/// <summary> /// This is the algorithm to find the Shortest Middle Snake (SMS). /// </summary> /// <param name="dataA">sequence A</param> /// <param name="lowerA">lower bound of the actual range in DataA</param> /// <param name="upperA">upper bound of the actual range in DataA (exclusive)</param> /// <param name="dataB">sequence B</param> /// <param name="lowerB">lower bound of the actual range in DataB</param> /// <param name="upperB">upper bound of the actual range in DataB (exclusive)</param> /// <param name="downVector">a vector for the (0,0) to (x,y) search. Passed as a parameter for speed reasons.</param> /// <param name="upVector">a vector for the (u,v) to (N,M) search. Passed as a parameter for speed reasons.</param> /// <returns>a MiddleSnakeData record containing x,y and u,v</returns> private ShortestMiddleSnakeReturnData ShortestMiddleSnake( DiffData dataA, int lowerA, int upperA, DiffData dataB, int lowerB, int upperB, int[] downVector, int[] upVector) { var ret = new ShortestMiddleSnakeReturnData(); int max = dataA.Length + dataB.Length + 1; int downK = lowerA - lowerB; // the k-line to start the forward search int upK = upperA - upperB; // the k-line to start the reverse search int delta = (upperA - lowerA) - (upperB - lowerB); bool oddDelta = (delta & 1) != 0; // The vectors in the publication accepts negative indexes. the vectors implemented here are 0-based // and are access using a specific offset: UpOffset UpVector and DownOffset for DownVektor int downOffset = max - downK; int upOffset = max - upK; int maxD = ((upperA - lowerA + upperB - lowerB) / 2) + 1; // Debug.Write(2, "SMS", String.Format("Search the box: A[{0}-{1}] to B[{2}-{3}]", LowerA, UpperA, LowerB, UpperB)); // init vectors downVector[downOffset + downK + 1] = lowerA; upVector[upOffset + upK - 1] = upperA; for (int d = 0; d <= maxD; d++) { // Extend the forward path. for (int k = downK - d; k <= downK + d; k += 2) { // Debug.Write(0, "SMS", "extend forward path " + k.ToString()); // find the only or better starting point int x; if (k == downK - d) { x = downVector[downOffset + k + 1]; // down } else { x = downVector[downOffset + k - 1] + 1; // a step to the right if ((k < downK + d) && (downVector[downOffset + k + 1] >= x)) { x = downVector[downOffset + k + 1]; // down } } int y = x - k; // find the end of the furthest reaching forward D-path in diagonal k. while ((x < upperA) && (y < upperB) && (dataA.Data[x] == dataB.Data[y])) { x++; y++; } downVector[downOffset + k] = x; // overlap ? if (oddDelta && (upK - d < k) && (k < upK + d)) { if (upVector[upOffset + k] <= downVector[downOffset + k]) { ret.X = downVector[downOffset + k]; ret.Y = downVector[downOffset + k] - k; return(ret); } } } // Extend the reverse path. for (int k = upK - d; k <= upK + d; k += 2) { // Debug.Write(0, "SMS", "extend reverse path " + k.ToString()); // find the only or better starting point int x, y; if (k == upK + d) { x = upVector[upOffset + k - 1]; // up } else { x = upVector[upOffset + k + 1] - 1; // left if ((k > upK - d) && (upVector[upOffset + k - 1] < x)) { x = upVector[upOffset + k - 1]; // up } } y = x - k; while ((x > lowerA) && (y > lowerB) && (dataA.Data[x - 1] == dataB.Data[y - 1])) { x--; y--; // diagonal } upVector[upOffset + k] = x; // overlap ? if (!oddDelta && (downK - d <= k) && (k <= downK + d)) { if (upVector[upOffset + k] <= downVector[downOffset + k]) { ret.X = downVector[downOffset + k]; ret.Y = downVector[downOffset + k] - k; return(ret); } } } } throw new Exception("the algorithm should never come here."); }
/// <summary> /// This is the algorithm to find the Shortest Middle Snake (SMS). /// </summary> /// <param name="dataA">sequence A</param> /// <param name="lowerA">lower bound of the actual range in DataA</param> /// <param name="upperA">upper bound of the actual range in DataA (exclusive)</param> /// <param name="dataB">sequence B</param> /// <param name="lowerB">lower bound of the actual range in DataB</param> /// <param name="upperB">upper bound of the actual range in DataB (exclusive)</param> /// <param name="downVector">a vector for the (0,0) to (x,y) search. Passed as a parameter for speed reasons.</param> /// <param name="upVector">a vector for the (u,v) to (N,M) search. Passed as a parameter for speed reasons.</param> /// <returns>a MiddleSnakeData record containing x,y and u,v</returns> private ShortestMiddleSnakeReturnData ShortestMiddleSnake( DiffData dataA, int lowerA, int upperA, DiffData dataB, int lowerB, int upperB, int[] downVector, int[] upVector) { var ret = new ShortestMiddleSnakeReturnData(); int max = dataA.Length + dataB.Length + 1; int downK = lowerA - lowerB; // the k-line to start the forward search int upK = upperA - upperB; // the k-line to start the reverse search int delta = (upperA - lowerA) - (upperB - lowerB); bool oddDelta = (delta & 1) != 0; // The vectors in the publication accepts negative indexes. the vectors implemented here are 0-based // and are access using a specific offset: UpOffset UpVector and DownOffset for DownVektor int downOffset = max - downK; int upOffset = max - upK; int maxD = ((upperA - lowerA + upperB - lowerB) / 2) + 1; // Debug.Write(2, "SMS", String.Format("Search the box: A[{0}-{1}] to B[{2}-{3}]", LowerA, UpperA, LowerB, UpperB)); // init vectors downVector[downOffset + downK + 1] = lowerA; upVector[upOffset + upK - 1] = upperA; for (int d = 0; d <= maxD; d++) { // Extend the forward path. for (int k = downK - d; k <= downK + d; k += 2) { // Debug.Write(0, "SMS", "extend forward path " + k.ToString()); // find the only or better starting point int x; if (k == downK - d) { x = downVector[downOffset + k + 1]; // down } else { x = downVector[downOffset + k - 1] + 1; // a step to the right if ((k < downK + d) && (downVector[downOffset + k + 1] >= x)) { x = downVector[downOffset + k + 1]; // down } } int y = x - k; // find the end of the furthest reaching forward D-path in diagonal k. while ((x < upperA) && (y < upperB) && (dataA.Data[x] == dataB.Data[y])) { x++; y++; } downVector[downOffset + k] = x; // overlap ? if (oddDelta && (upK - d < k) && (k < upK + d)) { if (upVector[upOffset + k] <= downVector[downOffset + k]) { ret.X = downVector[downOffset + k]; ret.Y = downVector[downOffset + k] - k; return ret; } } } // Extend the reverse path. for (int k = upK - d; k <= upK + d; k += 2) { // Debug.Write(0, "SMS", "extend reverse path " + k.ToString()); // find the only or better starting point int x, y; if (k == upK + d) { x = upVector[upOffset + k - 1]; // up } else { x = upVector[upOffset + k + 1] - 1; // left if ((k > upK - d) && (upVector[upOffset + k - 1] < x)) { x = upVector[upOffset + k - 1]; // up } } y = x - k; while ((x > lowerA) && (y > lowerB) && (dataA.Data[x - 1] == dataB.Data[y - 1])) { x--; y--; // diagonal } upVector[upOffset + k] = x; // overlap ? if (!oddDelta && (downK - d <= k) && (k <= downK + d)) { if (upVector[upOffset + k] <= downVector[downOffset + k]) { ret.X = downVector[downOffset + k]; ret.Y = downVector[downOffset + k] - k; return ret; } } } } throw new Exception("the algorithm should never come here."); }