Improve OCR replace list guessses

Qbort · Jun 14, 2020 · c78dda9 · c78dda9
1 parent a1c35e3
commit c78dda9
Show file tree

Hide file tree

Showing 4 changed files with 18 additions and 9 deletions.
diff --git a/Dictionaries/eng_OCRFixReplaceList.xml b/Dictionaries/eng_OCRFixReplaceList.xml
@@ -2820,7 +2820,7 @@
   </PartialWordsAlways>
   <PartialWords>
     <!-- Will be used to check words not in dictionary.
-          If new word(s) and longer than 5 chars and exists in spelling dictionary, it is (or they are) accepted -->
+          If new word(s) and longer than 4 chars and exists in spelling dictionary, it is (or they are) accepted -->
     <WordPart from="IVI" to="M" />
     <WordPart from="/" to="l" />
     <WordPart from="|" to="I" />

diff --git a/Dictionaries/pol_OCRFixReplaceList.xml b/Dictionaries/pol_OCRFixReplaceList.xml
@@ -4,7 +4,7 @@
   <PartialWordsAlways />
   <PartialWords>
     <!-- Will be used to check words not in dictionary.
-         If new word(s) and longer than 5 chars and exists
+         If new word(s) and longer than 4 chars and exists
          in spelling dictionary, it is (or they are) accepted -->
     <!-- "f " will be two words -->
     <WordPart from="f" to="f " />

diff --git a/libse/Dictionaries/OcrFixReplaceList.cs b/libse/Dictionaries/OcrFixReplaceList.cs
@@ -363,7 +363,7 @@ public IEnumerable<string> CreateGuessesFromLetters(string word)
                     if (word.Substring(i).StartsWith(letter, StringComparison.Ordinal))
                     {
 
-                        if (i == word.Length - letter.Length && !_partialWordReplaceList[letter].Contains(" "))
+                        if (i == word.Length - letter.Length && !_partialWordReplaceList[letter].Contains(' '))
                         {
                             var guess = word.Remove(i, letter.Length).Insert(i, _partialWordReplaceList[letter]);
                             AddToGuessList(list, guess);
@@ -379,14 +379,23 @@ public IEnumerable<string> CreateGuessesFromLetters(string word)
 
                 if (indexes.Count > 1)
                 {
-                    var multiGuess = word;
-                    for (int i = indexes.Count - 1; i >= 0; i--)
+                    if (!_partialWordReplaceList[letter].Contains(' '))
                     {
-                        var idx = indexes[i];
-                        multiGuess = multiGuess.Remove(idx, letter.Length).Insert(idx, _partialWordReplaceList[letter]);
-                        AddToGuessList(list, multiGuess);
+                        var multiGuess = word;
+                        for (int i = indexes.Count - 1; i >= 0; i--)
+                        {
+                            var idx = indexes[i];
+                            multiGuess = multiGuess.Remove(idx, letter.Length).Insert(idx, _partialWordReplaceList[letter]);
+                            AddToGuessList(list, multiGuess);
+                        }
+
+                        AddToGuessList(list, word.Replace(letter, _partialWordReplaceList[letter]));
                     }
                 }
+                else if (indexes.Count > 0)
+                {
+                    AddToGuessList(list, word.Replace(letter, _partialWordReplaceList[letter]));
+                }
 
                 if (indexes.Count > 0)
                 {

diff --git a/src/Logic/Ocr/OcrFixEngine.cs b/src/Logic/Ocr/OcrFixEngine.cs
@@ -1592,7 +1592,7 @@ public string FixUnknownWordsViaGuessOrPrompt(out int wordsNotFound, string line
                                 guesses.Add(wordWithVerticalLine);
                             }
 
-                            if (word.Length > 5 && autoGuess == AutoGuessLevel.Aggressive)
+                            if (word.Length > 4 && autoGuess == AutoGuessLevel.Aggressive)
                             {
                                 guesses.AddRange((List<string>)_ocrFixReplaceList.CreateGuessesFromLetters(word));