private static void TestSplit(CSpellApi cSpellApi)
{
// setup test case
// 10349.txt
//String inText = "sounding in my ear every time for along time.";
// 13864.txt
string inText = "I donate my self to be apart of this study.";
TextObj textObj = new TextObj(inText);
List <TokenObj> inTextList = textObj.GetTokenList();
List <TokenObj> nonSpaceTokenList = TextObj.GetNonSpaceTokenObjList(inTextList);
//int tarPos = 7;
int tarPos = 6;
TokenObj inTokenObj = nonSpaceTokenList[tarPos];
bool debugFlag = false;
Console.WriteLine("====== Real-Word One-To-One Correction Test =====");
Console.WriteLine("-- inTextList: [" + inText + "]");
Console.WriteLine("-- tarPos: [" + tarPos + "]");
Console.WriteLine("-- inTokenObj: [" + inTokenObj.ToString() + "]");
// get the correct term
TokenObj outTokenObj = GetCorrectTerm(inTokenObj, cSpellApi, debugFlag, tarPos, nonSpaceTokenList);
// print out
Console.WriteLine("--------- GetCorrectTermStr( ) -----------");
Console.WriteLine("-- outTokenObj: [" + outTokenObj.ToString() + "]");
}
// public methods
// private methods
private static void Test()
{
int tarIndex = 6; // target index
int startIndex = 4; // start index of merge
int endIndex = 6; // end index of merge
int tarPos = 3; // target pos
int startPos = 2; // start pos of merge
int endPos = 3; // end pos of merge
int mergeNo = 1; // total no of merged tokens
string tarWord = "gnosed"; // target term
string mergeWord = "diagnosed."; // suggested merged terms
string coreMergeWord = "diagnosed"; // core suggested merged terms
string orgMergeWord = "dia gnosed"; // org word b4 merge
MergeObj mergeObj = new MergeObj(tarWord, orgMergeWord, mergeWord, coreMergeWord, mergeNo, startIndex, tarIndex, endIndex, startPos, tarPos, endPos);
string inText = "He is dia gnosed last week.";
List <TokenObj> inTextList = TextObj.TextToTokenList(inText);
List <TokenObj> nonSpaceTextList = TextObj.GetNonSpaceTokenObjList(inTextList);
Console.WriteLine("------ Merge Obj -------");
Console.WriteLine(mergeObj.ToString());
Console.WriteLine("------ Non Merge Term -------");
string nonMergeTerm = GetNonMergeTerm(mergeObj, nonSpaceTextList);
Console.WriteLine("- inText: [" + inText + "]");
Console.WriteLine("- nonMergeTerm: [" + nonMergeTerm + "]");
}
private static void TestOnSet(Word2Vec w2vIm, Word2Vec w2vOm)
{
string inText = "He was diagnosed early on set dementia 3 years ago.";
TextObj textObj = new TextObj(inText);
List <TokenObj> inTextList = textObj.GetTokenList();
// remove space token from the list
List <TokenObj> nonSpaceTokenList = TextObj.GetNonSpaceTokenObjList(inTextList);
Console.WriteLine("==========================================");
Console.WriteLine("-- inTextList: [" + inText + "]");
int tarPos = 4;
int tarSize = 2; // "on set" has 2 tokens
int radius = 2;
bool word2VecSkipWord = true;
bool debugFlag = false;
// 1 context with window radius
DoubleVec contextVec = Word2VecContext.GetContextVec(tarPos, tarSize, nonSpaceTokenList, w2vIm, radius, word2VecSkipWord, debugFlag);
string str1 = "onset";
ContextScore s1 = new ContextScore(str1, contextVec, w2vOm);
string str2 = "on set";
ContextScore s2 = new ContextScore(str2, contextVec, w2vOm);
Console.WriteLine("- [" + str1 + "]: " + s1.ToString());
Console.WriteLine("- [" + str2 + "]: " + s2.ToString());
}
private static void Test1To1(CSpellApi cSpellApi)
{
// setup test case
// 51.txt
//String inText = "You'd thing that this is good.";
//String inText = "The doctor thing that this is good.";
string inText = "you would thing that is good.";
TextObj textObj = new TextObj(inText);
List <TokenObj> inTextList = textObj.GetTokenList();
List <TokenObj> nonSpaceTokenList = TextObj.GetNonSpaceTokenObjList(inTextList);
int tarPos = 2;
TokenObj inTokenObj = nonSpaceTokenList[tarPos];
bool debugFlag = false;
Console.WriteLine("====== Real-Word One-To-One Correction Test =====");
Console.WriteLine("-- inTextList: [" + inText + "]");
Console.WriteLine("-- tarPos: [" + tarPos + "]");
Console.WriteLine("-- inTokenObj: [" + inTokenObj.ToString() + "]");
// get the correct term
TokenObj outTokenObj = GetCorrectTerm(inTokenObj, cSpellApi, debugFlag, tarPos, nonSpaceTokenList);
// print out
Console.WriteLine("--------- GetCorrectTermStr( ) -----------");
Console.WriteLine("-- outTokenObj: [" + outTokenObj.ToString() + "]");
}
// public method
// process
public static List <TokenObj> Process(List <TokenObj> inTokenList, CSpellApi cSpellApi, bool debugFlag)
{
DebugPrint.PrintProcess("5. RealWord-Merge", debugFlag);
DebugPrint.PrintInText(TextObj.TokenListToText(inTokenList), debugFlag);
// pre-porcess
// update Pos for the inTokenList
TextObj.UpdateIndexPos(inTokenList);
// 1. remove non space-token and convert to non-space-token list
List <TokenObj> nonSpaceTokenList = TextObj.GetNonSpaceTokenObjList(inTokenList);
// 2. process: go through each token for detection and correction
// to find merge corrections (mergeObjList)
int index = 0;
List <MergeObj> mergeObjList = new List <MergeObj>();
int maxLegitTokenLength = cSpellApi.GetMaxLegitTokenLength();
while (index < inTokenList.Count)
{
TokenObj curTokenObj = inTokenList[index];
// update the tarPos
// SCR-3, use legit token
if (curTokenObj.IsLegitToken(maxLegitTokenLength) == true)
{
int tarPos = inTokenList[index].GetPos();
// correct term is the highest ranked candidates
MergeObj mergeObj = RealWordMergeCorrector.GetCorrectTerm(tarPos, nonSpaceTokenList, cSpellApi, debugFlag);
if (mergeObj == null) // no merge correction
{
index++;
}
else // has merge correction
{
mergeObjList.Add(mergeObj);
// next token after end token, this ensures no overlap merge
index = mergeObj.GetEndIndex() + 1;
}
}
else // space token
// update index
{
index++;
}
}
// update the output for merge for the whole inTokenList,
// has to update after the loop bz merge might
// happen to the previous token
// update the tokenObj up to the merge, then go to the next token
// update operation info also
List <TokenObj> outTokenList = MergeCorrector.CorrectTokenListByMerge(inTokenList, mergeObjList, TokenObj.HIST_RW_M, debugFlag, cSpellApi);
return(outTokenList);
}
// private method
// Test merge and Split
private static void Test(string inText, int tarPos, int tarSize, int radius, string mergedWord, string splitWords, Word2Vec w2vIm, Word2Vec w2vOm)
{
// 0. process the inText
TextObj textObj = new TextObj(inText);
List <TokenObj> inTextList = textObj.GetTokenList();
// remove space token from the list
List <TokenObj> nonSpaceTokenList = TextObj.GetNonSpaceTokenObjList(inTextList);
Console.WriteLine("==========================================");
Console.WriteLine("-- inTextList: [" + inText + "]");
bool word2VecSkipWord = true;
bool debugFlag = false;
// 1.a context with window radius
DoubleVec contextVec = Word2VecContext.GetContextVec(tarPos, tarSize, nonSpaceTokenList, w2vIm, radius, word2VecSkipWord, debugFlag);
// 1.b context with all inText
DoubleVec contextVecA = Word2VecContext.GetContextVec(tarPos, tarSize, nonSpaceTokenList, w2vIm, word2VecSkipWord, debugFlag);
// 1.c get score1
ContextScore score1 = new ContextScore(mergedWord, contextVec, w2vOm);
ContextScore score1a = new ContextScore(mergedWord, contextVecA, w2vOm);
Console.WriteLine(score1.ToString() + "|" + string.Format("{0,1:F8}", score1a.GetScore()));
// 2. split words
ContextScore score2 = new ContextScore(splitWords, contextVec, w2vOm);
ContextScore score2a = new ContextScore(splitWords, contextVecA, w2vOm);
Console.WriteLine(score2.ToString() + "|" + string.Format("{0,1:F8}", score2a.GetScore()));
// 3. 3. 3. Use avg. score on single words
// This method use different context for each single word
List <string> splitWordList = TermUtil.ToWordList(splitWords);
int index = 0;
double scoreSAvg = 0.0d; // radius
double scoreSAAvg = 0.0d; // all inText
//debugFlag = false;
foreach (string splitWord in splitWordList)
{
// window radius
DoubleVec contextVecS = Word2VecContext.GetContextVec(tarPos + index, 1, nonSpaceTokenList, w2vIm, radius, word2VecSkipWord, debugFlag);
ContextScore scoreS = new ContextScore(splitWord, contextVecS, w2vOm);
//System.out.println("-- " + scoreS.ToString());
scoreSAvg += scoreS.GetScore();
// all text
DoubleVec contextVecSA = Word2VecContext.GetContextVec(tarPos + index, 1, nonSpaceTokenList, w2vIm, word2VecSkipWord, debugFlag);
ContextScore scoreSA = new ContextScore(splitWord, contextVecSA, w2vOm);
//System.out.println("-- " + scoreSA.ToString());
scoreSAAvg += scoreSA.GetScore();
index++;
}
scoreSAvg = scoreSAvg / index; // window
scoreSAAvg = scoreSAAvg / index; // all text
Console.WriteLine("Avg. Single Word|" + string.Format("{0,1:F8}", scoreSAvg) + "|" + string.Format("{0,1:F8}", scoreSAAvg));
}
private static void TestGetCorrectTerm(CSpellApi cSpellApi)
{
// init
// all lowerCase
string inText = "Dur ing my absent.";
bool debugFlag = false;
List <TokenObj> inTokenList = TextObj.TextToTokenList(inText);
// 1. convert to the non-empty token list
List <TokenObj> nonSpaceTokenList = TextObj.GetNonSpaceTokenObjList(inTokenList);
// result
int tarPos = 0;
MergeObj mergeObj = NonWordMergeCorrector.GetCorrectTerm(tarPos, nonSpaceTokenList, cSpellApi, debugFlag);
// print out
Console.WriteLine("--------- GetCorrectTerm( ) -----------");
Console.WriteLine("In: [" + inText + "]");
Console.WriteLine("In nonSpaceTokenList: [" + nonSpaceTokenList.Count + "]");
Console.WriteLine("Out MergeObj: [" + mergeObj.ToString() + "]");
}
// public method
// Use: for loop, the latest and greatest implementation
// original implementation with for loop, To be deleted
// the core of spell-correction, include split
// inTokenList is the whole text
public static List <TokenObj> Process(List <TokenObj> inTokenList, CSpellApi cSpellApi, bool debugFlag)
{
DebugPrint.PrintProcess("3-4. NonWord-Split & 1To1", debugFlag);
DebugPrint.PrintInText(TextObj.TokenListToText(inTokenList), debugFlag);
// init the output TokenList
List <TokenObj> outTokenList = new List <TokenObj>();
// process: go through each token for detection and correction
// for the 1-to-1 and split correction
int tarPos = 0; // the position of the tokenObj in the inTokenList
// remove space token from the list
List <TokenObj> nonSpaceTokenList = TextObj.GetNonSpaceTokenObjList(inTokenList);
// use the inTokenList to keep the same spcae token
TokenObj outTokenObj = null;
int maxLegitTokenLength = cSpellApi.GetMaxLegitTokenLength();
foreach (TokenObj tokenObj in inTokenList)
{
/// <summary>
/// no context
/// TokenObj outTokenObj = SpellCorrector.GetCorrectTerm(tokenObj,
/// cSpellApi, debugFlag);
///
/// </summary>
// skip empty space tokens and long tokens
// SCR-3, use legit token
if (tokenObj.IsLegitToken(maxLegitTokenLength) == true)
{
// correct term is the highest ranked candidate
outTokenObj = NonWordCorrector.GetCorrectTerm(tokenObj, cSpellApi, debugFlag, tarPos, nonSpaceTokenList);
// used tarPos for context module
tarPos++;
}
else
{
outTokenObj = tokenObj;
}
// add the corrected tokenObj to the output token list
// use FlatMap because there might be a split
Split1To1Corrector.AddSplit1To1Correction(outTokenList, outTokenObj);
}
return(outTokenList);
}
// private methods
private static void Test()
{
// init cSpellApi
string configFile = "../data/Config/cSpell.properties";
CSpellApi cSpellApi = new CSpellApi(configFile);
Console.WriteLine("===== Unit Test of MergeCandidates =====");
//String inText = "He was dia gnosed early onset deminita 3 year ago.";
// example from 73.txt
//String inText = "I have seven live births with no problems dur ing my pregnancies. That is a dis appoint ment";
string inText = "That is a disa ppoint ment.";
List <TokenObj> inTextList = TextObj.TextToTokenList(inText);
string inStr = String.Join("|", inTextList.Select(obj => obj.GetTokenStr()));
Console.WriteLine(" - inTextList (" + inTextList.Count + "): [" + inStr + "]");
Console.WriteLine("-------------------------");
foreach (TokenObj tokenObj in inTextList)
{
Console.WriteLine(tokenObj.ToString());
}
int tarPos = 4;
Console.WriteLine("-------------------------");
Console.WriteLine("- tarPos: " + tarPos);
Console.WriteLine("- maxMergeNo: " + cSpellApi.GetCanNwMaxMergeNo());
Console.WriteLine("------ merge set -------");
// pre-Process: convert to the non-empty token list
List <TokenObj> nonSpaceTextList = TextObj.GetNonSpaceTokenObjList(inTextList);
// get the candidate for a specified target position
HashSet <MergeObj> mergeSet = GetCandidates(tarPos, nonSpaceTextList, cSpellApi);
// print out
foreach (MergeObj mergeObj in mergeSet)
{
Console.WriteLine(mergeObj.ToString());
}
}
private static void Tests(Word2Vec w2vIm, Word2Vec w2vOm)
{
string inText = "... last 10 years #$% was dianosed test123 yahoo.com early on set deminita 3 year ago.";
List <TokenObj> inTextList = TextObj.TextToTokenList(inText);
Console.WriteLine("======= Word2VecContext ======================");
Console.WriteLine(" - inText: [" + inText + "]");
string inStr = String.Join("|", inTextList.Select(obj => obj.GetTokenStr()));
Console.WriteLine(" - inTextList (" + inTextList.Count + "): [" + inStr + "]");
int tarPos = 0;
int tarSize = 1;
int index = 0;
int radius = 3;
bool debugFlag = false;
Console.WriteLine("------ Test GetContext (no skip), radius=3 ...");
// remove space token from the list
List <TokenObj> nonSpaceTokenList = TextObj.GetNonSpaceTokenObjList(inTextList);
foreach (TokenObj tokenObj in inTextList)
{
// not the space token
if (tokenObj.IsSpaceToken() == false)
{
string tokenStr = tokenObj.GetTokenStr();
// word2VecSkipWord = false (no skip)
List <string> contextList = GetContext(tarPos, tarSize, nonSpaceTokenList, w2vIm, radius, false, debugFlag);
string contextStr = String.Join("|", contextList);
Console.WriteLine(tarPos + "|" + index + "|" + tokenStr + ": [" + contextStr + "]");
tarPos++;
}
index++;
}
Console.WriteLine("------ Test GetContext (skip) , radius=3 ...");
Console.WriteLine(" - inText: [" + inText + "]");
tarPos = 0;
foreach (TokenObj tokenObj in inTextList)
{
// not the space token
if (tokenObj.IsSpaceToken() == false)
{
string tokenStr = tokenObj.GetTokenStr();
// word2VecSkipWord = true (skip)
List <string> contextList2 = GetContext(tarPos, tarSize, nonSpaceTokenList, w2vIm, radius, true, debugFlag);
string contextStr2 = String.Join("|", contextList2);
Console.WriteLine(tarPos + "|" + index + "|" + tokenStr + ": [" + contextStr2 + "]");
tarPos++;
}
index++;
}
Console.WriteLine("------ Test GetContext (skip) , all ...");
Console.WriteLine(" - inText: [" + inText + "]");
tarPos = 0;
// not the space token
foreach (TokenObj tokenObj in nonSpaceTokenList)
{
string tokenStr = tokenObj.GetTokenStr();
// word2VecSkipWord = true (skip)
List <string> contextList3 = GetContext(tarPos, tarSize, nonSpaceTokenList, w2vIm, true, debugFlag);
string contextStr3 = String.Join("|", contextList3);
Console.WriteLine(tarPos + "|" + tokenStr + ": [" + contextStr3 + "]");
tarPos++;
}
}
private static void Tests(Word2Vec w2vIm, Word2Vec w2vOm)
{
string inText = "for the last 10 years was dianosed\n early on set deminita 3 years ago";
List <TokenObj> inTextList = TextObj.TextToTokenList(inText);
// remove space token from the list
List <TokenObj> nonSpaceTokenList = TextObj.GetNonSpaceTokenObjList(inTextList);
List <string> testStrList = new List <string>();
testStrList.Add("diagnosed");
testStrList.Add("diagnose");
testStrList.Add("dianosed");
// init context
int tarPos = 6;
int tarSize = 1;
int radius = 2;
bool word2VecSkipWord = true;
bool debugFlag = false;
DoubleVec contextVec = Word2VecContext.GetContextVec(tarPos, tarSize, nonSpaceTokenList, w2vIm, radius, word2VecSkipWord, debugFlag);
Console.WriteLine("===== Test diagnosed|diagnose|dianosed (window-2) =====");
Console.WriteLine("inText: [" + inText + "]");
Console.WriteLine("============================================");
Console.WriteLine("Candidates|CBOW score|CBOW score 2|Similarity score");
Console.WriteLine("============================================");
foreach (string testStr in testStrList)
{
Test(testStr, contextVec, w2vIm, w2vOm);
}
Console.WriteLine("===== Test diagnosed|diagnose|dianosed (whole text) =====");
contextVec = Word2VecContext.GetContextVec(tarPos, tarSize, nonSpaceTokenList, w2vIm, word2VecSkipWord, debugFlag);
foreach (string testStr in testStrList)
{
Test(testStr, contextVec, w2vIm, w2vOm);
}
string inText1 = "Not all doctors know about this syndrome.";
List <TokenObj> inTextList1 = TextObj.TextToTokenList(inText1);
// remove space token from the list
List <TokenObj> nonSpaceTokenList1 = TextObj.GetNonSpaceTokenObjList(inTextList1);
Console.WriteLine("===== Test know about|know|about (window) =====");
List <string> testStrList1 = new List <string>();
testStrList1.Add("know about");
testStrList1.Add("know");
testStrList1.Add("about");
tarPos = 3;
tarSize = 2;
radius = 2;
contextVec = Word2VecContext.GetContextVec(tarPos, tarSize, nonSpaceTokenList1, w2vIm, radius, word2VecSkipWord, debugFlag);
Test(testStrList1[0], contextVec, w2vIm, w2vOm);
contextVec = Word2VecContext.GetContextVec(tarPos, tarSize, nonSpaceTokenList1, w2vIm, word2VecSkipWord, debugFlag);
Test(testStrList1[0], contextVec, w2vIm, w2vOm);
tarPos = 3;
tarSize = 1;
contextVec = Word2VecContext.GetContextVec(tarPos, tarSize, nonSpaceTokenList1, w2vIm, radius, word2VecSkipWord, debugFlag);
Test(testStrList1[1], contextVec, w2vIm, w2vOm);
contextVec = Word2VecContext.GetContextVec(tarPos, tarSize, nonSpaceTokenList1, w2vIm, word2VecSkipWord, debugFlag);
Test(testStrList1[1], contextVec, w2vIm, w2vOm);
tarPos = 4;
tarSize = 1;
contextVec = Word2VecContext.GetContextVec(tarPos, tarSize, nonSpaceTokenList1, w2vIm, radius, word2VecSkipWord, debugFlag);
Test(testStrList1[2], contextVec, w2vIm, w2vOm);
contextVec = Word2VecContext.GetContextVec(tarPos, tarSize, nonSpaceTokenList1, w2vIm, word2VecSkipWord, debugFlag);
Test(testStrList1[2], contextVec, w2vIm, w2vOm);
string inText2 = "for the last 10 years was diagnosed early on set dementia 3 years ago.";
List <TokenObj> inTextList2 = TextObj.TextToTokenList(inText2);
// remove space token from the list
List <TokenObj> nonSpaceTokenList2 = TextObj.GetNonSpaceTokenObjList(inTextList2);
List <string> testStrList2 = new List <string>();
testStrList2.Add("onset");
testStrList2.Add("on set");
Console.WriteLine("===== Test onset|on set (window-3) =====");
Console.WriteLine("inText: [" + inText + "]");
tarPos = 8;
tarSize = 2;
radius = 3;
contextVec = Word2VecContext.GetContextVec(tarPos, tarSize, nonSpaceTokenList2, w2vIm, radius, word2VecSkipWord, debugFlag);
foreach (string testStr in testStrList2)
{
Test(testStr, contextVec, w2vIm, w2vOm);
}
tarPos = 8;
tarSize = 1;
radius = 3;
contextVec = Word2VecContext.GetContextVec(tarPos, tarSize, nonSpaceTokenList2, w2vIm, radius, word2VecSkipWord, debugFlag);
Test("on", contextVec, w2vIm, w2vOm);
tarPos = 9;
contextVec = Word2VecContext.GetContextVec(tarPos, tarSize, nonSpaceTokenList2, w2vIm, radius, word2VecSkipWord, debugFlag);
Test("set", contextVec, w2vIm, w2vOm);
Console.WriteLine("===== Test onset|on set (whole text) =====");
radius = nonSpaceTokenList2.Count;
contextVec = Word2VecContext.GetContextVec(tarPos, tarSize, nonSpaceTokenList2, w2vIm, word2VecSkipWord, debugFlag);
foreach (string testStr in testStrList2)
{
Test(testStr, contextVec, w2vIm, w2vOm);
}
Console.WriteLine("===== Go through each tokens with diff radius 1-9) =====");
Console.WriteLine("tarPos|tarWord|r=1|r=2|r=3|r=4|r=5|r=6|r=7|r=8|r=9");
//String inText3 = "Broken bones can not sleep at night!";
string inText3 = "not xyxy all doctors know about this syndrome.";
List <TokenObj> inTextList3 = TextObj.TextToTokenList(inText3);
// remove space token from the list
List <TokenObj> nonSpaceTokenList3 = TextObj.GetNonSpaceTokenObjList(inTextList3);
tarPos = 0;
tarSize = 1;
radius = 0;
foreach (TokenObj tokenObj in nonSpaceTokenList3)
{
// skip the space token
string tokenStr = tokenObj.GetTokenStr();
string inStr = Word2VecContext.NormWordForWord2Vec(tokenStr);
Console.Write(tarPos + "|" + tokenStr + "|");
// print out all radius
for (int r = 1; r < 10; r++)
{
contextVec = Word2VecContext.GetContextVec(tarPos, tarSize, inTextList2, w2vIm, r, word2VecSkipWord, debugFlag);
TestWin(inStr, contextVec, w2vIm, w2vOm);
}
Console.WriteLine("");
tarPos++;
}
}
// private methods
// this test is not verified
private static int RunTest(bool detailFlag, int tarPos, int tarSize, int contextRadius, long limitNo)
{
// init dic
string configFile = "../data/Config/cSpell.properties";
CSpellApi cSpellApi = new CSpellApi(configFile);
cSpellApi.SetRankMode(CSpellApi.RANK_MODE_CONTEXT);
Word2Vec word2VecIm = cSpellApi.GetWord2VecIm();
Word2Vec word2VecOm = cSpellApi.GetWord2VecOm();
bool word2VecSkipWord = cSpellApi.GetWord2VecSkipWord();
ContextScoreComparator <ContextScore> csc = new ContextScoreComparator <ContextScore>();
// provide cmdLine interface
int returnValue = 0;
try {
StreamReader stdInput = new StreamReader(Console.OpenStandardInput());
try {
string inText = null;
Console.WriteLine("- Please input a text, only a spell error allowed (type \"Ctl-d\" to quit) > ");
while (!string.ReferenceEquals((inText = stdInput.ReadLine()), null))
{
// ---------------------------------
// Get spell correction on the input
// ---------------------------------
// convert input text to TokenObj
TextObj textObj = new TextObj(inText);
List <TokenObj> inTextList = textObj.GetTokenList();
// *2 because tokenList include space
string tarWord = inTextList[tarPos * 2].GetTokenStr();
for (int i = 1; i < tarSize; i++)
{
int ii = (tarPos + 1) * 2;
tarWord += " " + inTextList[ii].GetTokenStr();
}
Console.WriteLine("- input text: [" + inText + "]");
Console.WriteLine("- target: [" + tarPos + "|" + tarSize + "|" + tarWord + "]");
Console.WriteLine("- context radius: " + contextRadius);
// get all possible candidates
HashSet <string> candSet = NonWord1To1Candidates.GetCandidates(tarWord, cSpellApi);
candSet.Add(tarWord); // add the original word
Console.WriteLine("-- canSet.size(): " + candSet.Count);
// get final suggestion
// remove space token from the list
List <TokenObj> nonSpaceTokenList = TextObj.GetNonSpaceTokenObjList(inTextList);
string topRankStr = GetTopRankStr(tarWord, candSet, tarPos, tarSize, nonSpaceTokenList, word2VecIm, word2VecOm, word2VecSkipWord, contextRadius, detailFlag);
Console.WriteLine("- top rank str: " + topRankStr);
// print details
if (detailFlag == true)
{
HashSet <ContextScore> candScoreSet = GetCandidateScoreSet(candSet, tarPos, tarSize, inTextList, word2VecIm, word2VecOm, word2VecSkipWord, contextRadius, detailFlag);
Console.WriteLine("------ Suggestion List ------");
var list = candScoreSet.OrderBy(x => x, csc).Take((int)limitNo).Select(obj => obj.ToString());
foreach (var item in list)
{
Console.WriteLine(item);
}
}
// print the prompt
Console.WriteLine("- Please input a text, only a spell error allowed (type \"Ctl-d\" to quit) > ");
}
} catch (Exception e2) {
Console.Error.WriteLine(e2.Message);
returnValue = -1;
}
} catch (Exception e) {
Console.Error.WriteLine(e.Message);
returnValue = -1;
}
return(returnValue);
}