/// <summary>Gets a copy of this instance.</summary>
/// <returns></returns>
public Similars Copy()
{
Similars copySmls = new Similars(Count);
Array.Copy(_numbers, copySmls._numbers, Count);
return(copySmls);
}
/// <summary>Gets inizializeted similar words for given seed word.</summary>
/// <param name="seedNumber">Serial number of seed word.</param>
/// <returns>The total nubmer of filled similars.</returns>
/// <remarks>In Turbo mode getting Similars are saved and returned from memory by requery.</remarks>
public static Similars GetSimilars(int seedNumber)
{
if (_MismatchCount == 0)
{
return(null);
}
_Word = Words.GetWord(seedNumber);
int sum = 0;
if (_IsTurbo)
{
Similars smls = _ArrSimilars[seedNumber];
if (smls == null)
{
// create and fill Similars
_ArrSimilars[seedNumber] = smls = new Similars(_Capacity);
Fill_recurs(0, _MismatchCount, seedNumber, smls._numbers, ref sum, true);
}
return(smls);
}
else
{
Fill_recurs(0, _MismatchCount, seedNumber, _Similars._numbers, ref sum, true);
return(_Similars);
}
}
/// <summary>Initializes static members.</summary>
/// <param name="mismatchCount">The number of acceptable mismatches</param>
/// <param name="isTurbo">True if Turbo mode is defined.</param>
/// <param name="isDifferentWordLength">True if word's length is differ with previous session.</param>
public static void Init(byte mismatchCount, bool isTurbo, bool isDifferentWordLength)
{
_Capacity = GetCapacity(Words.WordLength, mismatchCount);
bool isDiffer = mismatchCount != _MismatchCount || isDifferentWordLength;
if (_IsTurbo = isTurbo)
{
// save filling Similars in case of the same MismatchCount & WordLength
_Similars = null;
if (isDiffer || _ArrSimilars == null)
{
_ArrSimilars = new Similars[Words.Count];
}
}
else
{
_ArrSimilars = null;
if (isDiffer || _Similars == null)
{
_Similars = new Similars(_Capacity);
}
}
_MismatchCount = mismatchCount;
_DepthOfClastering = DoC_FACTOR * Words.WordLength;
}
/// <summary>Reports the shrinked (without zero Frequency) Similars of the specified pattern.</summary>
/// <param name="seedNumber">The number of the seed-word.</param>
/// <returns>If Similars have zero-frequency items, the shrinked (without zero Frequency) Similars; otherwise initial Similars.</returns>
public Similars GetSimilars(int seedNumber)
{
Similars smls = Similars.GetSimilars(seedNumber);
bool isNotCopied = true;
// We have to remove from the total Similars all the numbers with zero Frequency
int j;
bool isFound = true;
for (int i = 0; i < smls.Count; i++)
{
// scan throw the whole collection to check for real existing similars
for (j = 0; j < Count; j++)
{
if (isFound = this[j].SeedNumber == smls[i])
{
break;
}
}
if (!isFound)
{
// if zero-frequence element is finded, Similars should be copied
// because it change its Count by shrinking
if (isNotCopied)
{
smls = smls.Copy();
isNotCopied = false;
}
smls[i] = Similars.UndefNumber;
}
}
return(isNotCopied ? smls : smls.Shrink());
}
/// <summary>Initialezas static member.</summary>
/// <param name="wordLength">The length of word.</param>
/// <param name="mismatchCount">The number of acceptable mismatches</param>
public static void Init(byte wordLength, byte mismatchCount)
{
_wordLength = wordLength;
_wordLastIndex = (byte)(wordLength - 1);
_count = (int)Math.Pow(Chars.Count, wordLength);
Similars.Init(mismatchCount);
// inizialise words
_wordsTrain = new byte[_count * wordLength];
SetRank_recurs(0, 1);
}
/// <summary>Initialezas static member.</summary>
/// <param name="wordLength">The length of word.</param>
/// <param name="mismatchCount">The number of acceptable mismatches</param>
/// <param name="isTurbo">True if Turbo mode is defined.</param>
public static void Init(byte wordLength, byte mismatchCount, bool isTurbo)
{
bool isDifferentLength = _wordLength != wordLength;
if (isDifferentLength)
{
_wordLength = wordLength;
_wordLastIndex = (byte)(wordLength - 1);
_count = GetCount(wordLength);
}
Similars.Init(mismatchCount, isTurbo, isDifferentLength); // _wordLength & _count are used in that method, so it could be the last
}
/// <summary>
/// Fills referenced array by the indexes of similar words of given seed word an returns number of indexes.
/// </summary>
/// <param name="words">Words sequence.</param>
/// <param name="wrdIndex">The index of the seed word in the words sequence.</param>
/// <param name="smlIndexes">Array of indexes of similar words.</param>
/// <returns></returns>
public static int Fill(byte[] words, int wrdIndex, ref Similars smlIndexes)
{
bool[] lockedLetters = new bool[Words.WordLength]; // array of signs indicated wich letter's position is locked
int currSmlIndex = 0;
if (_MismatchCount > 0)
{
FillUnder(ref words, wrdIndex, ref lockedLetters, 0, _MismatchCount, ref smlIndexes, ref currSmlIndex);
}
return(smlIndexes._count = currSmlIndex);
}
/// <summary>
/// Fills referenced array by the numbers of similar words for given seed word and returns number of similars.
/// </summary>
/// <param name="word">Seed word.</param>
/// <param name="smlIndexes">External uninitialised array of numbers of similar words.</param>
/// <returns>The total nubmer of filled similars.</returns>
public static int Fill(byte[] word, ref Similars smlIndexes)
{
_word = word;
_smlNumbers = smlIndexes;
_currSmlIndex = 0;
if (_MismatchCount > 0)
{
FillUnder_recurs(0, _MismatchCount);
}
return(smlIndexes._count = _currSmlIndex);
}
///// <summary>Removes all the elements with the zero Frequency from the instance.</summary>
///// <remarks>Sorts the collection by Frequency.</remarks>
//void Shrink ()
//{
// if (_isShrinked == true)
// return;
// Sort();
// //System.Console.Beep(400, 100);
// // start from the end because zero Frequencies are much more less then unzero
// for (int i = Count-1; i >= 0; i--)
// //if (this[i].Freq > 0)
// if (this[i] > 0)
// {
// Array.Resize(ref _coll, _insCount = i+1);
// break;
// }
// _isShrinked = true;
//}
#endregion
#region Statistic methods
/// <summary>Reports the shrinked (without zero Frequency) Similars of the specified unsorted pattern.</summary>
/// <param name="seedNumber">The number of the seed-word in the collection.</param>
/// <returns>If Similars have zero-frequency items, the shrinked (without zero Frequency) Similars; otherwise initial Similars.</returns>
public Similars GetSimilars(int seedNumber)
{
Similars smls = Similars.GetSimilars(seedNumber);
bool isNotCopied = true;
// We have to remove from the total Similars all the numbers with zero Frequency
// this is a variant used in ScanWindows
// we may use direct indexing to check zero-frequency seeds
for (int i = 0; i < smls.Count; i++)
{
if (this[smls[i]] == 0)
{
// if zero-frequence element is finded, Similars should be copied
// because it change its Count by shrinking
if (isNotCopied)
{
smls = smls.Copy();
isNotCopied = false;
}
smls[i] = Similars.UndefNumber;
}
}
return(isNotCopied ? smls : smls.Shrink());
}
/// <summary>Adds to referenced array the indexes of similar words of given seed word.</summary>
/// <param name="words">Words sequence.</param>
/// <param name="wrdIndex">The index of the seed word in the word's sequence.</param>
/// <param name="lockedLetters">Array of signs indicated wich letter's position is locked on this step.</param>
/// <param name="startPos">Started letter's position in seed word at wich search begins</param>
/// <param name="mismatchCnt">Number of mismathes.</param>
/// <param name="smlIndexes">Array of indexes of similar words.</param>
/// <param name="currSmlIndex">Current writing position in array of indexes of similar words.</param>
static void FillUnder(ref byte[] words, int wrdIndex, ref bool[] lockedLetters, int startPos, int mismatchCnt, ref Similars smlIndexes, ref int currSmlIndex)
{
byte j, originLetter, currLetter;
int ind, i;
mismatchCnt--; // reduce count of mismatches because of processing current mismatch's level
for (i = startPos; i < Words.WordLength; i++)
{
if (!lockedLetters[i]) // pass locked letter
{
ind = i + wrdIndex;
originLetter = words[ind]; // save original letter wich would be changed
// substitute other letters
for (j = 0; j < Chars.Count; j++)
{
//currLetter = Chars.GetCodeByIndex(j);
currLetter = j;
if (currLetter != originLetter)
{
words[ind] = currLetter; // substitute other letter
smlIndexes[currSmlIndex++] = Words.GetNumber(words, wrdIndex); // save words index
if (mismatchCnt > 0) // lock letter and pass down
{
lockedLetters[i] = true; // lock letter
FillUnder(ref words, wrdIndex, ref lockedLetters, i + 1, mismatchCnt, ref smlIndexes, ref currSmlIndex);
lockedLetters[i] = false; // unlock letter
}
}
}
words[ind] = originLetter; // restore original letter
}
}
mismatchCnt++; // restore count of mismatches
}
/// <summary>Initializes an instance of the Patterns.</summary>
/// <param name="worker">The BackgroundWorker in wich Processing is run.</param>
/// <param name="sequence">The 'names' sequence.</param>
/// <param name="startIndex">The index in the inputArray at which scanning begins.</param>
/// <param name="strictLength">The precise length of range of inputArray for scanning.</param>
/// <param name="strictFreqs">The external array of strict frequencies; needed to avoid memory overallotment.</param>
/// <param name="loopsPerStep">The amount of loops per one progress bar step; -1 if without progress bar steps.</param>
/// <param name="maxFreqNumber">Return number of Pattern with max Frequence: if init value = -1, then no return.</param>
/// <param name="isFirstCall">True if method is called first time in cycle; otherwise, false.</param>
/// <param name="isDrawPlot">True if plot is drawing; otherwise, false.</param>
/// <returns>Maximum frequency value in the instance; if init maxFreqNumber param = -1, then 0.</returns>
int Init(
BackgroundWorker worker, byte[] sequence,
int startIndex, int strictLength, int[] strictFreqs, int loopsPerStep, ref int maxFreqNumber,
bool isFirstCall, bool isDrawPlot)
{
#region Comments
/*
* Patterns are generated from inputArray sequence on 2 steps.
*
* Step 1: Initialization strict frequency
* We create a new collection with the maximum possible length and an temporary array of strict frequencies.
* Index of each element for both collections is equal to absolute word number,
* so we scan through the input sequence and increase strict frequency in array directly by indexing.
*
* Step 2: Initialization Frequency
* From each pattern in collection we added the Frequency of all his Similars.
* In ALL POSSIBLE mode we search all the patterns, because their Similars may have Frequency > 0;
* in FROM INPUT mode we skip zero-frequency patterns.
*
* In TURBO mode all the Similars are keeping in memory after first generation,
* so we should go through the given Similars collection and adding thier Frequency directly by word number, equls to index in this instance.
* In ORDINARY mode we receive the sum of Frequencies of all Similars through single call of Similars.SumUp().
*
* So we receive the frequencies with mismatches for all the Patterns in this instance.
*
* This is the maximum fastest method because of using only direct indexing.
* We scan input sequence just once? and no searches through collection are required.
*/
#endregion
//System.Windows.Forms.MessageBox.Show(loopsPerStep.ToString());
int i, maxFreq = 0;
worker.ReportProgress(TreatSCANNING);
OnPrgBarReseat();
this.Clear();
Array.Clear(strictFreqs, 0, strictFreqs.Length);
//System.Windows.Forms.MessageBox.Show(startIndex.ToString("startIndex= 0 ") + strictLength.ToString("strictLength= 0 "));
//*** Step_1: Look over all the patterns in a range of sequence
for (i = 0; i <= strictLength; i++)
{
strictFreqs[Words.GetNumber(sequence, startIndex + i)]++;
}
//*** Step_2: expanding to mismatches
int k;
Similars smls;
//Pattern ptn;
// compare each Pattern to all the Pattern's in the same collection with a glance of mismathes;
// collection is still unsorted
for (i = 0; i < Count; i++)
{
if (worker.CancellationPending) // Canceled by user.
{
throw new ApplicationException(string.Empty);
}
if (IsAllPossible || (strictFreqs[i] > 0))
{
//ptn = this[i];
if (Similars.IsTurbo)
{
smls = Similars.GetSimilars(i);
for (k = 0; k < smls.Count; k++)
{
//ptn.Freq += strictFreqs[smls[k]];
this[i] += strictFreqs[smls[k]];
}
}
else
{
//ptn.Freq += Similars.SumUp(i, strictFreqs);
this[i] += Similars.SumUp(i, strictFreqs);
}
//ptn.WordNumber = i; // initialize number
//ptn.Freq += strictFreqs[i]; // frequency itself
this[i] += strictFreqs[i]; // frequency itself
//this[i] = ptn;
// keep index of maximum Freq without sorting: needs only in case of Max Average Choice or Global task.
//if (maxFreqNumber >= 0 && ptn.Freq > maxFreq)
if (maxFreqNumber >= 0 && this[i] > maxFreq)
{
//maxFreq = ptn.Freq;
maxFreq = this[i];
maxFreqNumber = i;
}
}
OnPrgBarIncreased(worker, loopsPerStep);
}
if (isDrawPlot)
{
// if it needs to keep sorting by name, we should clone collection for drawing,
// because it will be sorted by Freq during drawing
worker.ReportProgress(TreatSORTING);
worker.ReportProgress(Convert.ToInt32(isFirstCall), Clone());
}
return(maxFreq);
}
