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