Introduction
I was recently working on a Indian language natural speech engine. So it was important to have a Transliteration engine in place that can convert English characters to corresponding Hindi characters ( not translation) . So if I type "mera" it should should show "मेरा " and not trying to find meaning of it.
Interestingly Devnagari texts are UTF encoded text. We have an UTF encoding table which can help map English characters to be converted to Devanagari using UTF mapping. However my aim was much simpler approch: Something of sort of a lookup table. I searched for entire eternity and yet could not manage any suitable table or code snippet in either Java or C#. Of course Java script version of UTF-Ascii Devanagari mapping is available for free and one can easily convert the java scipt to Java/C# code with little effort.
However I really wanted the most inefficient way of doing it: Using a lookup table. You may have your arguments of such an unwise method, but many NLP guys will agree that creating a Hashtable with Hindi-English strings comes handy in several processings. A lookup table in my case is helpful to have speech files with the same name as English chunks. So once user types in Hindi, the part of English chunks are known. I can simply concat the speech files with the name to get a final audio out. For example if user types 'ख', it's English equivalent is 'kh'. So I can have a speech file by name 'kh.wav' which can be played when ख is to be pronounced.
Background
The fascination of a lookup table ( Hashtable) based transliteration leads to this work. It is not really something that is written to get a Noble Prize, but aims at helping countless of those starters who are wondering a simple way of doing it.
Using the code
My implementation is in Java. I have not used any Hashtable here to keep it as platform independent as possible. So that the basic code can be reused in different platforms. I have selected most common Hindi characters including Yuktakshars, the set amy not be a complete set but is most sufficient set while working with Hindi Transliteration. There are two tables: DevText and English and there are two methods, one for returning a Hindi character associated with group of English character and the other does just inverse.
public class HindiTyping
{
public static void SetSingleLetterRight()
{
for(int i=0;i<DevText.length;i++)
{
if(DevText[i].length()==1)
{
DevText[i]=DevText[i]+"a";
}
}
}
static String[] DevText={"अ","आ","इ","ई","ऒ","ओ","उ","ऊ","ऎ","ए","ऐ","व","वा","वि","वी","वॆ","वे","वॊ","वो","वु","वू","रा","रॆ","रे","रि","री","रॊ","रो","रु","रू","र्य","र्या","र्यि","र्यी","र्ये","र्यु","र्यू","र्यॊ","र्यो","त","ता","तॆ","ते","ति","ती","तॊ","तो","तु","तू","त्य","त्या","त्यॆ","त्ये","त्यु","त्यू","त्यॊ","त्यो","त्यि","त्यी","य","या","यॆ","ये","यि","यी",
"यु","यू","यॊ","यो","प","पा","पॆ","पे","पि","पी","पु","पू","पॊ","पो","प्य","प्या","प्यॆ","प्ये","प्यि","प्यी","प्यु","प्यू","प्यॊ","प्यो","प्रा","प्री","प्रि","प्रॆ","प्रे","प्रु","प्रू","प्रॊ","प्रो","प्रु","प्रू","सा","सि","सी","सु","सू","सॊ","सो","सॆ","से","स्र","स्रा","स्रि","स्री","स्रु","स्रू","स्रॊ","स्रो","स्रॆ","स्रे","स्य","स्या","स्यि","स्यी","स्यॆ","स्ये",
"स्यु","स्यू","स्यॊ","स्यो","द","दा","दि","दी","दॊ","दो","दु","दू","दॆ","दे","ध","धा","धि","धी","धॊ","धो","धु","धू","धॆ","धे","फ","फा","फि","फी","फॊ","फो","फु","फू","फॆ","फे","फ्र","फ्रा","फ्रि","फ्री","फ्रॊ","फ्रो","फ्रु","फ्रू","फ्रॆ","फ्रे","ग","गा","गि","गी","गॊ","गो","गु","गू","गॆ","गे","ग्य","ग्या","ग्यि","ग्यी","ग्यो","ग्यॊ",
"ग्यु","ग्यू","ग्यॆ","ग्ये","ग्रा","ग्रि","ग्री","ग्रॊ","ग्रो","ग्रु","ग्रू","ग्रॆ","ग्रे","घ","घा","घि","घी","घॊ","घो","घु","घू","घॆ","घे","ह","हा","हि","ही","हॊ","हो","हु","हू","हॆ","हे","ज","जा","जि","जी","जॊ","जो","जु","जू","जॆ","जे","ज्र","ज्रा","ज्रि","ज्री","ज्रॊ","ज्रो","ज्रु","ज्रू","ज्रॆ","ज्रे","ज्य","ज्या","ज्यि","ज्यी","ज्यॊ","ज्यो","ज्यु","ज्यू",
"ज्यॆ","ज्ये","झ","झा","झि","झी","झॊ","झो","झु","झू","झॆ","झे","क","का","कि","की","कॊ","को","कु","कू","कॆ","के","क्य","क्या","क्यि","क्यी","क्यॊ","क्यो","क्यु","क्यू","क्यॆ","क्ये","क्र","क्रा","क्रि","क्री","क्रॊ","क्रो","क्रु","क्रू","क्रॆ","क्रे","क्व","क्वा","क्वि","क्वी","क्वॊ","क्वो","क्वु","क्वू","क्वॆ","कवे","त्व","त्वा",
"त्वि","त्वी","त्वॊ","त्वो","त्वु","त्वू","त्वॆ","त्वे","र्व","र्वा","र्वि","र्वी","र्वु","र्वू","र्वॊ","र्वो","र्वॆ","र्वे","प्व","प्वा","प्वि","प्वी","प्वॊ","प्वो","प्वु","प्वू","प्वॆ","प्वे","स्व","स्वा","स्वि","स्वी","स्वॊ","स्वो","स्वु","स्वू","स्वॆ","स्वे","","फ्व","फ्वा","फ्वि","फ्वी","फ्वॊ","फ्वो","फ्वु","फ्वू","फ्वॆ","फ्वे","ग्व","ग्वा","ग्वि","ग्वी",
"ग्वॊ","ग्वो","ग्वु","ग्वू","ग्वॆ","ग्वे","ह्व","","ज्व","झा","ज्वि","ज्वी","ज्वॊ","ज्वो","ज्वु","ज्वू","ज्वॆ","ज्वे","ल्व","ल्वा","ल्वि","ल्वी","ल्वॊ","ल्वो","ल्वु","ल्वू","ल्वॆ","ल्वे","ल","ला","लि","ली","लॊ","लो","लु","लू","लॆ","ले","ल्र","ल्र","ल्रा","ल्रि","ल्री","ल्रॊ","ल्रो","ल्रु","ल्रू","ल्रॆ","ल्रे","ल्य","ल्या","ल्यि","ल्यी","ल्यॊ",
"ल्यो","ल्यु","ल्यू","ल्यॆ","ल्ये","ल्ह","ल्हा","ल्हि","ल्ही","ल्हॊ","ल्हो","ल्हु","ल्हू","ल्हॆ","ल्हे","क्श","क्शा","क्शि","क्शी","क्शॊ","क्शो","क्शु","क्शू","क्शॆ","क्शे","च","चा","चि","ची","चॊ","चो","चु","चू","चॆ","चे","छ","छा","छि","छी","छॊ","छो","छु","छू","छॆ","छे","छ्र्","छ्रा","छ्रि","छ्री","छ्रॊ","छ्रो","छ्रु","छ्रू","छ्रॆ",
"छ्रे","च्र्","च्रा","च्रि","च्री","च्रॊ","च्रो","च्रु","च्रू","च्रॆ","च्रे","च्य","च्या","च्यि","च्यी","च्यॊ","च्यो","च्यु","च्यू","च्यॆ","च्ये","छ्य","व","वा","वि","वी","वॊ","वो","वु","वू","वॆ","वे","व्य","व्या","व्यि","व्यी","व्यॊ","व्यो","व्यु","व्यू","व्यॆ","व्ये","व्र","व्रा","व्रि","व्री","व्रु","व्रू","व्रॊ","व्रो","व्रॆ","व्रे","व्ह","व्हा","व्हि","व्ही","व्हॊ",
"व्हो","व्हु","व्हू","व्हॆ","व्हे","ब","बा","बि","बी","बॊ","बो","बु","बू","बॆ","बे","भ","भा","भि","भी","भॊ","भो","भु","भू","भॆ","भे","भ्र","भ्रा","भ्रि","भ्री","भ्रॊ","भ्रो","भ्रु","भ्रू","भ्रॆ","भ्रे","ब्र","ब्रा","ब्रि","ब्री","ब्रॊ","ब्रो","ब्रु","ब्रू","ब्रॆ","ब्रे","न","ना","नि","नी","नु","नू","नॊ","नो","नॆ","ने","न्ह","न्हा","न्हि","न्ही","न्हॊ","न्हो","न्हु","न्हू",
"न्हॆ","न्हे","न्य","न्या","न्यु","न्यू","न्यि","न्यी","न्यॊ","न्यो","न्यॆ","न्ये","न्व","न्वा","न्वि","न्वी","न्वॊ","न्वो","न्वु","न्वू","न्वॆ","न्वे","ञ","ञा","ञि","ञी","ञॊ","ञो","ञु","ञू","ञॆ","ञे","ङ","ङा","ङि","ङी","ङॊ","ङो","ङु","ङू","ङॆ","ङे","मा","मि","मी","मॊ","मो","मु","मू","मॆ","मे","म्य","म्या","म्यि","म्यी","म्यॊ",
"म्यो","म्यु","म्यू","म्यॆ","म्ये","म्र","म्रा","म्रॊ","म्रो","म्रि","म्री","म्रु","म्रू","म्रॆ","म्रे","म्व","म्व","म्वा","म्वि","म्वी","म्वॊ","म्वो","म्वु","म्वू","म्वॆ","म्वे","है"};
static String [] English={
"a","aa","i","ii","o","o","u","uu","e","e","ai","v","vaa","vi","vii","ve","ve","vo","vo","vu","vuu","raa","re","re","ri","rii","ro","ro","ru","ruu","ry","ryaa","ryi","ryii","rye","ryu","ryuu","ryo","ryo","t","taa","te","te","ti","tii","to","to","tu","tuu","ty","tyaa","tye","tye","tyu","tyuu","tyo","tyo","tyi","tyii","y","yaa","ye","ye","yi","yii","yu","yuu","yo","yo","p","paa","pe","pe","pi","pii","pu","puu","po","po","py","pyaa","pye","pye","pyi","pyii","pyu","pyuu","pyo","pyo","praa","prii","pri","pre","pre","pru","pruu","pro","pro","pru","pruu","saa","si","sii","su","suu","so","so","se","se","sr","sraa","sri","srii","sru","sruu","sro","sro","sre","sre","sy","syaa","syi","syii","sye","sye","syu","syuu","syo","syo","d","daa","di","dii","do","do","du","duu","de","de","dh","dhaa","dhi","dhii","dho","dho","dhu","dhuu","dhe","dhe","ph","phaa","phi","phii","pho","pho","phu","phuu","phe","phe","phr","phraa","phri","phrii","phro","phro","phru","phruu","phre","phre","g","gaa","gi","gii","go","go","gu","guu","ge","ge","gy","gyaa","gyi","gyii","gyo","gyo","gyu","gyuu","gye","gye","graa","gri","grii","gro","gro","gru","gruu","gre","gre","gh","ghaa","ghi","ghii","gho","gho","ghu","ghuu","ghe","ghe","h","haa","hi","hii","ho","ho","hu","huu","he","he","j","jaa","ji","jii","jo","jo","ju","juu","je","je","jr","jraa","jri","jrii","jro","jro","jru","jruu","jre","jre","jy","jyaa","jyi","jyii","jyo","jyo","jyu","jyuu","jye","jye","jh","jhaa","jhi","jhii","jho","jho","jhu","jhuu","jhe","jhe","k","kaa","ki","kii","ko","ko","ku","kuu","ke","ke","ky","kyaa","kyi","kyii","kyo","kyo","kyu","kyuu","kye","kye","kr","kraa","kri","krii","kro","kro","kru","kruu","kre","kre","kv","kvaa","kvi","kvii","kvo","kvo","kvu","kvuu","kve","kve","tv","tvaa","tvi","tvii","tvo","tvo","tvu","tvuu","tve","tve","rv","rvaa","rvi","rvii","rvu","rvuu","rvo","rvo","rve","rve","pv","pvaa","pvi","pvii","pvo","pvo","pvu","pvuu","pve","pve","sv","svaa","svi","svii","svo","svo","svu","svuu","sve","sve","","phv","phvaa","phvi","phvii","phvo","phvo","phvu","phvuu","phve","phve","gv","gvaa","gvi","gvii","gvo","gvo","gvu","gvuu","gve","gve","hv","","jv","jhaa","jvi","jvii","jvo","jvo","jvu","jvuu","jve","jve","lv","lvaa","lvi","lvii","lvo","lvo","lvu","lvuu","lve","lve","l","laa","li","lii","lo","lo","lu","luu","le","le","lr","lr","lraa","lri","lrii","lro","lro","lru","lruu","lre","lre","ly","lyaa","lyi","lyii","lyo","lyo","lyu","lyuu","lye","lye","lh","lhaa","lhi","lhii","lho","lho","lhu","lhuu","lhe","lhe","ksh","kshaa","kshi","kshii","ksho","ksho","kshu","kshuu","kshe","kshe","c","caa","ci","cii","co","co","cu","cuu","ce","ce","ch","chaa","chi","chii","cho","cho","chu","chuu","che","che","chr","chraa","chri","chrii","chro","chro","chru","chruu","chre","chre","cr","craa","cri","crii","cro","cro","cru","cruu","cre","cre","cy","cyaa","cyi","cyii","cyo","cyo","cyu","cyuu","cye","cye","chy","v","vaa","vi","vii","vo","vo","vu","vuu","ve","ve","vy","vyaa","vyi","vyii","vyo","vyo","vyu","vyuu","vye","vye","vr","vraa","vri","vrii","vru","vruu","vro","vro","vre","vre","vh","vhaa","vhi","vhii","vho","vho","vhu","vhuu","vhe","vhe","b","baa","bi","bii","bo","bo","bu","buu","be","be","bh","bhaa","bhi","bhii","bho","bho","bhu","bhuu","bhe","bhe","bhr","bhraa","bhri","bhrii","bhro","bhro","bhru","bhruu","bhre","bhre","br","braa","bri","brii","bro","bro","bru","bruu","bre","bre","n","naa","ni","nii","nu","nuu","no","no","ne","ne","nh","nhaa","nhi","nhii","nho","nho","nhu","nhuu","nhe","nhe","ny","nyaa","nyu","nyuu","nyi","nyii","nyo","nyo","nye","nye","nv","nvaa","nvi","nvii","nvo","nvo","nvu","nvuu","nve","nve","ny","nyaa","nyi","nyii","nyo","nyo","nyu","nyuu","nye","nye","ng","ngaa","ngi","ngii","ngo","ngo","ngu","nguu","nge","nge","maa","mi","mii","mo","mo","mu","muu","me","me","my","myaa","myi","myii","myo","myo","myu","myuu","mye","mye","mr","mraa","mro","mro","mri","mrii","mru","mruu","mre","mre","mv","mv","mvaa","mvi","mvii","mvo","mvo","mvu","mvuu","mve","mve","hai"};
public static String EnglishToHindi(String english)
{
for(int i=0;i<DevText.length;i++)
{
if(English[i].equals(english))
return DevText[i];
}
return "";
}
public static String Hindi2English(String hindi)
{
for(int i=0;i<DevText.length;i++)
{
if(English[i].equals(hindi))
return English[i];
}
return "";
}
}
There are indeed better ways of table mapping like hashtable and better seraching techniques, and you are free to use them. Anybody working in Indian language NLP may find this as an easy and quick starting point.
There is also a commented main section which really acts like a spoiler, revealing the secret of the code. A wise programmer ( as all CP programmers are) you now know how on earth such tables can be generated. This will also illuminate you as how to go about doing things in other languages.
Points of Interest
I would try to incorporate some other interesting mapping techniques. Users can check out http://www.quillpad.in/ for other Indian languages.
History
Published on 07-Aug-2014
