Import last change for jmdict-0.9

[jmdict-cli.git] / kana2romaji.cpp

diff --git a/kana2romaji.cpp b/kana2romaji.cpp
index 2ba77fdb984d1a1f58c5bbbf0d0aab9716c3940d..8508f764c1794ad46e27605236bac4f83427635d 100644 (file)
--- a/kana2romaji.cpp
+++ b/kana2romaji.cpp
@@ -27,15 +27,15 @@ using namespace std;
 
 namespace  {
 void utfchar(const string& from, string::size_type pos, string& to) {
-    string::value_type first = from[pos];
+    unsigned first = from[pos];
     if ((first & 0x80) == 0)
         to = from[pos];
     else {
         string::size_type len = 0;
         while (first & 0x80) {
             ++len;
-           first <<= 1;
-       }
+            first <<= 1;
+        }
         to = from.substr(pos, len);
     }
 }
@@ -218,7 +218,7 @@ void initRomaji() {
     romaji["プ"] = "pu";
     romaji["ペ"] = "pe";
     romaji["ポ"] = "po";
-    romaji["ã\83¼"] = "";
+    romaji["ã\83»"] = " ";
     
     // -- double width letters ------
     romaji["Ａ"] = "A";
@@ -292,9 +292,7 @@ void initRomaji() {
     romaji["＄"] = "$";
     romaji["％"] = "%";
     romaji["＆"] = "&";
-    romaji["＇"] = "'"; // TODO:
-    romaji["（"] = "(";
-    romaji["）"] = ")";
+    romaji["＇"] = "'";
     romaji["＊"] = "*";
     romaji["＋"] = "+";
     romaji["，"] = ",";
@@ -310,39 +308,60 @@ void initRomaji() {
     romaji["？"] = "?";
     romaji["＠"] = "@";
 
+    romaji["〔"] = "(";
+    romaji["〕"] = ")";
+    romaji["（"] = "(";
+    romaji["）"] = ")";
     romaji["［"] = "[";
-    romaji["＼"] = "\\";
     romaji["］"] = "]";
+    romaji["【"] = "[";
+    romaji["】"] = "]";
+    romaji["｛"] = "{";
+    romaji["｝"] = "}";
+    romaji["＼"] = "\\";
     romaji["＾"] = "^";
     romaji["＿"] = "_";
     romaji["｀"] = "`";
-
-    romaji["｛"] = "{";
     romaji["｜"] = "|";
-    romaji["｝"] = "}";
     romaji["～"] = "~";
-
-
-    // don't know where those belong to
+    romaji["ー"] = "";
+    romaji["。"] = ".";
+    romaji["、"] = ",";
     romaji["〜"] = "~";
-    romaji["、"] = ","; // TODO:
     romaji["−"] = "-";
-    
-    romaji["　"] = " ";
     romaji["―"] = "-";
-    romaji["・"] = "-"; // FIXME
+    romaji["　"] = " ";
+
+    romaji["ー"] = "\3";
+}
+
+void remove_quote_1(
+  string::size_type const pos,
+  string &rom)
+{
+  // if we encounter something like
+  // "ki" + '\1' 
+  // remove the previous character of \1 and the \1
+  // if \1 is followed by an 'y' remove that also
+  rom.erase(
+    pos - 1,
+    (pos + 1 < rom.size()
+     && rom[pos + 1] == 'y'
+    )
+      ? 3
+      : 2);
 }
 
 void kana2romaji(const string& kana, string& rom) {
     rom.clear();
     for (string::size_type pos = 0; pos < kana.size(); ) {
-       string ch;
+        string ch;
         utfchar(kana, pos, ch);
         romaji_map::const_iterator trans = romaji.find(ch);
         if (trans == romaji.end()) {
             rom += ch;
             if (ch.size() > 1)
-                cout << "Don't know how to translate '" << ch << "' in '" << kana << "' to romaji." << endl;
+                cout << "Don't know how to translate '" << ch << "' in '" << kana << "' to romaji.\n";
         }
         else
             rom += trans->second;
@@ -350,13 +369,142 @@ void kana2romaji(const string& kana, string& rom) {
     }
     for (string::size_type pos = 0; pos < rom.size(); ++pos)
         if (rom[pos] == '\1') {
-               string::size_type from = pos, count = 1;
-            if (pos > 1 && (rom[pos - 2] == 'h' || rom[pos - 2] == 'j')) {
-                --from;
-                count = (pos + 1 < rom.size() && rom[pos + 1] == 'y') ? 3 : 2;
+
+            if (pos > 2) {
+               string const pred = rom.substr(pos - 3, 3);
+               if(pred == "chi" ||
+                  pred == "shi" ||
+                  pred == "dzi" ||
+                  pred == "tsu" ||
+                  pred == "shi"
+               ) {
+                 remove_quote_1(pos, rom);
+                 pos -= 2;
+                 continue;
+               }
+            }
+            if (pos > 1) {
+             
+               string const pred = rom.substr(pos - 2, 2);
+               if(pred == "ki" ||
+                  pred == "ni" ||
+                  pred == "mi" ||
+                  pred == "ri" ||
+                  pred == "gi" ||
+                  pred == "ji" ||
+                  pred == "hi" ||
+                  pred == "bi" ||
+                  pred == "pi"
+                )
+               {
+                 // shorten "ji\1y" to "j"
+                 // otherwise remove "\1" and the preceding character
+                 // but not the y
+                 rom.erase(
+                   pos - 1,
+                   (pos + 1 < rom.size()
+                    && rom[pos + 1] == 'y'
+                    && pred[0] == 'j')
+                      ? 3
+                      : 2);
+                  pos -= 2;
+                 continue;
+               }
+               else if(
+                  pred == "fu" ||
+                  pred == "de" ||
+                  pred == "te" ||
+                  pred == "vu")
+               {
+                  remove_quote_1(pos, rom);
+                  pos -= 2;
+                 continue;
+               }
+               else if(
+                 pred == "su" ||
+                 pred == "zu" ||
+                 pred == "te" ||
+                 pred == "de" ||
+                 pred == "ku" ||
+                 pred == "gu" ||
+                 pred == "mu"
+               )
+               {
+                 rom[pos - 1] = 'w'; 
+                 rom.erase(pos, 1);
+                 --pos;
+                 continue;
+               }
+               else if(
+                 pred == "to" ||
+                 pred == "do"
+               )
+               {
+                 rom[pos - 1] = 'h';
+                 rom.erase(pos, 1);
+                 --pos;
+                 continue;
+               }
+            }
+
+            if (pos > 0) {
+              char const pred = rom[pos - 1];
+
+              switch(pred)
+              {
+              case 'a':
+              case 'i':
+              case 'u':
+              case 'e':
+              case 'o':
+                rom.erase(pos, 1);
+                --pos;
+                break;
+              default:
+                cout << "Encountered a special character in " << kana << " but don't know what to do with it.\n";
+              }
+            }
+            else
+            {
+              rom.erase(pos, 1);
+              --pos;
             }
-            rom.erase(from, count);
         }
-        else if (rom[pos] == '\2' && pos + 1 < rom.size())
+        else if (rom[pos] == '\2')
+        {
+         // two tsu may follow each other, so just remove them
+          if(pos + 1 < rom.size() && rom[pos + 1] != '\2')
             rom[pos] = rom[pos + 1];
+          else
+          {
+            rom.erase(pos, 1);
+            --pos;
+          }
+        }
+        else if (rom[pos] == '\3')
+        {
+          if(pos == 0)
+          {
+            if(rom.size() == 1)
+              rom = "chouon";
+            else
+            {
+              cout << "ー is the first letter of " << kana << ". Don't know how to translate this.\n";
+              rom.erase(pos, 1);
+              --pos;
+            }
+          }
+          else
+            rom[pos] = rom[pos-1];
+        }
+
+  for (string::size_type pos = 0; pos < rom.size(); ++pos)
+    switch(rom[pos])
+    {
+    case '\1':
+    case '\2':
+    case '\3':
+      cout << "Failed to translate " << kana << '\n';
+      return;
+    }
 }