From 09241865ab8b15892117a931be7031ac6d399714 Mon Sep 17 00:00:00 2001
From: hkb <hiamnshubarnwal26@gmail.com>
Date: Mon, 14 Oct 2019 23:28:58 +0530
Subject: [PATCH] String based advance Algorithms

---
 Algorithms/String/Btree.cpp                   | 158 ++++++++++++++++++
 .../Finite_Automata_String_Matching.cpp       |  53 ++++++
 Algorithms/String/KMPsearch.cpp               |  61 +++++++
 Algorithms/String/Rabin-Karpsearch.cpp        |  34 ++++
 Algorithms/String/naivesearch.cpp             |  23 +++
 Algorithms/String/substring_find.cpp          |  16 ++
 6 files changed, 345 insertions(+)
 create mode 100644 Algorithms/String/Btree.cpp
 create mode 100644 Algorithms/String/Finite_Automata_String_Matching.cpp
 create mode 100644 Algorithms/String/KMPsearch.cpp
 create mode 100644 Algorithms/String/Rabin-Karpsearch.cpp
 create mode 100644 Algorithms/String/naivesearch.cpp
 create mode 100644 Algorithms/String/substring_find.cpp

diff --git a/Algorithms/String/Btree.cpp b/Algorithms/String/Btree.cpp
new file mode 100644
index 00000000..16ba6437
--- /dev/null
+++ b/Algorithms/String/Btree.cpp
@@ -0,0 +1,158 @@
+#include<iostream>
+#include<conio.h>
+using namespace std;
+int d;
+struct btnode
+{
+	int *key;
+	btnode **child;
+};
+void print(btnode* t)
+{
+	if(t==NULL) return;
+	int i;
+	for(i=0;i<2*d;i++)
+	{
+		print(t->child[i]);
+		if(t->key[i]!=-1)
+		cout<<t->key[i]<<" ";
+	}
+    print(t->child[2*d]);
+}
+bool check_leaf(btnode* T)
+{
+	int flag=0;
+	for(int i=0;i<2*d+1;i++)
+	if(T->child[i]!=NULL)
+	flag=1;
+	if(flag==0)
+	return (1);
+	else
+	return (0);
+}
+void add_to_node(btnode* old_node,int k,btnode* &new_node,int pos)
+{
+	int i,j;
+	for(i=0;i<2*d&&old_node->key[i]!=-1;i++);
+	for(j=i-1;j>=pos&&old_node->key[j]>k;j--)
+	{
+		old_node->key[j+1]=old_node->key[j];
+		old_node->child[j+2]=old_node->child[j+1];
+	}
+	old_node->key[pos]=k;
+	old_node->child[pos+1]=new_node;
+}
+void create_newnode(btnode* &T)
+{
+	int i;
+	T=new btnode;
+	T->key=new int[2*d];
+	T->child=new btnode*[2*d+1];
+	for(i=0;i<2*d;i++) T->key[i]=-1;
+	for(i=0;i<2*d+1;i++) T->child[i]=NULL;
+}
+void nodesplit(btnode* old_node,btnode* &new_node,int &k,int pos)
+{
+	btnode* N=NULL;
+	create_newnode(N);
+	if(pos<=d)
+	{
+		for(int i=d;i<2*d;i++)
+		{
+			N->key[i-d]=old_node->key[i];
+			N->child[i-d+1]=old_node->child[i+1];
+			old_node->key[i]=-1;
+			old_node->child[i+1]=NULL;
+		}
+		N->child[0]=old_node->child[d];
+		add_to_node(old_node,k,new_node,pos);
+	}
+	else
+	{
+		for(int i=d+1;i<2*d;i++)
+		{
+			N->key[i-(d+1)]=old_node->key[i];
+			N->child[i-d]=old_node->child[i+1];
+			old_node->key[i]=-1;
+			old_node->child[i+1]=NULL;
+		}
+		N->child[0]=old_node->child[d+1];
+        old_node->child[d+1]=NULL;
+		add_to_node(N,k,new_node,pos-(d+1));
+	}
+	k=old_node->key[d];
+	old_node->key[d]=-1;
+	new_node=N;
+}
+int setflag(btnode* T,int &k,btnode* &new_node)
+{
+	int pos;
+	if(T==NULL)
+	{
+		return (1);
+	}
+	for(pos=0;pos<2*d&&T->key[pos]!=-1;pos++)
+	{
+		if(k<T->key[pos]) break;
+	}
+	if(check_leaf(T))
+	{
+		if(T->key[2*d-1]==-1)
+		{
+			add_to_node(T,k,new_node,pos);
+			return (0);
+		}
+		else
+		{
+			nodesplit(T,new_node,k,pos);
+			return (1);
+		}
+	}
+	else
+	{
+		int flag=setflag(T->child[pos],k,new_node);
+		if(flag==1)
+		{
+			if(T->key[2*d-1]==-1)
+			{
+				add_to_node(T,k,new_node,pos);
+				return (0);
+			}
+		    else
+		    {
+			    nodesplit(T,new_node,k,pos);
+			    return (1);
+		    }
+		}
+	}
+}
+btnode* create_root(btnode* T,btnode* new_node,int k)
+{
+	btnode* root=NULL;
+        create_newnode(root);
+	    root->key[0]=k;
+		root->child[0]=T;
+		root->child[1]=new_node;
+		return (root);
+}
+void insert(btnode* &T,int k,btnode* new_node)
+{
+	btnode* root=NULL;
+	int flag=setflag(T,k,new_node);
+	if(flag) T=create_root(T,new_node,k);
+}
+int main()
+{
+	btnode *root=NULL;
+	cout<<"enter d\n";
+	cin>>d;
+	cout<<"enter elements to be inserted \n";
+	int next;
+	while(cin>>next)
+	{
+		if(next==-1) break;
+		else insert(root,next,NULL);
+	}
+	print(root);
+	return (0);
+}
diff --git a/Algorithms/String/Finite_Automata_String_Matching.cpp b/Algorithms/String/Finite_Automata_String_Matching.cpp
new file mode 100644
index 00000000..b80afd65
--- /dev/null
+++ b/Algorithms/String/Finite_Automata_String_Matching.cpp
@@ -0,0 +1,53 @@
+#include<bits/stdc++.h>
+using namespace std;
+const int characters=256;
+int getNextState(char pat[],int M,int i,int x){
+	if(i<M && pat[i]==x){
+		return i+1;
+	}
+	for(int j=i;j>0;j--){
+		if(pat[j-1]==x){
+			int s=0;
+			for(s=0;s<j-1;s++){
+				if(pat[s]!=pat[i-j+1+s]){
+					break;
+				}
+			}
+			if(s==j-1){
+				return j;
+			}
+		}
+	}
+	return 0;
+}
+void computefa(char pat[],int M,int fa[][characters]){
+	for(int i=0;i<=M;i++){
+		for(int ch=0;ch<256;ch++){
+			fa[i][ch]=getNextState(pat,M,i,ch);
+		}
+	}
+}
+int main() 
+{ 
+    char txt[]="AABAACAADAABAAABAA";
+    char pat[]="ABAA";
+    int N=strlen(txt);
+    int M=strlen(pat);
+    int fa[M+1][characters];
+    computefa(pat,M,fa);
+    for(int i=0;i<=M;i++){
+    	for(int j=0;j<characters;j++){
+    		cout<<fa[i][j]<<" ";
+		}
+		cout<<endl;
+	}
+	cout<<endl<<endl;
+	int state=0;
+	for(int i=0;i<N;i++){
+		state=fa[state][txt[i]];
+		if(state==M){
+			cout<<i-M+1<<endl;
+		}
+	}
+    return 0;
+} 
diff --git a/Algorithms/String/KMPsearch.cpp b/Algorithms/String/KMPsearch.cpp
new file mode 100644
index 00000000..cf9ee939
--- /dev/null
+++ b/Algorithms/String/KMPsearch.cpp
@@ -0,0 +1,61 @@
+#include<bits/stdc++.h>
+using namespace std;
+
+void KMPsearch(char txt[],char pat[],int lps[],int M){
+	int i=0,j=0;
+	int N=strlen(txt);
+	while(i<N){
+		if(txt[i]!=pat[j]){
+			if(j!=0)
+                j=lps[j-1];
+            else
+                i++;
+		}
+		else if(txt[i]==pat[j]){
+			i++;
+			j++;
+		}
+		if(j==M){
+			cout<<i-j<<endl;
+			j==lps[j-1];
+		}
+		
+	}
+}
+
+void computeLPSArray(char pat[],int M,int lps[]){
+	int len=0;
+	lps[0]=0;
+	int i=1;
+	while(i<M){
+		if(pat[i]==pat[len]){
+			len++;
+			lps[i]=len;
+			i++;
+		}
+		else{
+			if(len==0){
+				lps[i]=0;
+				i++;
+			}
+			else{
+				len=lps[len-1];
+			}
+		}
+	}
+}
+
+int main() 
+{ 
+	int M=9;
+	int lps[M];
+    char txt[]="ABABDABACDABABCABAB";
+    char pat[]="ABABCABAB";
+    computeLPSArray(pat,M,lps);
+    for(int i=0;i<M;i++){
+    	cout<<lps[i]<<" ";
+	}
+	cout<<endl;
+	KMPsearch(txt,pat,lps,M);
+    return 0; 
+} 
diff --git a/Algorithms/String/Rabin-Karpsearch.cpp b/Algorithms/String/Rabin-Karpsearch.cpp
new file mode 100644
index 00000000..b1fe4ebf
--- /dev/null
+++ b/Algorithms/String/Rabin-Karpsearch.cpp
@@ -0,0 +1,34 @@
+#include<bits/stdc++.h>
+using namespace std;
+int main(){
+	char txt[]="ABABDABACDABABCABAB";
+    char pat[]="ABABCABAB";
+    int M=strlen(pat);
+    int N=strlen(txt);
+    int i=0,j=0;
+    int h=1,d=256,q=101;
+    for(i=0;i<M-1;i++){
+        h=(h*d)%q;
+    }
+    int p=0,t=0;
+    for(i=0;i<M;i++){
+        p=(d*p+pat[i])%q;
+        t=(d*t+txt[i])%q;
+    }
+    for(i=0;i<=N-M;i++){
+        if(p==t){
+            for(j=0;j<M;j++){
+                if(txt[i+j]!=pat[j])
+                	break;
+            }
+            if(j==M){
+                cout<<i<<endl;
+        	}        
+        }
+        if(i<N-M ){
+            t=(d*(t-txt[i]*h)+txt[i+M])%q;
+            if(t<0)
+            	t=(t+q);
+		}
+    }
+}
diff --git a/Algorithms/String/naivesearch.cpp b/Algorithms/String/naivesearch.cpp
new file mode 100644
index 00000000..56646e35
--- /dev/null
+++ b/Algorithms/String/naivesearch.cpp
@@ -0,0 +1,23 @@
+#include<bits/stdc++.h>
+using namespace std;
+int main(){
+	char txt[]="ABABDABACDABABCABAB";
+    char pat[]="ABABCABAB";
+    int M=strlen(pat);
+    int N=strlen(txt);
+    int i=0,j=0,l=0;
+    while(l<N){
+    	if(txt[i]==pat[j]){
+    		i++;
+    		j++;
+		}
+		else{
+			l++;
+			i=l;
+			j=0;
+		}
+		if(j==M){
+			cout<<i-j<<endl;
+		}
+	}
+}
diff --git a/Algorithms/String/substring_find.cpp b/Algorithms/String/substring_find.cpp
new file mode 100644
index 00000000..5708b35a
--- /dev/null
+++ b/Algorithms/String/substring_find.cpp
@@ -0,0 +1,16 @@
+#include<stdio.h>
+int main(){
+	char str[]="hello world";
+	int i=0,j=0;
+	char substr[]="or";
+	int m=0,n=0;
+	
+	for(i=0;str[i]!='\0';i++){
+		for(j=0;substr[j]!='\0' && str[i+j]==substr[j];j++){
+			if(i>=2){
+				printf("%d ",i);
+			}
+		}
+	}
+	return 0;
+}