Add punctuation handling: implement isPunctuation and trimTrailingPunctuation functions

lib/Epub/Epub/hyphenation/HyphenationCommon.cpp

@@ -57,10 +57,54 @@ bool isCyrillicVowel(uint32_t cp) {
 
 bool isCyrillicConsonant(const uint32_t cp) { return isCyrillicLetter(cp) && !isCyrillicVowel(cp); }
 
-bool isAlphabetic(const uint32_t cp) { return isLatinLetter(cp) || isCyrillicLetter(cp); }
+bool isAlphabetic(const uint32_t cp) { return isLatinLetter(cp) || isCyrillicLetter(cp) || isPunctuation(cp); }
 
 bool isVowel(const uint32_t cp) { return isLatinVowel(cp) || isCyrillicVowel(cp); }
 
+bool isPunctuation(const uint32_t cp) {
+  switch (cp) {
+    case '.':
+    case ',':
+    case '!':
+    case '?':
+    case ';':
+    case ':':
+    case '"':
+    case '\'':
+    case ')':
+    case '(':
+    case '«':
+    case '»':
+    case '‘':
+    case '’':
+    case '“':
+    case '”':
+    case '—':
+    case '-':
+    case '–':
+    case '―':
+    case '[':
+    case ']':
+    case '{':
+    case '}':
+    case '/':
+    case 0x2019:  // ’
+    case 0x201D:  // ”
+    case 0x00BB:  // »
+    case 0x203A:  // ›
+    case 0x2026:  // …
+      return true;
+    default:
+      return false;
+  }
+}
+
+void trimTrailingPunctuation(std::vector<CodepointInfo>& cps) {
+  while (!cps.empty() && isPunctuation(cps.back().value)) {
+    cps.pop_back();
+  }
+}
+
 Script detectScript(const std::vector<CodepointInfo>& cps) {
   bool hasLatin = false;
   bool hasCyrillic = false;

lib/Epub/Epub/hyphenation/HyphenationCommon.h

@@ -27,5 +27,7 @@ bool isCyrillicConsonant(uint32_t cp);
 
 bool isAlphabetic(uint32_t cp);
 bool isVowel(uint32_t cp);
+bool isPunctuation(uint32_t cp);
+void trimTrailingPunctuation(std::vector<CodepointInfo>& cps);
 
 Script detectScript(const std::vector<CodepointInfo>& cps);

lib/Epub/Epub/hyphenation/Hyphenator.cpp

@@ -94,6 +94,7 @@ std::vector<size_t> Hyphenator::breakOffsets(const std::string& word, const bool
   }
 
   auto cps = collectCodepoints(word);
+  trimTrailingPunctuation(cps);
   if (cps.size() < MIN_PREFIX_CP + MIN_SUFFIX_CP) {
     return byteOffsets;
   }