[GitHub] [lucene-solr] dweiss commented on a change in pull request #2320: LUCENE-9742: Hunspell: suggest dictionary entries similar to the misspelled word

[GitBox <gi...@apache.org>]

dweiss commented on a change in pull request #2320:
URL: https://github.com/apache/lucene-solr/pull/2320#discussion_r572349453



##########
File path: lucene/analysis/common/src/java/org/apache/lucene/analysis/hunspell/GeneratingSuggester.java
##########
@@ -0,0 +1,321 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements.  See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License.  You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+package org.apache.lucene.analysis.hunspell;
+
+import java.io.IOException;
+import java.util.ArrayList;
+import java.util.EnumSet;
+import java.util.List;
+import java.util.Objects;
+import java.util.Set;
+import java.util.TreeSet;
+import java.util.stream.Collectors;
+import org.apache.lucene.util.IntsRef;
+import org.apache.lucene.util.fst.IntsRefFSTEnum;
+
+/**
+ * A class that traverses the entire dictionary and applies affix rules to check if those yield
+ * correct suggestions similar enough to the given misspelled word
+ */
+class GeneratingSuggester {
+  private static final int MAX_ROOTS = 100;
+  private static final int MAX_GUESSES = 100;
+  private final Dictionary dictionary;
+
+  GeneratingSuggester(Dictionary dictionary) {
+    this.dictionary = dictionary;
+  }
+
+  List<String> suggest(String word, WordCase originalCase, Set<String> prevSuggestions) {
+    List<WeightedWord> roots = findSimilarDictionaryEntries(word, originalCase);
+    List<WeightedWord> expanded = expandRoots(word, roots);
+    TreeSet<WeightedWord> bySimilarity = rankBySimilarity(word, expanded);
+    return getMostRelevantSuggestions(bySimilarity, prevSuggestions);
+  }
+
+  private List<WeightedWord> findSimilarDictionaryEntries(String word, WordCase originalCase) {
+    try {
+      IntsRefFSTEnum<IntsRef> fstEnum = new IntsRefFSTEnum<>(dictionary.words);
+      TreeSet<WeightedWord> roots = new TreeSet<>();
+
+      IntsRefFSTEnum.InputOutput<IntsRef> mapping;
+      while ((mapping = fstEnum.next()) != null) {
+        IntsRef key = mapping.input;
+        if (Math.abs(key.length - word.length()) > 4 || !isSuitableRoot(mapping.output)) continue;
+
+        String root = toString(key);
+        if (originalCase == WordCase.LOWER
+            && WordCase.caseOf(root) == WordCase.TITLE
+            && !dictionary.hasLanguage("de")) {
+          continue;
+        }
+
+        String lower = dictionary.toLowerCase(root);
+        int sc =
+            ngram(3, word, lower, EnumSet.of(NGramOptions.LONGER_WORSE)) + commonPrefix(word, root);
+
+        roots.add(new WeightedWord(root, sc));
+      }
+      return roots.stream().limit(MAX_ROOTS).collect(Collectors.toList());
+    } catch (IOException e) {
+      throw new RuntimeException(e);
+    }
+  }
+
+  private static String toString(IntsRef key) {
+    char[] chars = new char[key.length];
+    for (int i = 0; i < key.length; i++) {
+      chars[i] = (char) key.ints[i + key.offset];
+    }
+    return new String(chars);
+  }
+
+  private boolean isSuitableRoot(IntsRef forms) {
+    for (int i = 0; i < forms.length; i += dictionary.formStep()) {
+      int entryId = forms.ints[forms.offset + i];
+      if (dictionary.hasFlag(entryId, dictionary.needaffix)
+          || dictionary.hasFlag(entryId, dictionary.forbiddenword)
+          || dictionary.hasFlag(entryId, Dictionary.HIDDEN_FLAG)
+          || dictionary.hasFlag(entryId, dictionary.onlyincompound)) {
+        continue;
+      }
+      return true;
+    }
+
+    return false;
+  }
+
+  private List<WeightedWord> expandRoots(String word, List<WeightedWord> roots) {
+    int thresh = calcThreshold(word);
+
+    TreeSet<WeightedWord> expanded = new TreeSet<>();
+    for (WeightedWord weighted : roots) {
+      String guess = weighted.word;
+      String lower = dictionary.toLowerCase(guess);
+      int sc =
+          ngram(word.length(), word, lower, EnumSet.of(NGramOptions.ANY_MISMATCH))
+              + commonPrefix(word, guess);
+      if (sc > thresh) {
+        expanded.add(new WeightedWord(guess, sc));
+      }
+    }
+    return expanded.stream().limit(MAX_GUESSES).collect(Collectors.toList());
+  }
+
+  // find minimum threshold for a passable suggestion
+  // mangle original word three different ways
+  // and score them to generate a minimum acceptable score
+  private static int calcThreshold(String word) {
+    int thresh = 0;
+    for (int sp = 1; sp < 4; sp++) {
+      char[] mw = word.toCharArray();
+      for (int k = sp; k < word.length(); k += 4) {
+        mw[k] = '*';
+      }
+
+      thresh += ngram(word.length(), word, new String(mw), EnumSet.of(NGramOptions.ANY_MISMATCH));
+    }
+    return thresh / 3 - 1;
+  }
+
+  private TreeSet<WeightedWord> rankBySimilarity(String word, List<WeightedWord> expanded) {
+    double fact = (10.0 - dictionary.maxDiff) / 5.0;
+    TreeSet<WeightedWord> bySimilarity = new TreeSet<>();
+    for (WeightedWord weighted : expanded) {
+      String guess = weighted.word;
+      String lower = dictionary.toLowerCase(guess);
+      if (lower.equals(word)) {
+        bySimilarity.add(new WeightedWord(guess, weighted.score + 2000));
+        break;
+      }
+
+      int re =
+          ngram(2, word, lower, EnumSet.of(NGramOptions.ANY_MISMATCH, NGramOptions.WEIGHTED))
+              + ngram(2, lower, word, EnumSet.of(NGramOptions.ANY_MISMATCH, NGramOptions.WEIGHTED));
+
+      int score =
+          2 * lcs(word, lower)
+              - Math.abs(word.length() - lower.length())
+              + commonCharacterPositionScore(word, lower)
+              + commonPrefix(word, lower)
+              + ngram(4, word, lower, EnumSet.of(NGramOptions.ANY_MISMATCH))
+              + re
+              + (re < (word.length() + lower.length()) * fact ? -1000 : 0);
+      bySimilarity.add(new WeightedWord(guess, score));
+    }
+    return bySimilarity;
+  }
+
+  private List<String> getMostRelevantSuggestions(
+      TreeSet<WeightedWord> bySimilarity, Set<String> prevSuggestions) {
+    List<String> result = new ArrayList<>();
+    boolean hasExcellent = false;
+    for (WeightedWord weighted : bySimilarity) {
+      if (weighted.score > 1000) {
+        hasExcellent = true;
+      } else if (hasExcellent) {
+        break; // leave only excellent suggestions, if any
+      }
+
+      boolean bad = weighted.score < -100;
+      // keep the best ngram suggestions, unless in ONLYMAXDIFF mode
+      if (bad && (!result.isEmpty() || dictionary.onlyMaxDiff)) {
+        break;
+      }
+
+      if (prevSuggestions.stream().noneMatch(weighted.word::contains)
+          && result.stream().noneMatch(weighted.word::contains)) {
+        result.add(weighted.word);
+        if (result.size() > dictionary.maxNGramSuggestions) {
+          break;
+        }
+      }
+
+      if (bad) {
+        break;
+      }
+    }
+    return result;
+  }
+
+  private static int commonPrefix(String s1, String s2) {

Review comment:
       I don't know of any.




----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@lucene.apache.org
For additional commands, e-mail: issues-help@lucene.apache.org