After a little work, I've finally ironed this out... Though could I get someone to go over this and see if there is a better way to present this library within the code, before I publish it?
Premise: Kimchi (a Korean transliteration library)
Kimchi, apart from being a delicious Korean side dish is also a Korean transliteration library that takes romanized input and transliterates (not translates) said text into the equivalent Hangeul character syllable blocks.
There are currently two commands, kimchi.all() and kimchi.convert(str); kimchi.convert(str) will convert the str string variable into the corresponding Korean syllable blocks. kimchi.all() will go through the DOM and apply kimchi.convert() on the innerHTML component of each <kimchi> tag. Basically treat <kimchi> tags the same way as you would a <span> tag.
The library (save as kimchi.js):
var kimchi = {
convert: function(fromText) {
// Internal search function, avoids issues with non existent Array.indexOf() functionality in certain versions
var inArr=function(a,s){for(var x=-1,y=0;y<a.length;y++)if(a[y]==s){x=y;break}return x};
// Lower case the input text, build LVT arrays and init output text
fromText = fromText.toLowerCase(),
aLead = "g0kk0n0d0tt0r0m0b0pp0s0ss0-0j0jj0ch0k0t0p0h".split(0),
aVowel = "a0ae0ya0yae0eo0e0yeo0ye0o0wa0wae0oe0yo0u0wo0we0wi0yu0eu0ui0i".split(0),
aTail = "g0gg0gs0n0nj0nh0d0l0lg0lm0lb0ls0lt0lp0lh0m0b0bs0s0ss0ng0j0c0k0t0p0h".split(0),
toText = "";
// Split into Alpha/Dash and non-Alpha/Dash arrays
// Then it's simply a matter of flipping between the two arrays and displaying accordingly.
aKorean = fromText.split(/[^a-z\-]+/) || [],
aNotKorean = fromText.split(/[a-z\-]+/) || [],
iType = 3,
iPhase = 0;
// Loop through each Korean word.
for (var iLoop = 0; iLoop < aKorean.length; iLoop++) {
sWord = aKorean[iLoop];
// Determine if first letter is vowel, if so latch a dash on its front.
sWord = (~"aeiouwy".indexOf(sWord[0]) ? "-" : "") + sWord;
// Loop through in reverse and extract letter blocks.
for (var sLetterExt = "", bFoundLetter = !1, sLetterBlock = "", iLoop2 = sWord.length - 1; - 1 < iLoop2; iLoop2--)
for (var iLocationLast = iLoop2, iLen = 1; !bFoundLetter;) {
// Attempt to find letter in chosen array
// This is important, as there is no chance that a letter will be missed...
// At least I hope not...
sExtract = sWord.substr(iLocationLast - iLen + 1, iLen);
iLocation = inArr([aLead, aVowel, aTail][iType - 1], sExtract);
if (0 > iLocation && "" == sLetterExt) {
// If the letter isn't found, move to next logical array,
// move cursor back and break out of current loop for restart.
iType = 1 > iType - 1 ? 3 : iType - 1;
iLoop2++;
break
}
if (0 > iLocation && "" == sLetterExt && 1 == iPhase) {
// If the letter found is a - and the previous letter found is a lead,
// ignore it as this is plainly a divisionary - and not a mute.
// Move to tail array, move cursor forward and break out of
// current loop for restart.
iType = 3;
break
} else if (0 > iLocation && "" != sLetterExt && 0 == iLocationLast && 1 == iLen && 1 == iType) {
// It's gone as far as it can, and it can't go any further...
// Flag letter found and assign letter type, then relocate cursor.
bLetterFound = !0;
sLetterBlock = (1 == iType ? " " : "") + (1 < iType ? "/" : "") + sLetterExt + sLetterBlock;
iLoop2 = iLocationLast - iLen + 1;
sLetterExt = "";
break
} else if (0 > iLocation && "" != sLetterExt)
// It's gone as far as it can, and it can't go any further...
// Flag letter found and assign letter type, then relocate cursor.
bLetterFound = !0,
iType == iPhase && "" != sLetterBlock && (sLetterBlock = " -" + sLetterBlock),
sLetterBlock = (1 == iType ? " " : "") + (1 < iType ? "/" : "") + sLetterExt + sLetterBlock,
iLoop2 = iLocationLast - iLen + 1,
sLetterExt = "",
iPhase = iType;
else if (sLetterExt = sExtract, 0 == iLocationLast) {
bLetterFound = !0;
sLetterBlock = sExtract + sLetterBlock;
iType = 3;
break
} else iLen++
}
sKoreanConv = "";
aConvKorean = sLetterBlock.split(" ");
// Determine the Unicode character value for the syllable block...
for (var a = 0; a < aConvKorean.length; a++)
"" != aConvKorean[a] && (aSyllable = aConvKorean[a].split("/"), sKoreanConv += "&#" + ((2 == aSyllable.length ? 0 : inArr(aTail, aSyllable[2]) + 1) + 28 * inArr(aVowel, aSyllable[1]) + 588 * inArr(aLead, aSyllable[0]) + 44032) + ";");
aKorean[iLoop] = sKoreanConv;
sKoreanConv = "";
}
bNotKoreanStart = aNotKorean ? 0 == fromText.indexOf(aNotKorean[0]) ? !0 : !1 : !1;
iCounter = 0;
if (bNotKoreanStart && aNotKorean.length == aKorean.length)
for (var iDisp = 0; iDisp < aNotKorean.length; iDisp++) toText += aNotKorean[iDisp] + aKorean[iDisp];
else if (!bNotKoreanStart && aNotKorean.length == aKorean.length)
for (iDisp = 0; iDisp < aNotKorean.length; iDisp++) toText += aKorean[iDisp] + aNotKorean[iDisp];
else if (bNotKoreanStart) {
if (toText += aNotKorean[0], 1 < aNotKorean.length)
for (iDisp = 0; iDisp < aKorean.length; iDisp++) toText += aKorean[iDisp] + aNotKorean[iDisp + 1]
} else if (!bNotKoreanStart && (toText += aKorean[0], 1 < aKorean.length))
for (iDisp = 0; iDisp < aNotKorean.length; iDisp++) toText += aNotKorean[iDisp] + aKorean[iDisp + 1];
return toText;
},
all: function() {
// Affects all <kimchi> tags...
var z = document.getElementsByTagName('kimchi');
for (x = z.length; x--;) {
z[x].innerHTML = kimchi.convert(z[x].childNodes[0].data);
}
}
}
Usage:
<html>
<head>
<meta charset="utf-8" />
</head>
<script src="kimchi.js"></script>
<body onLoad="kimchi.all();">
<kimchi>hangug-eo, radio, seoul, pyeong-yang, kimchi</kimchi> - some random words in Korean<br/>
<kimchi>Kim Yu Suk</kimchi>: the name of a genuine Korean Olympic high jumper<br />
<span id="test"></span>
<script>
document.getElementById("test").innerHTML = kimchi.convert("pyeong-yang");
</script>
</body>
</html>
