Difference between revisions of "User talk:Perry Mizin"
Perry Mizin (talk | contribs) m (→OK) |
Perry Mizin (talk | contribs) m (→OK) |
||
| Line 486: | Line 486: | ||
if(contains(body, "lt>") != 0) | if(contains(body, "lt>") != 0) | ||
{ // Yahoo Spell Checker correction | { // Yahoo Spell Checker correction | ||
integer httsplID = llList2Integer(spellchecks, | integer httsplID = llList2Integer(spellchecks, 1); | ||
corrections = (corrections=[]) + corrections + [httsplID, llGetSubString(body, llSubStringIndex(body, "t>") + 2, llSubStringIndex(body, "</") - 1)]; | corrections = (corrections=[]) + corrections + [httsplID, llGetSubString(body, llSubStringIndex(body, "t>") + 2, llSubStringIndex(body, "</") - 1)]; | ||
llOwnerSay(llGetSubString(body, llSubStringIndex(body, "t>") + 2, llSubStringIndex(body, "</") - 1)); | llOwnerSay(llGetSubString(body, llSubStringIndex(body, "t>") + 2, llSubStringIndex(body, "</") - 1)); | ||
Revision as of 05:26, 23 December 2009
Bienvenid@
Hola, soy Perry Mizin y este será mi humilde intento de contribución al proyecto "Traductor Universal de Google" creado por Hank Ramos.
El siguiente código fué originálmente creado por Hank Ramos, empezaré modificando algunas funciones a mí manera... No ós riais
Welcome
Hello, I am Perry Mizin and this will be my humble try to contribute with the project "Universal Google Translator" created by Hank Ramos.
The next code was originally created by Hank Ramos, I will start modifiying some functions on my way... Don´t laugh
OK
Bueno, despues de respiñar todo lo que pude, creo que el uso de memoria y la velocidad están prácticamente igual qué en la versión original pero bueno... en la variedad está el gusto no? Ya he probado un poco el código, y parece que funciona aunque estoy teniendo algunos problemas... Por ejemplo creo que el sistema de traduccion no funciona corréctamente, y no me refiero al sistema de traduccion de Google (en mi opinion, si no es el mejor, es uno de los mejores del mundo), sino a que aún queda mucho trabajo por hacer en el script...
El siguiente paso sera intentar dividir el script en dos, o si es posible en tres para repartir el uso de memoria. Tambien me gustaria probar un truco que uso a menudo para traducir páginas en un idioma que no entiendo, lo que hago es que en vez de traducir diréctamente al español, primero se traduce a inglés, y luego desde inglés al español. Parece una tontería y un gasto innecesario, pero si lo hago es porque así suelo entender mejor las traducciones. Quiero por un lado el sistema de comunicacion y evaluacion entre traductores, y por otro el sistema de traduccion entre avatares, ¿es muy dificil? no sé ya lo veremos...
Well, after saving as much as I could, I think the memory usage and the speed are technically the same as in the original script, but well, in the variety is the taste, isnt it?. I have tested a bit the code above already, and it seems to work altought i´m having some problems... For instance i think the translation system is not working properly, and i dont mean the Google´s translation system (in my opinion, if is not the best, one of the bests), but that there is still much work to do in the script...
The next step will be to try to divide the script in 2, or 3 if is possible, so we can spread the memory usage. Also I would like to test a trick i use often for to translate languages that i dont understand, what I do first is to translate that language into English, and then from English to Spanish, it seems a nonsense and unnecesary waste, but if i do it is becouse so i usually understand the translations better. I want by one part the translators communications and evaluation system, and by other the translation system between avatars, is it very difficult? I don´t know , we will see...
Es dificil... me está costando pero ya veremos... Its difficult... i am struggling but we will see...
chu chu chu
De Güatemala a Güatepeor, Sin comentarios...
From Güatebad to Güateworse, No comments...
Buscando en Internet he encontrado este API de Yahoo comprobador de ortografía, aunque creo que por ahora sólo soporta el lenguage inglés pero sería interesante añadirlo al proyecto
Searching on the web I crossed in my way with this spell checker Yahoo´s API , I think that it only supports English language by now althought it would be interesting to add it to the project
Estoy probando el comprobador de deletreo de Yahoo aunque la verdad es que no estoy teniendo buenos resultados, parece que Google es mucho más rapido y necesito corregir antes de traducir. También estoy intentando solucionar el error que aparece con el símbolo (apostrofe) ´ en las frases con genitivo sajon traducidas al inglés y tampoco doy con la tecla... Ya que esto me está dando muchos dolores de cabeza quizás deje el proyecto en standby hasta después de navidades.
I am testing the Yahoo´s spell checker althought to be honest I´m not having good success, it seems Google is much faster and I need to correct before translate. Also I am triying to fix the error showed in the phrases translated to English with the saxons genitive apostrophe ´ and still triying and triying. As this is driving me nuts, maybe I will leave the project in standby until new year.
ESTO ES EL TRADUCTOR "un lío" por ahora... Necesito rehacer algunas cosas
THIS IS THE TRANSLATOR "mess" by now... I need to redo some things
<lsl>
//Personal Translator (Mod made in Perry Mizin)
//Universal Translator
//Version 1.9.0
//November 12, 2009
//LSL Script copyright 2006-2009 by Hank Ramos
//Translation Web Services powered by Google
//Spell Checker Web Services powered by Yahoo
//Variables
list requestedTranslations;
list tempDetections;
list agentsInTranslation;
list spellchecks;
list corrections;
integer M = 1;
integer listenID;
integer isMaster = 1;
integer autoLanguage = TRUE;
integer enabled = FALSE;
integer showTranslation = FALSE;
integer tranObjects = TRUE;
integer isInitialized = FALSE;
string options;
string tempcorrection;
key tempValue;
// Functions
integer contains(string haystack, string needle) // http://wiki.secondlife.com/wiki/llSubStringIndex
{
return ~llSubStringIndex(haystack, needle);
}
sendIM(key imid, string imstr) {
// llInstantMessage(key id , string message) causes the script to sleep for 2 seconds
if (llGetAgentSize(imid) == ZERO_VECTOR)
{ }
else
{
if (imid != llGetOwner())
{
llOwnerSay(imstr + " @" + llKey2Name(imid)); // llOwnerSay(string message) has no delay
}
}
if (M == 8)
{
llMessageLinked(LINK_THIS, 2334420, imstr, imid);
M = 1;
}
else if (M == 7)
{
llMessageLinked(LINK_THIS, 2334430, imstr, imid);
M = 8;
}
else if (M == 6)
{
llMessageLinked(LINK_THIS, 2334440, imstr, imid);
M = 7;
}
else if (M == 5)
{
llMessageLinked(LINK_THIS, 2334450, imstr, imid);
M = 6;
}
else if (M == 4)
{
llMessageLinked(LINK_THIS, 2334460, imstr, imid);
M = 5;
}
else if (M == 3)
{
llMessageLinked(LINK_THIS, 2334470, imstr, imid);
M = 4;
}
else if (M == 2)
{
llMessageLinked(LINK_THIS, 2334480, imstr, imid);
M = 3;
}
else if (M == 1)
{
llMessageLinked(LINK_THIS, 2334490, imstr, imid);
M = 2;
}
}
DetectLanguage(string dtcmsg, key dtcid) {
tempValue = llHTTPRequest("http://ajax.googleapis.com/ajax/services/language/detect?v=1.0&q=" + llEscapeURL(dtcmsg), [HTTP_METHOD, "GET", HTTP_MIMETYPE, "plain/text;charset=utf-8"], " ");
if (tempValue != NULL_KEY)
{
tempDetections = (tempDetections=[]) + tempDetections + [tempValue, dtcid];
}
}
CheckSpell(integer splnum, string splmsg) {
key spellValue;
spellValue = llHTTPRequest("http://search.yahooapis.com/WebSearchService/V1/spellingSuggestion?appid=DNPjiabV34EOgXQ3kAu6.d7kf6v.JUGCn1bFUcSTFMSGCyCy_ALzbepYstlI7DrOGP0-&query=" + splmsg, [HTTP_METHOD, "GET", HTTP_MIMETYPE, "plain/text;charset=utf-8"], "");
if (spellValue != NULL_KEY)
{ // Translation
spellchecks = (spellchecks=[]) + spellchecks + [spellValue, splnum];
}
}
Translate(integer chkn, string chkmsg, list chkli) {
key returnValue;
returnValue = llHTTPRequest("http://ajax.googleapis.com/ajax/services/language/translate?v=1.0&q=" + chkmsg, [HTTP_METHOD, "GET", HTTP_MIMETYPE, "plain/text;charset=utf-8"], "");
if (returnValue != NULL_KEY)
{ // Translation
requestedTranslations = (requestedTranslations=[]) + requestedTranslations + [returnValue, llList2CSV(chkli)];
}
else
{
llSleep(25.0); // Something has gone horribly wrong, sleep 25 seconds to clear throttle
}
}
string checkLanguage(string chklantmpStr) {
if (chklantmpStr == "")
{
chklantmpStr = "en";
}
else if (chklantmpStr == "und")
{
chklantmpStr = "el";
}
if (llGetSubString(chklantmpStr, 0, 1) == "zh")
{
chklantmpStr = "zh-CN";
}
else
{
chklantmpStr = llGetSubString(chklantmpStr, 0, 1);
}
return chklantmpStr;
}
addAgent(key avid, string avlan, integer chklan) {
integer avliPos;
integer idNum;
if (~llListFindList(agentsInTranslation, (list)avid))
{
avliPos = llListFindList(agentsInTranslation, (list)avid);
agentsInTranslation = llListReplaceList(agentsInTranslation, [avlan, avlan, chklan], avliPos + 1, avliPos + 3);
}
else
{
@at;
idNum = llRound(llFrand(2000000)) + 1;
if(~llListFindList(agentsInTranslation, (list)idNum))
{
jump at;
}
agentsInTranslation = (agentsInTranslation=[]) + agentsInTranslation + [avid, avlan, avlan, chklan, idNum];
// llMessageLinked(LINK_THIS, 64562349, avlan, avid);
}
}
string addNewAgent(key newavid) {
string speakerLan;
if (llGetAgentSize(newavid) == ZERO_VECTOR)
{
speakerLan = "en";
addAgent(newavid, speakerLan, 4);
}
else
{
speakerLan = checkLanguage(llGetAgentLanguage(newavid));
if (speakerLan == "")
{
speakerLan = "en";
addAgent(newavid, speakerLan, 0);
}
else
{
addAgent(newavid, speakerLan, 0);
}
}
return speakerLan;
}
key getAgentKey(integer agentID) {
if(~llListFindList(agentsInTranslation, (list)agentID))
{
integer gkliPos = llListFindList(agentsInTranslation, (list)agentID);
return llList2Key(agentsInTranslation, gkliPos - 4);
}
else
{
return "";
}
}
processHTTPResponse(integer ntype, string bodyStr, list httli) {
key recepientID;
key speakerID;
list paramtmpli;
list recepientList;
string recepientLanguage;
string languagePair;
string speakerName;
string speakerLanguage;
string translatedText;
string lantmpStr;
integer z;
//===================
// Process Translation
//===================
speakerID = llList2Key(httli, 1);
speakerName = llKey2Name(speakerID);
if (speakerName == "")
{
speakerName = llList2String(llGetObjectDetails(speakerID, [OBJECT_NAME]), 0);
}
recepientList = llParseString2List(llList2String(httli, 2), ["@"], []);
paramtmpli = llParseStringKeepNulls(llList2String(httli, 3), ["|"],[]);
recepientLanguage = llList2String(paramtmpli, 1);
languagePair = llDumpList2String(paramtmpli, ">>");
// Perform Text Cleanup
translatedText = llGetSubString(bodyStr, llSubStringIndex(bodyStr, "{\"translatedText\":\"") + 18, llSubStringIndex(bodyStr, "\",\"detectedSourceLanguage\":\""));
// Reverse order if Recepient Language is Hebrew or Arabic
if ((recepientLanguage == "iw") || (recepientLanguage == "ar"))
{
lantmpStr = "";
for(z = llStringLength(translatedText);z >= 0; z--)
{
lantmpStr = (lantmpStr="") + lantmpStr + llGetSubString(translatedText, z, z);
}
translatedText = lantmpStr;
}
lantmpStr = speakerName + "(" + languagePair + "): " + translatedText;
for (z = 0; z < llGetListLength(recepientList); z += 1)
{
recepientID = getAgentKey(llList2Integer(recepientList, z));
if (recepientID != "")
{
recepientLanguage = llList2String(agentsInTranslation, llListFindList(agentsInTranslation, [recepientID]) + 1);
if (recepientLanguage == speakerLanguage)
{ }
else
{
sendIM(recepientID, lantmpStr);
}
}
}
return;
}
default {
state_entry()
{
llSetTimerEvent(5);
}
link_message(integer sender_num, integer num, string lstr, key lid)
{
list ltmpli;
if (num == 8434532)
{
enabled = (integer)lstr;
}
else if (num == 3342976)
{ // Send Preferences
options = lstr;
if (isInitialized == FALSE)
{
return;
}
ltmpli = llCSV2List(options);
showTranslation = llList2Integer(ltmpli, 0);
tranObjects = llList2Integer(ltmpli, 1);
autoLanguage = llList2Integer(ltmpli, 2);
}
}
sensor(integer num_detected)
{
integer s;
key sid;
for (s = 0; s < num_detected; s += 1)
{
sid = llDetectedKey(s);
if (~llListFindList(agentsInTranslation, (list)sid))
{ }
else
{
addNewAgent(sid);
}
}
}
no_sensor()
{ }
listen(integer channel, string name, key id, string message)
{
integer w;
string spkrLan;
string recepLan;
integer recepID;
integer lstliPos;
string lanpair;
list translationCache;
// Translator Engine Code
integer msgid = llRound(llFrand(2000000)) + 1;
CheckSpell(msgid , llEscapeURL(message));
llSleep(1.0);
if ((!enabled) && (isMaster == 1))
{
return;
}
if (!tranObjects)
{
if (llGetAgentSize(id) == ZERO_VECTOR)
{
return;
}
else
{
DetectLanguage(message, id);
}
}
else
{
DetectLanguage(message, id);
}
if (~llListFindList(agentsInTranslation, (list)id))
{
lstliPos = llListFindList(agentsInTranslation, (list)id);
spkrLan = llList2String(agentsInTranslation, lstliPos + 1);
}
else
{
spkrLan = addNewAgent(id);
}
//===============================
// Formulate Translation Requests
//===============================
if (id == llGetOwner())
{ // If speaker is owner
// I will do this dirty cheat by now
// There should exist another clean way to filter speakers & receivers, I need to think about this
for (w = 0; w < llGetListLength(agentsInTranslation); w += 5)
{ // Loop through translation group and do appropriate translations as needed
recepID = llList2Integer(agentsInTranslation, w + 4);
recepLan = checkLanguage(llList2String(agentsInTranslation, w + 1));
if (spkrLan == recepLan)
{ }
else
{
lanpair = spkrLan + "|" + recepLan;
if (~llListFindList(translationCache, (list)lanpair))
{
lstliPos = llListFindList(translationCache, (list)lanpair);
translationCache = llListReplaceList(translationCache, [llList2String(translationCache, lstliPos + 1) + "@" + (string)recepID], lstliPos + 1, lstliPos + 1);
}
else
{
translationCache = (translationCache=[]) + translationCache + [lanpair, recepID];
}
}
}
if (translationCache != [])
{ // Process Requests
if (~llListFindList(corrections, (list)msgid))
{
message = llList2String(corrections, 1);
corrections = [];
}
else
{
corrections = [];
}
for (w = 0; w < llGetListLength(translationCache); w += 2)
{ //====================================
//Translation
//====================================
//Forumulate and Send Translation Request
lanpair = "|" + llList2String(llParseStringKeepNulls(llList2String(translationCache, w), ["|"],[]), 1);
Translate(0, llEscapeURL(message) + "&langpair=" + llEscapeURL(lanpair), [llGetTime(), id , llList2String(translationCache, w + 1), llList2String(translationCache, w)]);
}
}
}
else
{ // if speaker is not owner
lstliPos = llListFindList(agentsInTranslation, (list)llGetOwner());
recepID = llList2Integer(agentsInTranslation, lstliPos + 4);
recepLan = checkLanguage(llList2String(agentsInTranslation , lstliPos + 1 ));
if (spkrLan == recepLan)
{ }
else
{
lanpair = spkrLan + "|" + recepLan;
if (~llListFindList(translationCache, (list)lanpair))
{
lstliPos = llListFindList(translationCache, (list)lanpair);
translationCache = llListReplaceList(translationCache, [llList2String(translationCache, lstliPos + 1) + "@" + (string)recepID], lstliPos + 1, lstliPos + 1);
}
else
{
translationCache = (translationCache=[]) + translationCache + [lanpair, recepID];
}
if (translationCache != [])
{ //====================================
//Translation
//====================================
//Forumulate and Send Translation Request
lanpair = "|" + llList2String(llParseStringKeepNulls(llList2String(translationCache, lstliPos), ["|"],[]), 1);
Translate(0, llEscapeURL(message) + "&langpair=" + llEscapeURL(lanpair), [llGetTime(), id , llList2String(translationCache, lstliPos + 1), llList2String(translationCache, lstliPos)]);
}
}
}
}
http_response(key request_id, integer status, list metadata, string body)
{ // Process Response Code
integer httliPos;
if (~llListFindList(spellchecks, (list)request_id))
{
if(contains(body, "lt>") != 0)
{ // Yahoo Spell Checker correction
integer httsplID = llList2Integer(spellchecks, 1);
corrections = (corrections=[]) + corrections + [httsplID, llGetSubString(body, llSubStringIndex(body, "t>") + 2, llSubStringIndex(body, "</") - 1)];
llOwnerSay(llGetSubString(body, llSubStringIndex(body, "t>") + 2, llSubStringIndex(body, "</") - 1));
spellchecks = [];
}
else
{
spellchecks = [];
}
}
if (~llListFindList(requestedTranslations, (list)request_id))
{ // Translation
httliPos = llListFindList(requestedTranslations, (list)request_id);
processHTTPResponse(0, body, llCSV2List(llList2String(requestedTranslations, httliPos + 1)));
requestedTranslations = llDeleteSubList(requestedTranslations, httliPos, httliPos + 1);
return;
}
if (~llListFindList(tempDetections, (list)request_id))
{ // Language Detection
integer tmpliPos = llListFindList(tempDetections, (list)request_id);
key speakerID = llList2Key(tempDetections, tmpliPos + 1);
tempDetections = llDeleteSubList(tempDetections, tmpliPos, tmpliPos + 3);
string htttmpStr = llToLower(llGetSubString(body, llSubStringIndex(body, "{\"language\":\"") + 13, llSubStringIndex(body, "\",\"isReliable\":") - 1));
if (contains(htttmpStr, "err") == 0 && contains(htttmpStr, "unn") == 0)
{ // We Trust the Google system, the Google´s guys are good and I am sure they will keep improving :-)
// Language Detected
htttmpStr = checkLanguage(htttmpStr);
if (~llListFindList(agentsInTranslation, (list)speakerID))
{
integer tmpavliPos = llListFindList(agentsInTranslation, (list)speakerID);
string actlan = llList2String(agentsInTranslation, tmpavliPos + 1); string tmplan = llList2String(agentsInTranslation, tmpavliPos + 2);
if (llList2Integer(agentsInTranslation, tmpavliPos + 3) >= 2)
{ // if anyone speaks on any different language 3 times in a row
addAgent(speakerID, htttmpStr, 0);
}
if (actlan == htttmpStr)
{ // reset language checker for speaker
agentsInTranslation = llListReplaceList(agentsInTranslation, [actlan, 0 ], tmpavliPos + 2, tmpavliPos + 3);
}
else
{
if ( htttmpStr == tmplan)
{ // if language of speaker is not the same as detected language
agentsInTranslation = llListReplaceList(agentsInTranslation,[llList2Integer(agentsInTranslation,tmpavliPos + 3) + 1] , tmpavliPos + 3, tmpavliPos + 3);
}
else
{
agentsInTranslation = llListReplaceList(agentsInTranslation,[htttmpStr, 0] , tmpavliPos + 2, tmpavliPos + 3);
}
}
}
}
}
}
timer()
{
integer r;
vector tmrVct;
list newList;
float timeElapsed;
llSetTimerEvent(0.0); // we better reset timer by way of precaution
if (isInitialized == FALSE)
{
enabled = TRUE;
listenID = llListen(0, "", "", "");
llMessageLinked(LINK_THIS, 6877259, (string)enabled, NULL_KEY);
isInitialized = TRUE;
string mylanguage = llGetAgentLanguage(llGetOwner());
if (llStringLength(mylanguage) < 1)
{
mylanguage = "en";
}
if(~llListFindList(agentsInTranslation, (list)llGetOwner()))
{
integer ownerliPos = llListFindList(agentsInTranslation, (list)llGetOwner());
agentsInTranslation = llListReplaceList(agentsInTranslation, [mylanguage] , ownerliPos + 1, ownerliPos + 1);
}
else
{
agentsInTranslation = (agentsInTranslation=[]) + agentsInTranslation + [llGetOwner() , mylanguage, mylanguage , 0 , llRound(llFrand(2000000)) + 1];
}
}
for (r = 0; r < llGetListLength(tempDetections); r += 2)
{
timeElapsed = llGetTime() - llList2Float(llCSV2List(llList2String(tempDetections, r + 1)), 0);
if (timeElapsed < 20.0)
{
newList = (newList=[]) + newList + llList2List(tempDetections, r, r + 1);
}
}
tempDetections = newList;
newList = [];
for (r = 0; r < llGetListLength(requestedTranslations); r += 2)
{
timeElapsed = llGetTime() - llList2Float(llCSV2List(llList2String(requestedTranslations, r + 1)), 0);
if (timeElapsed < 20.0)
{
newList = (newList=[]) + newList + llList2List(requestedTranslations, r, r + 1);
}
}
requestedTranslations = newList;
newList = [];
if (isMaster > 0)
{
for (r = 0; r < llGetListLength(agentsInTranslation); r += 5)
{
tmrVct = llList2Vector(llGetObjectDetails(llList2Key(agentsInTranslation, r), [OBJECT_POS]), 0);
if (llVecDist(llGetPos(), tmrVct) <= 20.0)
{
newList = (newList=[]) + newList + llList2List(agentsInTranslation, r, r + 4);
}
}
agentsInTranslation = newList;
newList = [];
}
if ((autoLanguage) && isMaster > 0)
{
llSensor("", NULL_KEY, AGENT, 20.0, PI);
}
llSetTimerEvent(4.0); // restart timer
}
}
</lsl>