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
  70     EntrySource entrySource;
  71
  72     // final Set<String> alreadyDone = new HashSet<String>();
  73
  74     public DictFileParser(final Charset charset, boolean flipCols,
  75                           final Pattern fieldSplit, final Pattern subfieldSplit,
  76                           final DictionaryBuilder dictBuilder) {
  77         this.charset = charset;
  78         this.flipCols = flipCols;
  79         this.fieldSplit = fieldSplit;
  80         this.subfieldSplit = subfieldSplit;
  81         this.dictBuilder = dictBuilder;
  82     }
  83
  84     @Override
  85     public void parse(final File file, final EntrySource entrySouce, final int pageLimit) throws IOException {
  86         this.entrySource = entrySouce;
  87         final BufferedReader reader = new BufferedReader(new InputStreamReader(new FileInputStream(file), charset));
  88         String line;
  89         int count = 0;
  90         while ((line = reader.readLine()) != null) {
  91             if (pageLimit >= 0 && count >= pageLimit) {
  92                 return;
  93             }
  94             if (count % 10000 == 0) {
  95                 logger.info("count=" + count + ", line=" + line);
  96             }
  97             parseLine(line);
  98             ++count;
  99         }
 100     }
 101
 102     private void parseLine(final String line) {
 103         if (line.startsWith("#") || line.isEmpty()) {
 104             logger.info("Skipping comment line: " + line);
 105             return;
 106         }
 107         final String[] fields = fieldSplit.split(line);
 108         if (fields.length < 2 || fields.length > 4) {
 109             logger.warning("Malformed line, expected 3 or 4 fields, got " + fields.length + ": " + line);
 110             return;
 111         }
 112
 113         fields[0] = SPACES.matcher(fields[0]).replaceAll(" ").trim();
 114         fields[1] = SPACES.matcher(fields[1]).replaceAll(" ").trim();
 115         if (flipCols) {
 116             final String temp = fields[0];
 117             fields[0] = fields[1];
 118             fields[1] = temp;
 119         }
 120
 121         final String[][] subfields = new String[2][];
 122         if (subfieldSplit != null) {
 123             subfields[0] = subfieldSplit.split(fields[0]);
 124             subfields[1] = subfieldSplit.split(fields[1]);
 125             if (subfields[0].length != subfields[1].length) {
 126                 logger.warning("Number of subfields doesn't match: " + line);
 127                 return;
 128             }
 129         } else {
 130             subfields[0] = new String[] { fields[0] };
 131             subfields[1] = new String[] { fields[1] };
 132         }
 133
 134         final PairEntry pairEntry = new PairEntry(entrySource);
 135         for (int i = 0; i < subfields[0].length; ++i) {
 136             subfields[0][i] = subfields[0][i].trim();
 137             subfields[1][i] = subfields[1][i].trim();
 138             if (subfields[0][i].isEmpty() && subfields[1][i].isEmpty()) {
 139                 logger.warning("Empty pair: " + line);
 140                 continue;
 141             }
 142             if (subfields[0][i].isEmpty()) {
 143                 subfields[0][i] = "__";
 144             }
 145             if (subfields[1][i].isEmpty()) {
 146                 subfields[1][i] = "__";
 147             }
 148             pairEntry.pairs.add(new PairEntry.Pair(subfields[0][i], subfields[1][i]));
 149         }
 150         final IndexedEntry entryData = new IndexedEntry(pairEntry);
 151         entryData.isValid = true;
 152
 153         for (int l = 0; l < 2; ++l) {
 154             // alreadyDone.clear();
 155
 156             final IndexBuilder indexBuilder = dictBuilder.indexBuilders.get(l);
 157             for (int j = 0; j < subfields[l].length; ++j) {
 158                 String subfield = subfields[l][j];
 159                 if (indexBuilder.index.sortLanguage == Language.de) {
 160                     subfield = parseField_DE(indexBuilder, subfield, entryData, j);
 161                 } else if (indexBuilder.index.sortLanguage == Language.en) {
 162                     subfield = parseField_EN(indexBuilder, subfield, entryData, j);
 163                 }
 164                 parseFieldGeneric(indexBuilder, subfield, entryData, j, subfields[l].length);
 165             }
 166         }
 167     }
 168
 169     private void parseFieldGeneric(final IndexBuilder indexBuilder, String field,
 170                                    final IndexedEntry entryData, final int subfieldIdx, final int numSubFields) {
 171         // remove bracketed and parenthesized stuff.
 172         final StringBuilder bracketed = new StringBuilder();
 173         final StringBuilder parenthesized = new StringBuilder();
 174
 175         if (field.indexOf('[') != -1) {
 176             StringBuilder stripped = new StringBuilder(field.length());
 177             Matcher matcher = BRACKETED.matcher(field);
 178             while (matcher.find()) {
 179                 bracketed.append(matcher.group(1)).append(" ");
 180                 matcher.appendReplacement(stripped, " ");
 181             }
 182             stripped = matcher.appendTail(stripped);
 183             field = stripped.toString();
 184         }
 185
 186         if (field.indexOf('(') != -1) {
 187             StringBuilder stripped = new StringBuilder(field.length());
 188             Matcher matcher = PARENTHESIZED.matcher(field);
 189             while (matcher.find()) {
 190                 parenthesized.append(matcher.group(1)).append(" ");
 191                 matcher.appendReplacement(stripped, " ");
 192             }
 193             stripped = matcher.appendTail(stripped);
 194             field = stripped.toString();
 195         }
 196
 197         field = field.trim();
 198
 199         // split words on non -A-z0-9, do them.
 200         final String[] tokens = NON_CHAR_DASH.split(field);
 201
 202         final EntryTypeName entryTypeName;
 203         if (numSubFields == 1) {
 204             assert subfieldIdx == 0;
 205             if (tokens.length == 1) {
 206                 entryTypeName = EntryTypeName.ONE_WORD;
 207             } else if (tokens.length == 2) {
 208                 entryTypeName = EntryTypeName.TWO_WORDS;
 209             } else if (tokens.length == 3) {
 210                 entryTypeName = EntryTypeName.THREE_WORDS;
 211             } else if (tokens.length == 4) {
 212                 entryTypeName = EntryTypeName.FOUR_WORDS;
 213             } else {
 214                 entryTypeName = EntryTypeName.FIVE_OR_MORE_WORDS;
 215             }
 216         } else {
 217             assert numSubFields > 1;
 218             if (subfieldIdx == 0) {
 219                 if (tokens.length == 1) {
 220                     entryTypeName = EntryTypeName.MULTIROW_HEAD_ONE_WORD;
 221                 } else {
 222                     entryTypeName = EntryTypeName.MULTIROW_HEAD_MANY_WORDS;
 223                 }
 224             } else {
 225                 assert subfieldIdx > 0;
 226                 if (tokens.length == 1) {
 227                     entryTypeName = EntryTypeName.MULTIROW_TAIL_ONE_WORD;
 228                 } else {
 229                     entryTypeName = EntryTypeName.MULTIROW_TAIL_MANY_WORDS;
 230                 }
 231             }
 232         }
 233
 234         for (String token : tokens) {
 235             token = TRIM_PUNC.matcher(token).replaceAll("");
 236             if (/*!alreadyDone.contains(token) && */!token.isEmpty()) {
 237                 indexBuilder.addEntryWithTokens(entryData, Collections.singleton(token), entryTypeName);
 238                 // alreadyDone.add(token);
 239
 240                 // also split words on dashes, do them, too.
 241                 if (token.indexOf('-') != -1) {
 242                     final String[] dashed = token.split("-");
 243                     for (final String dashedToken : dashed) {
 244                         if (/*!alreadyDone.contains(dashedToken) && */!dashedToken.isEmpty()) {
 245                             indexBuilder.addEntryWithTokens(entryData, Collections.singleton(dashedToken), EntryTypeName.PART_OF_HYPHENATED);
 246                         }
 247                     }
 248                 }
 249
 250             }  // if (!alreadyDone.contains(token)) {
 251         }  // for (final String token : tokens) {
 252
 253         // process bracketed stuff (split on spaces and dashes always)
 254         if (bracketed.length() > 0) {
 255             final String[] bracketedTokens = NON_CHAR.split(bracketed.toString());
 256             for (final String token : bracketedTokens) {
 257                 assert token.indexOf("-") == -1;
 258                 if (/*!alreadyDone.contains(token) && */!token.isEmpty()) {
 259                     indexBuilder.addEntryWithTokens(entryData, Collections.singleton(token), EntryTypeName.BRACKETED);
 260                 }
 261             }
 262         }
 263
 264         // process paren stuff
 265         if (parenthesized.length() > 0) {
 266             final String[] parenTokens = NON_CHAR.split(parenthesized.toString());
 267             for (final String token : parenTokens) {
 268                 assert token.indexOf("-") == -1;
 269                 if (/*!alreadyDone.contains(token) && */!token.isEmpty()) {
 270                     indexBuilder.addEntryWithTokens(entryData, Collections.singleton(token), EntryTypeName.PARENTHESIZED);
 271                 }
 272             }
 273         }
 274
 275     }
 276
 277     private String parseField_DE(final IndexBuilder indexBuilder, String field,
 278                                  final IndexedEntry entryData, final int subfieldIdx) {
 279
 280 //    final Matcher matcher = DE_NOUN.matcher(field);
 281 //    while (matcher.find()) {
 282 //      final String noun = matcher.group(1);
 283         //final String gender = matcher.group(2);
 284 //      if (alreadyDone.add(noun)) {
 285         // System.out.println("Found DE noun " + noun + ", " + gender);
 286 //        final List<EntryData> entries = indexBuilder.getOrCreateEntries(noun, EntryTypeName.NOUN);
 287 //        entries.add(entryData);
 288 //      }
 289 //    }
 290
 291         // In English, curly braces are used for different tenses.
 292         field = CURLY_BRACED.matcher(field).replaceAll(" ");
 293
 294         return field;
 295     }
 296
 297     private String parseField_EN(final IndexBuilder indexBuilder, String field,
 298                                  final IndexedEntry entryData, final int subfieldIdx) {
 299         if (field.startsWith("to ")) {
 300             field = field.substring(3);
 301         }
 302         return field;
 303     }
 304
 305     public static Set<String> tokenize(final String text, final Pattern pattern) {
 306         final String[] split = pattern.split(text);
 307         final Set<String> result = new LinkedHashSet<>(Arrays.asList(split));
 308         result.remove("");
 309         return result;
 310     }
 311
 312
 313 }