37 import com.hughes.android.dictionary.engine.Language;
  38 import com.hughes.android.dictionary.engine.PairEntry;
  39 import com.hughes.android.dictionary.engine.PairEntry.Pair;
  40
  41 public class DictFileParser implements Parser {
  42
  43     static final Logger logger = Logger.getLogger(DictFileParser.class.getName());
  44
  45     // Dictcc
  46     public static final Pattern TAB = Pattern.compile("\\t");
  47
  48     // Chemnitz
  49     public static final Pattern DOUBLE_COLON = Pattern.compile(" :: ");
  50     public static final Pattern PIPE = Pattern.compile("\\|");
  51
  52     static final Pattern SPACES = Pattern.compile("\\s+");
  53
  54     static final Pattern BRACKETED = Pattern.compile("\\[([^]]+)\\]");
  55     static final Pattern PARENTHESIZED = Pattern.compile("\\(([^)]+)\\)");
  56     static final Pattern CURLY_BRACED = Pattern.compile("\\{([^}]+)\\}");
  57
  58     // http://www.regular-expressions.info/unicode.html
  59     static final Pattern NON_CHAR_DASH = Pattern.compile("[^-'\\p{L}\\p{M}\\p{N}]+");
  60     public static final Pattern NON_CHAR = Pattern.compile("[^\\p{L}\\p{M}\\p{N}]+");
  61
  62     static final Pattern TRIM_PUNC = Pattern.compile("^[^\\p{L}\\p{M}\\p{N}]+|[^\\p{L}\\p{M}\\p{N}]+$");
  63
  64     final Charset charset;
  65     final boolean flipCols;
  66
  67     final Pattern fieldSplit;
  68     final Pattern subfieldSplit;
  69
  70     final DictionaryBuilder dictBuilder;
  71     final IndexBuilder[] langIndexBuilders;
  72     final IndexBuilder bothIndexBuilder;
  73
  74     EntrySource entrySource;
  75
  76     // final Set<String> alreadyDone = new HashSet<String>();
  77
  78     public DictFileParser(final Charset charset, boolean flipCols,
  79                           final Pattern fieldSplit, final Pattern subfieldSplit,
  80                           final DictionaryBuilder dictBuilder, final IndexBuilder[] langIndexBuilders,
  81                           final IndexBuilder bothIndexBuilder) {
  82         this.charset = charset;
  83         this.flipCols = flipCols;
  84         this.fieldSplit = fieldSplit;
  85         this.subfieldSplit = subfieldSplit;
  86         this.dictBuilder = dictBuilder;
  87         this.langIndexBuilders = langIndexBuilders;
  88         this.bothIndexBuilder = bothIndexBuilder;
  89     }
  90
  91     @Override
  92     public void parse(final File file, final EntrySource entrySouce, final int pageLimit) throws IOException {
  93         this.entrySource = entrySouce;
  94         final BufferedReader reader = new BufferedReader(new InputStreamReader(new FileInputStream(file), charset));
  95         String line;
  96         int count = 0;
  97         while ((line = reader.readLine()) != null) {
  98             if (pageLimit >= 0 && count >= pageLimit) {
  99                 return;
 100             }
 101             if (count % 10000 == 0) {
 102                 logger.info("count=" + count + ", line=" + line);
 103             }
 104             parseLine(line);
 105             ++count;
 106         }
 107     }
 108
 109     private void parseLine(final String line) {
 110         if (line.startsWith("#") || line.length() == 0) {
 111             logger.info("Skipping comment line: " + line);
 112             return;
 113         }
 114         final String[] fields = fieldSplit.split(line);
 115         // dictcc now has a part of speech field as field #3.
 116         if (fields.length < 2 || fields.length > 3) {
 117             logger.warning("Malformed line: " + line);
 118             return;
 119         }
 120
 121         fields[0] = SPACES.matcher(fields[0]).replaceAll(" ").trim();
 122         fields[1] = SPACES.matcher(fields[1]).replaceAll(" ").trim();
 123         if (flipCols) {
 124             final String temp = fields[0];
 125             fields[0] = fields[1];
 126             fields[1] = temp;
 127         }
 128
 129         final String[][] subfields = new String[2][];
 130         if (subfieldSplit != null) {
 131             subfields[0] = subfieldSplit.split(fields[0]);
 132             subfields[1] = subfieldSplit.split(fields[1]);
 133             if (subfields[0].length != subfields[1].length) {
 134                 logger.warning("Number of subfields doesn't match: " + line);
 135                 return;
 136             }
 137         } else {
 138             subfields[0] = new String[] { fields[0] };
 139             subfields[1] = new String[] { fields[1] };
 140         }
 141
 142         final PairEntry pairEntry = new PairEntry(entrySource);
 143         for (int i = 0; i < subfields[0].length; ++i) {
 144             subfields[0][i] = subfields[0][i].trim();
 145             subfields[1][i] = subfields[1][i].trim();
 146             if (subfields[0][i].length() == 0 && subfields[1][i].length() == 0) {
 147                 logger.warning("Empty pair: " + line);
 148                 continue;
 149             }
 150             if (subfields[0][i].length() == 0) {
 151                 subfields[0][i] = "__";
 152             }
 153             if (subfields[1][i].length() == 0) {
 154                 subfields[1][i] = "__";
 155             }
 156             pairEntry.pairs.add(new Pair(subfields[0][i], subfields[1][i]));
 157         }
 158         final IndexedEntry entryData = new IndexedEntry(pairEntry);
 159         entryData.isValid = true;
 160
 161         for (int l = 0; l < 2; ++l) {
 162             // alreadyDone.clear();
 163
 164             for (int j = 0; j < subfields[l].length; ++j) {
 165                 String subfield = subfields[l][j];
 166                 final IndexBuilder indexBuilder = langIndexBuilders[l];
 167                 if (indexBuilder.index.sortLanguage == Language.de) {
 168                     subfield = parseField_DE(indexBuilder, subfield, entryData, j);
 169                 } else if (indexBuilder.index.sortLanguage == Language.en) {
 170                     subfield = parseField_EN(indexBuilder, subfield, entryData, j);
 171                 }
 172                 parseFieldGeneric(indexBuilder, subfield, entryData, j, subfields[l].length);
 173             }
 174         }
 175     }
 176
 177     private void parseFieldGeneric(final IndexBuilder indexBuilder, String field,
 178                                    final IndexedEntry entryData, final int subfieldIdx, final int numSubFields) {
 179         // remove bracketed and parenthesized stuff.
 180         final StringBuilder bracketed = new StringBuilder();
 181         final StringBuilder parenthesized = new StringBuilder();
 182
 183         Matcher matcher;
 184         while ((matcher = BRACKETED.matcher(field)).find()) {
 185             bracketed.append(matcher.group(1)).append(" ");
 186             field = matcher.replaceFirst(" ");
 187         }
 188
 189         while ((matcher = PARENTHESIZED.matcher(field)).find()) {
 190             parenthesized.append(matcher.group(1)).append(" ");
 191             field = matcher.replaceFirst(" ");
 192         }
 193
 194         field = SPACES.matcher(field).replaceAll(" ").trim();
 195
 196         // split words on non -A-z0-9, do them.
 197         final String[] tokens = NON_CHAR_DASH.split(field);
 198
 199         final EntryTypeName entryTypeName;
 200         if (numSubFields == 1) {
 201             assert subfieldIdx == 0;
 202             if (tokens.length == 1) {
 203                 entryTypeName = EntryTypeName.ONE_WORD;
 204             } else if (tokens.length == 2) {
 205                 entryTypeName = EntryTypeName.TWO_WORDS;
 206             } else if (tokens.length == 3) {
 207                 entryTypeName = EntryTypeName.THREE_WORDS;
 208             } else if (tokens.length == 4) {
 209                 entryTypeName = EntryTypeName.FOUR_WORDS;
 210             } else {
 211                 entryTypeName = EntryTypeName.FIVE_OR_MORE_WORDS;
 212             }
 213         } else {
 214             assert numSubFields > 1;
 215             if (subfieldIdx == 0) {
 216                 if (tokens.length == 1) {
 217                     entryTypeName = EntryTypeName.MULTIROW_HEAD_ONE_WORD;
 218                 } else {
 219                     entryTypeName = EntryTypeName.MULTIROW_HEAD_MANY_WORDS;
 220                 }
 221             } else {
 222                 assert subfieldIdx > 0;
 223                 if (tokens.length == 1) {
 224                     entryTypeName = EntryTypeName.MULTIROW_TAIL_ONE_WORD;
 225                 } else {
 226                     entryTypeName = EntryTypeName.MULTIROW_TAIL_MANY_WORDS;
 227                 }
 228             }
 229         }
 230
 231         for (String token : tokens) {
 232             token = TRIM_PUNC.matcher(token).replaceAll("");
 233             if (/*!alreadyDone.contains(token) && */token.length() > 0) {
 234                 indexBuilder.addEntryWithTokens(entryData, Collections.singleton(token), entryTypeName);
 235                 // alreadyDone.add(token);
 236
 237                 // also split words on dashes, do them, too.
 238                 if (token.contains("-")) {
 239                     final String[] dashed = token.split("-");
 240                     for (final String dashedToken : dashed) {
 241                         if (/*!alreadyDone.contains(dashedToken) && */dashedToken.length() > 0) {
 242                             indexBuilder.addEntryWithTokens(entryData, Collections.singleton(dashedToken), EntryTypeName.PART_OF_HYPHENATED);
 243                         }
 244                     }
 245                 }
 246
 247             }  // if (!alreadyDone.contains(token)) {
 248         }  // for (final String token : tokens) {
 249
 250         // process bracketed stuff (split on spaces and dashes always)
 251         final String[] bracketedTokens = NON_CHAR.split(bracketed.toString());
 252         for (final String token : bracketedTokens) {
 253             assert !token.contains("-");
 254             if (/*!alreadyDone.contains(token) && */token.length() > 0) {
 255                 indexBuilder.addEntryWithTokens(entryData, Collections.singleton(token), EntryTypeName.BRACKETED);
 256             }
 257         }
 258
 259         // process paren stuff
 260         final String[] parenTokens = NON_CHAR.split(parenthesized.toString());
 261         for (final String token : parenTokens) {
 262             assert !token.contains("-");
 263             if (/*!alreadyDone.contains(token) && */token.length() > 0) {
 264                 indexBuilder.addEntryWithTokens(entryData, Collections.singleton(token), EntryTypeName.PARENTHESIZED);
 265             }
 266         }
 267
 268     }
 269
 270     private String parseField_DE(final IndexBuilder indexBuilder, String field,
 271                                  final IndexedEntry entryData, final int subfieldIdx) {
 272
 273 //    final Matcher matcher = DE_NOUN.matcher(field);
 274 //    while (matcher.find()) {
 275 //      final String noun = matcher.group(1);
 276         //final String gender = matcher.group(2);
 277 //      if (alreadyDone.add(noun)) {
 278         // System.out.println("Found DE noun " + noun + ", " + gender);
 279 //        final List<EntryData> entries = indexBuilder.getOrCreateEntries(noun, EntryTypeName.NOUN);
 280 //        entries.add(entryData);
 281 //      }
 282 //    }
 283
 284         // In English, curly braces are used for different tenses.
 285         field = CURLY_BRACED.matcher(field).replaceAll(" ");
 286
 287         return field;
 288     }
 289
 290     private String parseField_EN(final IndexBuilder indexBuilder, String field,
 291                                  final IndexedEntry entryData, final int subfieldIdx) {
 292         if (field.startsWith("to ")) {
 293             field = field.substring(3);
 294         }
 295         return field;
 296     }
 297
 298     public static final Set<String> tokenize(final String text, final Pattern pattern) {
 299         final String[] split = pattern.split(text);
 300         final Set<String> result = new LinkedHashSet<String>(Arrays.asList(split));
 301         result.remove("");
 302         return result;
 303     }
 304
 305
 306 }