Merge pull request #114 from jaumeortola/run-on-camelcase

improve run-on suggestions for camel case words
morfologik · Apr 20, 2022 · fcaa33e · fcaa33e
2 parents 3364cf6 + dbc11c3
commit fcaa33e
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 1 deletion.
diff --git a/morfologik-speller/src/main/java/morfologik/speller/Speller.java b/morfologik-speller/src/main/java/morfologik/speller/Speller.java
@@ -335,7 +335,9 @@ public List<CandidateData> replaceRunOnWordCandidates(final String original) {
         // chop from left to right
         final String prefix = wordToCheck.substring(0, i);
         final String suffix = wordToCheck.substring(i);
-        if (isInDictionary(suffix)) {
+        if (isInDictionary(suffix) 
+            // camel case words: e.g. GreatElephant
+            || (!isNotCapitalizedWord(suffix) && isInDictionary(suffix.toLowerCase(locale)))) {
           if (isInDictionary(prefix)) {
             addReplacement(candidates, prefix + " " + suffix);
           } else if (Character.isUpperCase(prefix.charAt(0)) && isInDictionary(prefix.toLowerCase(locale))) {

diff --git a/morfologik-speller/src/test/java/morfologik/speller/SpellerTest.java b/morfologik-speller/src/test/java/morfologik/speller/SpellerTest.java
@@ -51,6 +51,10 @@ public void testRunonWords() throws IOException {
     final URL url2 = getClass().getResource("single-char-word.dict");
     final Speller spell2 = new Speller(Dictionary.read(url2));
     assertTrue(spell2.replaceRunOnWords("alot").contains("a lot"));
+    assertTrue(spell2.replaceRunOnWords("Alot").contains("A lot"));
+    assertTrue(spell2.replaceRunOnWords("ALot").contains("A Lot"));
+    assertTrue(spell2.replaceRunOnWords("LotAmusement").contains("Lot Amusement"));
+    //TODO? assertTrue(spell2.replaceRunOnWords("LOTAMUSEMENT").contains("LOT AMUSEMENT"));
     assertTrue(spell2.replaceRunOnWords("aalot").contains("aa lot"));
     assertTrue(spell2.replaceRunOnWords("aamusement").contains("a amusement"));
     assertTrue(spell2.replaceRunOnWords("clot").isEmpty());