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         if (fields.length < 2 || fields.length > 4) {
 114             logger.warning("Malformed line, expected 3 or 4 fields, got " + fields.length + ": " + line);
 115             return;
 116         }
 117
 118         fields[0] = SPACES.matcher(fields[0]).replaceAll(" ").trim();
 119         fields[1] = SPACES.matcher(fields[1]).replaceAll(" ").trim();
 120         if (flipCols) {
 121             final String temp = fields[0];
 122             fields[0] = fields[1];
 123             fields[1] = temp;
 124         }
 125
 126         final String[][] subfields = new String[2][];
 127         if (subfieldSplit != null) {
 128             subfields[0] = subfieldSplit.split(fields[0]);
 129             subfields[1] = subfieldSplit.split(fields[1]);
 130             if (subfields[0].length != subfields[1].length) {
 131                 logger.warning("Number of subfields doesn't match: " + line);
 132                 return;
 133             }
 134         } else {
 135             subfields[0] = new String[] { fields[0] };
 136             subfields[1] = new String[] { fields[1] };
 137         }
 138
 139         final PairEntry pairEntry = new PairEntry(entrySource);
 140         for (int i = 0; i < subfields[0].length; ++i) {
 141             subfields[0][i] = subfields[0][i].trim();
 142             subfields[1][i] = subfields[1][i].trim();
 143             if (subfields[0][i].length() == 0 && subfields[1][i].length() == 0) {
 144                 logger.warning("Empty pair: " + line);
 145                 continue;
 146             }
 147             if (subfields[0][i].length() == 0) {
 148                 subfields[0][i] = "__";
 149             }
 150             if (subfields[1][i].length() == 0) {
 151                 subfields[1][i] = "__";
 152             }
 153             pairEntry.pairs.add(new PairEntry.Pair(subfields[0][i], subfields[1][i]));
 154         }
 155         final IndexedEntry entryData = new IndexedEntry(pairEntry);
 156         entryData.isValid = true;
 157
 158         for (int l = 0; l < 2; ++l) {
 159             // alreadyDone.clear();
 160
 161             for (int j = 0; j < subfields[l].length; ++j) {
 162                 String subfield = subfields[l][j];
 163                 final IndexBuilder indexBuilder = langIndexBuilders[l];
 164                 if (indexBuilder.index.sortLanguage == Language.de) {
 165                     subfield = parseField_DE(indexBuilder, subfield, entryData, j);
 166                 } else if (indexBuilder.index.sortLanguage == Language.en) {
 167                     subfield = parseField_EN(indexBuilder, subfield, entryData, j);
 168                 }
 169                 parseFieldGeneric(indexBuilder, subfield, entryData, j, subfields[l].length);
 170             }
 171         }
 172     }
 173
 174     private void parseFieldGeneric(final IndexBuilder indexBuilder, String field,
 175                                    final IndexedEntry entryData, final int subfieldIdx, final int numSubFields) {
 176         // remove bracketed and parenthesized stuff.
 177         final StringBuilder bracketed = new StringBuilder();
 178         final StringBuilder parenthesized = new StringBuilder();
 179
 180         Matcher matcher;
 181         while ((matcher = BRACKETED.matcher(field)).find()) {
 182             bracketed.append(matcher.group(1)).append(" ");
 183             field = matcher.replaceFirst(" ");
 184         }
 185
 186         while ((matcher = PARENTHESIZED.matcher(field)).find()) {
 187             parenthesized.append(matcher.group(1)).append(" ");
 188             field = matcher.replaceFirst(" ");
 189         }
 190
 191         field = SPACES.matcher(field).replaceAll(" ").trim();
 192
 193         // split words on non -A-z0-9, do them.
 194         final String[] tokens = NON_CHAR_DASH.split(field);
 195
 196         final EntryTypeName entryTypeName;
 197         if (numSubFields == 1) {
 198             assert subfieldIdx == 0;
 199             if (tokens.length == 1) {
 200                 entryTypeName = EntryTypeName.ONE_WORD;
 201             } else if (tokens.length == 2) {
 202                 entryTypeName = EntryTypeName.TWO_WORDS;
 203             } else if (tokens.length == 3) {
 204                 entryTypeName = EntryTypeName.THREE_WORDS;
 205             } else if (tokens.length == 4) {
 206                 entryTypeName = EntryTypeName.FOUR_WORDS;
 207             } else {
 208                 entryTypeName = EntryTypeName.FIVE_OR_MORE_WORDS;
 209             }
 210         } else {
 211             assert numSubFields > 1;
 212             if (subfieldIdx == 0) {
 213                 if (tokens.length == 1) {
 214                     entryTypeName = EntryTypeName.MULTIROW_HEAD_ONE_WORD;
 215                 } else {
 216                     entryTypeName = EntryTypeName.MULTIROW_HEAD_MANY_WORDS;
 217                 }
 218             } else {
 219                 assert subfieldIdx > 0;
 220                 if (tokens.length == 1) {
 221                     entryTypeName = EntryTypeName.MULTIROW_TAIL_ONE_WORD;
 222                 } else {
 223                     entryTypeName = EntryTypeName.MULTIROW_TAIL_MANY_WORDS;
 224                 }
 225             }
 226         }
 227
 228         for (String token : tokens) {
 229             token = TRIM_PUNC.matcher(token).replaceAll("");
 230             if (/*!alreadyDone.contains(token) && */token.length() > 0) {
 231                 indexBuilder.addEntryWithTokens(entryData, Collections.singleton(token), entryTypeName);
 232                 // alreadyDone.add(token);
 233
 234                 // also split words on dashes, do them, too.
 235                 if (token.contains("-")) {
 236                     final String[] dashed = token.split("-");
 237                     for (final String dashedToken : dashed) {
 238                         if (/*!alreadyDone.contains(dashedToken) && */dashedToken.length() > 0) {
 239                             indexBuilder.addEntryWithTokens(entryData, Collections.singleton(dashedToken), EntryTypeName.PART_OF_HYPHENATED);
 240                         }
 241                     }
 242                 }
 243
 244             }  // if (!alreadyDone.contains(token)) {
 245         }  // for (final String token : tokens) {
 246
 247         // process bracketed stuff (split on spaces and dashes always)
 248         final String[] bracketedTokens = NON_CHAR.split(bracketed.toString());
 249         for (final String token : bracketedTokens) {
 250             assert !token.contains("-");
 251             if (/*!alreadyDone.contains(token) && */token.length() > 0) {
 252                 indexBuilder.addEntryWithTokens(entryData, Collections.singleton(token), EntryTypeName.BRACKETED);
 253             }
 254         }
 255
 256         // process paren stuff
 257         final String[] parenTokens = NON_CHAR.split(parenthesized.toString());
 258         for (final String token : parenTokens) {
 259             assert !token.contains("-");
 260             if (/*!alreadyDone.contains(token) && */token.length() > 0) {
 261                 indexBuilder.addEntryWithTokens(entryData, Collections.singleton(token), EntryTypeName.PARENTHESIZED);
 262             }
 263         }
 264
 265     }
 266
 267     private String parseField_DE(final IndexBuilder indexBuilder, String field,
 268                                  final IndexedEntry entryData, final int subfieldIdx) {
 269
 270 //    final Matcher matcher = DE_NOUN.matcher(field);
 271 //    while (matcher.find()) {
 272 //      final String noun = matcher.group(1);
 273         //final String gender = matcher.group(2);
 274 //      if (alreadyDone.add(noun)) {
 275         // System.out.println("Found DE noun " + noun + ", " + gender);
 276 //        final List<EntryData> entries = indexBuilder.getOrCreateEntries(noun, EntryTypeName.NOUN);
 277 //        entries.add(entryData);
 278 //      }
 279 //    }
 280
 281         // In English, curly braces are used for different tenses.
 282         field = CURLY_BRACED.matcher(field).replaceAll(" ");
 283
 284         return field;
 285     }
 286
 287     private String parseField_EN(final IndexBuilder indexBuilder, String field,
 288                                  final IndexedEntry entryData, final int subfieldIdx) {
 289         if (field.startsWith("to ")) {
 290             field = field.substring(3);
 291         }
 292         return field;
 293     }
 294
 295     public static Set<String> tokenize(final String text, final Pattern pattern) {
 296         final String[] split = pattern.split(text);
 297         final Set<String> result = new LinkedHashSet<>(Arrays.asList(split));
 298         result.remove("");
 299         return result;
 300     }
 301
 302
 303 }