From 56d2a0ea243ed8411271d5fdb97517bf3fb9936b Mon Sep 17 00:00:00 2001
From: Petr Wudi <petr.wudi@gmail.com>
Date: Fri, 17 Nov 2017 20:26:02 +0100
Subject: [PATCH] Split Needleman-Wunsch into two classes to ease
implementation of Smith-Waterman
---
.../AbstractSequenceComparator.java | 183 ++++++++++++++++++
.../NeedlemanWunsch.java | 175 +----------------
.../SequenceComparator.java | 13 ++
3 files changed, 201 insertions(+), 170 deletions(-)
create mode 100644 src/main/java/edu/cvut/fit/kw/vmm/backend/sequences_comparement/AbstractSequenceComparator.java
create mode 100644 src/main/java/edu/cvut/fit/kw/vmm/backend/sequences_comparement/SequenceComparator.java
diff --git a/src/main/java/edu/cvut/fit/kw/vmm/backend/sequences_comparement/AbstractSequenceComparator.java b/src/main/java/edu/cvut/fit/kw/vmm/backend/sequences_comparement/AbstractSequenceComparator.java
new file mode 100644
index 0000000..5423829
--- /dev/null
+++ b/src/main/java/edu/cvut/fit/kw/vmm/backend/sequences_comparement/AbstractSequenceComparator.java
@@ -0,0 +1,183 @@
+package edu.cvut.fit.kw.vmm.backend.sequences_comparement;
+
+import edu.cvut.fit.kw.vmm.alignment_solution.AlignmentSolution;
+
+import java.io.PrintStream;
+
+/**
+ * Contains methods and variables useful for both Needleman-Wunsch and Smith-Waterman algorithms
+ */
+abstract class AbstractSequenceComparator implements SequenceComparator {
+
+ protected static final int DISTANCE_PENALTY = -1;
+ protected int[][] scores;
+ protected NeedlemanWunsch.TableDirection[][] directions;
+ protected String str0;
+ protected String str1;
+
+ protected AbstractSequenceComparator(String str0, String str1) {
+ this.str0 = str0;
+ this.str1 = str1;
+ }
+
+ /**
+ * Solves the problem and finds similarity
+ * @return best score
+ */
+ public AlignmentSolution solve() {
+ initScoreMatrix();
+ initDirectionMatrix();
+ for(int y = 0; y < str0.length(); y++) {
+ for(int x = 0; x < str1.length(); x++) {
+ solveItem(y,x);
+ }
+ }
+ return getAlignmentSolution();
+ }
+
+ /**
+ * Computes item of the scores and stores it. It also stores how this value were achieved (which direction came the
+ * value from) in direction matrix.
+ * @param pos0 position of cahr in fist string
+ * @param pos1 position of char in second string
+ * @return score of the computed solution
+ */
+
+ protected int solveItem(int pos0, int pos1) {
+ char char0 = str0.charAt(pos0);
+ char char1 = str1.charAt(pos1);
+ int scoreTop = scores[pos0][pos1 + 1] + DISTANCE_PENALTY;
+ int scoreLeft = scores[pos0 + 1][pos1] + DISTANCE_PENALTY;
+ int scoreDiagonal = scores[pos0][pos1] + getSimilarity(char0, char1);
+ int bestScore = scoreLeft;
+ TableDirection bestDirection = TableDirection.HORIZONTAL;
+ if(scoreDiagonal >= scoreLeft && scoreDiagonal >= scoreTop) {
+ bestScore = scoreDiagonal;
+ bestDirection = TableDirection.DIAGONAL;
+ }
+ else if(scoreTop >= scoreDiagonal && scoreTop >= scoreLeft) {
+ bestScore = scoreTop;
+ bestDirection = TableDirection.VERTICAL;
+ }
+ scores[pos0 + 1][pos1 + 1] = bestScore;
+ directions[pos0 + 1][pos1 + 1] = bestDirection;
+ return bestScore;
+ }
+
+ /**
+ * Initializes matrix of partial similarity scores.
+ * Allocates it and fills first row and column.
+ */
+ protected abstract void initScoreMatrix();
+
+ /**
+ * Initializes directions of steps in score matrix
+ * Allocates it and fills first row and column.
+ */
+ private void initDirectionMatrix() {
+ directions = new NeedlemanWunsch.TableDirection[str0.length() + 1][str1.length() + 1];
+ directions[0][0] = null;
+ for(int i = 1; i <= str0.length(); i++) {
+ directions[i][0] = NeedlemanWunsch.TableDirection.VERTICAL;
+ }
+ for(int i = 1; i <= str1.length(); i++) {
+ directions[0][i] = NeedlemanWunsch.TableDirection.HORIZONTAL;
+ }
+ }
+ /**
+ * Generates Alignment solution instance from the solution stored in the matrices.
+ * @return Alignment solution of the two strings (those which are set up during class construction)
+ */
+ protected abstract AlignmentSolution getAlignmentSolution();
+
+ /**
+ * Prints score and direction matrices to stdout (useful for debugging)
+ */
+ public void printMatrices() {
+ printScoreMatrix();
+ printDirectionMatrix();
+ }
+
+ /**
+ * Prints score matrix to stdout (useful for debugging)
+ */
+ public void printScoreMatrix() {
+ PrintStream stream = System.out;
+ // Print header (first string)
+ stream.format(" ");
+ for(int i = 0; i < str1.length(); i++) {
+ stream.format("%5c", str1.charAt(i));
+ }
+ stream.println();
+ // Print actual matrix
+ for(int y = 0; y <= str0.length(); y++) {
+ if(y != 0) {
+ stream.format("%c", str0.charAt(y - 1));
+ }
+ else {
+ stream.format(" ");
+ }
+ for(int x = 0; x <= str1.length(); x++) {
+ stream.format("%4d ", scores[y][x]);
+ }
+ stream.println();
+ }
+ }
+
+ /**
+ * Prints score matrix to stdout (useful for debugging)
+ */
+ public void printDirectionMatrix() {
+ PrintStream stream = System.out;
+ // Print header (first string)
+ stream.format(" ");
+ for(int i = 0; i < str1.length(); i++) {
+ stream.format("%c ", str1.charAt(i));
+ }
+ stream.println();
+ // Print actual matrix
+ for(int y = 0; y <= str0.length(); y++) {
+ if(y != 0) {
+ stream.format("%c", str0.charAt(y - 1));
+ }
+ else {
+ stream.format(" ");
+ }
+ for(int x = 0; x <= str1.length(); x++) {
+ NeedlemanWunsch.TableDirection direction = directions[y][x];
+ if (direction == null) {
+ stream.format(" ");
+ }
+ else{
+ switch (direction) {
+ case DIAGONAL:
+ stream.format("\\ ");
+ break;
+ case VERTICAL:
+ stream.format("| ");
+ break;
+ case HORIZONTAL:
+ stream.format("– ");
+ break;
+ }
+ }
+ }
+ stream.println();
+ }
+ }
+
+ /**
+ * Returns similarity score between two specified characters
+ * @param a first character to compare
+ * @param b another character to compare
+ * @return similarity - positive or negative value or zero
+ */
+ protected int getSimilarity(char a, char b) {
+ return a == b ? 1 : -1;
+ }
+
+ protected enum TableDirection {
+ HORIZONTAL, VERTICAL, DIAGONAL
+ }
+
+}
diff --git a/src/main/java/edu/cvut/fit/kw/vmm/backend/sequences_comparement/NeedlemanWunsch.java b/src/main/java/edu/cvut/fit/kw/vmm/backend/sequences_comparement/NeedlemanWunsch.java
index bb800e3..5501cae 100644
--- a/src/main/java/edu/cvut/fit/kw/vmm/backend/sequences_comparement/NeedlemanWunsch.java
+++ b/src/main/java/edu/cvut/fit/kw/vmm/backend/sequences_comparement/NeedlemanWunsch.java
@@ -3,71 +3,16 @@ package edu.cvut.fit.kw.vmm.backend.sequences_comparement;
import edu.cvut.fit.kw.vmm.alignment_solution.AlignmentSolution;
import edu.cvut.fit.kw.vmm.alignment_solution.StringOperation;
-import java.io.PrintStream;
import java.util.LinkedList;
import java.util.List;
-public class NeedlemanWunsch {
-
- private static final int DISTANCE_PENALTY = -1;
- private int[][] scores;
- private TableDirection[][] directions;
- private String str0;
- private String str1;
+public class NeedlemanWunsch extends AbstractSequenceComparator implements SequenceComparator {
public NeedlemanWunsch(String str0, String str1) {
- this.str0 = str0;
- this.str1 = str1;
- }
-
- /**
- * Solves the problem and finds simliarity
- * @return best score
- */
- public AlignmentSolution solve() {
- initScoreMatrix();
- initDirectionMatrix();
- for(int y = 0; y < str0.length(); y++) {
- for(int x = 0; x < str1.length(); x++) {
- solveItem(y,x);
- }
- }
- return getAlignmentSolution();
+ super(str0, str1);
}
- /**
- * Computes item of the scores and stores it. It also stores how this value were achieved (which direction came the
- * value from) in direction matrix.
- * @param pos0 position of cahr in fist string
- * @param pos1 position of char in second string
- * @return score of the computed solution
- */
- private int solveItem(int pos0, int pos1) {
- char char0 = str0.charAt(pos0);
- char char1 = str1.charAt(pos1);
- int scoreTop = scores[pos0][pos1 + 1] + DISTANCE_PENALTY;
- int scoreLeft = scores[pos0 + 1][pos1] + DISTANCE_PENALTY;
- int scoreDiagonal = scores[pos0][pos1] + getSimilarity(char0, char1);
- int bestScore = scoreLeft;
- TableDirection bestDirection = TableDirection.HORIZONTAL;
- if(scoreDiagonal >= scoreLeft && scoreDiagonal >= scoreTop) {
- bestScore = scoreDiagonal;
- bestDirection = TableDirection.DIAGONAL;
- }
- else if(scoreTop >= scoreDiagonal && scoreTop >= scoreLeft) {
- bestScore = scoreTop;
- bestDirection = TableDirection.VERTICAL;
- }
- scores[pos0 + 1][pos1 + 1] = bestScore;
- directions[pos0 + 1][pos1 + 1] = bestDirection;
- return bestScore;
- }
-
- /**
- * Initializes matrix of partial similarity scores.
- * Allocates it and fills first row and column.
- */
- private void initScoreMatrix() {
+ protected void initScoreMatrix() {
scores = new int[str0.length() + 1][str1.length() + 1];
scores[0][0] = 0;
for(int i = 1; i <= str0.length(); i++) {
@@ -78,26 +23,7 @@ public class NeedlemanWunsch {
}
}
- /**
- * Initializes directions of steps in score matrix
- * Allocates it and fills first row and column.
- */
- private void initDirectionMatrix() {
- directions = new TableDirection[str0.length() + 1][str1.length() + 1];
- directions[0][0] = null;
- for(int i = 1; i <= str0.length(); i++) {
- directions[i][0] = TableDirection.VERTICAL;
- }
- for(int i = 1; i <= str1.length(); i++) {
- directions[0][i] = TableDirection.HORIZONTAL;
- }
- }
-
- /**
- * Generates Alignment solution instance from the solution stored in the matrices.
- * @return Alignment solution of the two strings (those which are set up during class construction)
- */
- private AlignmentSolution getAlignmentSolution() {
+ protected AlignmentSolution getAlignmentSolution() {
int bestScore = scores[str0.length()][str1.length()];
List<StringOperation> operations = getStringTransformations();
return new AlignmentSolution(operations, bestScore);
@@ -112,7 +38,7 @@ public class NeedlemanWunsch {
int y = str0.length();
int x = str1.length();
while(y > 0 || x > 0) {
- TableDirection direction = directions[y][x];
+ NeedlemanWunsch.TableDirection direction = directions[y][x];
switch (direction) {
case DIAGONAL:
if(str0.charAt(y - 1) == str1.charAt(x - 1)) {
@@ -137,95 +63,4 @@ public class NeedlemanWunsch {
return operations;
}
- /**
- * Prints score and direction matrices to stdout (useful for debugging)
- */
- public void printMatrices() {
- printScoreMatrix();
- printDirectionMatrix();
- }
-
- /**
- * Prints score matrix to stdout (useful for debugging)
- */
- public void printScoreMatrix() {
- PrintStream stream = System.out;
- // Print header (first string)
- stream.format(" ");
- for(int i = 0; i < str1.length(); i++) {
- stream.format("%5c", str1.charAt(i));
- }
- stream.println();
- // Print actual matrix
- for(int y = 0; y <= str0.length(); y++) {
- if(y != 0) {
- stream.format("%c", str0.charAt(y - 1));
- }
- else {
- stream.format(" ");
- }
- for(int x = 0; x <= str1.length(); x++) {
- stream.format("%4d ", scores[y][x]);
- }
- stream.println();
- }
- }
-
- /**
- * Prints score matrix to stdout (useful for debugging)
- */
- public void printDirectionMatrix() {
- PrintStream stream = System.out;
- // Print header (first string)
- stream.format(" ");
- for(int i = 0; i < str1.length(); i++) {
- stream.format("%c ", str1.charAt(i));
- }
- stream.println();
- // Print actual matrix
- for(int y = 0; y <= str0.length(); y++) {
- if(y != 0) {
- stream.format("%c", str0.charAt(y - 1));
- }
- else {
- stream.format(" ");
- }
- for(int x = 0; x <= str1.length(); x++) {
- TableDirection direction = directions[y][x];
- if (direction == null) {
- stream.format(" ");
- }
- else{
- switch (direction) {
- case DIAGONAL:
- stream.format("\\ ");
- break;
- case VERTICAL:
- stream.format("| ");
- break;
- case HORIZONTAL:
- stream.format("– ");
- break;
- }
- }
- }
- stream.println();
- }
- }
-
-
- /**
- * Returns similarity score between two specified characters
- * @param a first character to compare
- * @param b another character to compare
- * @return similarity - positive or negative value or zero
- */
- private int getSimilarity(char a, char b) {
- return a == b ? 1 : -1;
- }
-
- protected enum TableDirection {
- HORIZONTAL, VERTICAL, DIAGONAL
- }
-
}
diff --git a/src/main/java/edu/cvut/fit/kw/vmm/backend/sequences_comparement/SequenceComparator.java b/src/main/java/edu/cvut/fit/kw/vmm/backend/sequences_comparement/SequenceComparator.java
new file mode 100644
index 0000000..731fc57
--- /dev/null
+++ b/src/main/java/edu/cvut/fit/kw/vmm/backend/sequences_comparement/SequenceComparator.java
@@ -0,0 +1,13 @@
+package edu.cvut.fit.kw.vmm.backend.sequences_comparement;
+
+import edu.cvut.fit.kw.vmm.alignment_solution.AlignmentSolution;
+
+/**
+ * Compares sequences and aligns them. The sequences to compare are passed as constructor parameter or any another way
+ * (it depends on the actual class).
+ */
+public interface SequenceComparator {
+
+ AlignmentSolution solve();
+
+}
--
GitLab