Work on wordlists

Dictionaries/en_US_se.xml

@@ -191,6 +191,7 @@
   <word>cyanotic</word>
   <word>darmstadtium</word>
   <word>darndest</word>
+  <word>dealmaking</word>
   <word>debride</word>
   <word>debridement</word>
   <word>decompensating</word>
@@ -418,6 +419,7 @@
   <word>mycelial</word>
   <word>nah</word>
   <word>namaste</word>
+  <word>narrowminded</word>
   <word>nasties</word>
   <word>neighbour</word>
   <word>neighbourhood</word>
@@ -499,6 +501,7 @@
   <word>polenta</word>
   <word>polonium</word>
   <word>polysulfide</word>
+  <word>postsurgical</word>
   <word>postulator</word>
   <word>potassium</word>
   <word>praseodymium</word>
@@ -511,6 +514,7 @@
   <word>pseudoachondroplasia</word>
   <word>pupillary</word>
   <word>purée</word>
+  <word>pushback</word>
   <word>radium</word>
   <word>radon</word>
   <word>ragdoll</word>
@@ -576,6 +580,7 @@
   <word>slipspace</word>
   <word>smartphone</word>
   <word>smartphones</word>
+  <word>snakelet</word>
   <word>snuck</word>
   <word>sociopathic</word>
   <word>sodium</word>
@@ -603,6 +608,7 @@
   <word>stepmom</word>
   <word>stereotactic</word>
   <word>sternotomy</word>
+  <word>storyboarded</word> 
   <word>strontium</word>
   <word>subclavian</word>
   <word>subdural hematoma</word>
@@ -708,6 +714,7 @@
   <word>where'd</word>
   <word>where're</word>
   <word>which</word>
+  <word>whodunit</word>
   <word>why'd</word>
   <word>why's</word>
   <word>wizarding</word>

Dictionaries/en_names.xml

@@ -1437,6 +1437,7 @@
   <name>Rafferty</name>
   <name>Raiden</name>
   <name>Raina</name>
+  <name>Rajan</name>
   <name>Ramiro</name>
   <name>Rashad</name>
   <name>Rayan</name>

src/libse/Dictionaries/StringWithoutSpaceSplitToWords.cs

@@ -1,5 +1,6 @@
 using System;
 using System.Collections.Generic;
+using System.IO;
 using System.Linq;
 using Nikse.SubtitleEdit.Core.Common;
 
@@ -38,7 +39,13 @@ namespace Nikse.SubtitleEdit.Core.Dictionaries
             var s = input;
             var check = s;
             var spaces = new List<int>();
-            for (int i = 0; i < words.Length; i++)
+
+            if (words.Contains(input))
+            {
+                return input;
+            }
+
+            for (var i = 0; i < words.Length; i++)
             {
                 var w = words[i];
                 if (w.Length >= input.Length)
@@ -77,5 +84,32 @@ namespace Nikse.SubtitleEdit.Core.Dictionaries
 
             return s.Trim();
         }
+
+        public static string[] LoadWordSplitList(string threeLetterIsoLanguageName, NameList nameList)
+        {
+            var fileName = $"{Configuration.DictionariesDirectory}{threeLetterIsoLanguageName}_WordSplitList.txt";
+            if (!File.Exists(fileName))
+            {
+                return Array.Empty<string>();
+            }
+
+            var wordList = File.ReadAllText(fileName).SplitToLines().Where(p => p.Trim().Length > 0).ToList();
+
+            if (threeLetterIsoLanguageName == "eng")
+            {
+                wordList.AddRange(new List<string>
+                {
+                    // Ignore list
+                    "Andor", "honour", "putain", "whoah", "eastside", "Starpath", "comlink"
+                });
+            }
+
+            if (nameList != null)
+            {
+                wordList.AddRange(nameList.GetNames().Where(p => p.Length > 4));
+            }
+
+            return wordList.OrderByDescending(p => p.Length).ToArray();
+        }
     }
 }

src/ui/Forms/SpellCheck.cs

@@ -340,14 +340,7 @@ namespace Nikse.SubtitleEdit.Forms
 
             _wordSplitListLanguage = languageName;
             var threeLetterIsoLanguageName = Iso639Dash2LanguageCode.GetThreeLetterCodeFromTwoLetterCode(twoLetterLanguageName);
-            var fileName = $"{Configuration.DictionariesDirectory}{threeLetterIsoLanguageName}_WordSplitList.txt";
-            if (!File.Exists(fileName))
-            {
-                return Array.Empty<string>();
-            }
-
-            var wordList = File.ReadAllText(fileName).SplitToLines().Where(p => p.Trim().Length > 0).ToList();
-            return wordList.ToArray();
+            return StringWithoutSpaceSplitToWords.LoadWordSplitList(threeLetterIsoLanguageName, null);
         }
 
         private void FillSpellCheckDictionaries(string languageName)

src/ui/Logic/Ocr/OcrFixEngine.cs

@@ -313,7 +313,7 @@ namespace Nikse.SubtitleEdit.Logic.Ocr
             _nameListWithApostrophe = new HashSet<string>();
             var nameListWithPeriods = new List<string>();
             _abbreviationList = new HashSet<string>();
-            _wordSplitList = LoadWordSplitList(threeLetterIsoLanguageName, _nameListObj);
+            _wordSplitList = StringWithoutSpaceSplitToWords.LoadWordSplitList(threeLetterIsoLanguageName, _nameListObj);
 
             var isEnglish = threeLetterIsoLanguageName.Equals("eng", StringComparison.OrdinalIgnoreCase);
             foreach (var name in _nameList)
@@ -415,19 +415,6 @@ namespace Nikse.SubtitleEdit.Logic.Ocr
             }
         }
 
-        private static string[] LoadWordSplitList(string threeLetterIsoLanguageName, NameList nameList)
-        {
-            var fileName = $"{Configuration.DictionariesDirectory}{threeLetterIsoLanguageName}_WordSplitList.txt";
-            if (!File.Exists(fileName))
-            {
-                return Array.Empty<string>();
-            }
-
-            var wordList = File.ReadAllText(fileName).SplitToLines().Where(p => p.Trim().Length > 0).ToList();
-            wordList.AddRange(nameList.GetNames().Where(p => p.Length > 4));
-            return wordList.OrderByDescending(p => p.Length).ToArray();
-        }
-
         public string SpellCheckDictionaryName
         {
             get