update ch5

AceCoooool · AceCoooool · commit a9cf3dd75ed1 · 2018-07-09T16:37:02.000+08:00
diff --git a/README.md b/README.md
@@ -112,12 +112,15 @@ Working directory: youroot/algs4cplusplus/ch1
 
 ## ch5. Strings
 
-|                             REF                             |                           PROGRAM                            |    DESCRIPTION / C++DOC    |                             REF                             |                           PROGRAM                            |       DESCRIPTION / C++DOC        |
-| :---------------------------------------------------------: | :----------------------------------------------------------: | :------------------------: | :---------------------------------------------------------: | :----------------------------------------------------------: | :-------------------------------: |
-|  [-](https://algs4.cs.princeton.edu/50strings/index.php#-)  | [Alphabet.java](https://algs4.cs.princeton.edu/50strings/Alphabet.java.html) |          alphabet          |  [-](https://algs4.cs.princeton.edu/50strings/index.php#-)  | [Count.java](https://algs4.cs.princeton.edu/50strings/Count.java.html) |          alphabet client          |
-| [5.1](https://algs4.cs.princeton.edu/51radix/index.php#5.1) |                   [LSD.h](ch5/head/LSD.h)                    |       LSD radix sort       | [5.2](https://algs4.cs.princeton.edu/51radix/index.php#5.2) |                   [MSD.h](ch5/head/MSD.h)                    |          MSD radix sort           |
-|   [-](https://algs4.cs.princeton.edu/51radix/index.php#-)   |            [InplaceMSD.h](ch5/head/InplaceMSD.h)             |  In-place MSD radix sort1  | [5.3](https://algs4.cs.princeton.edu/51radix/index.php#5.3) |          [Quick3string.h](ch5/head/Quick3string.h)           |      3-way string quicksort       |
-|   [-](https://algs4.cs.princeton.edu/51radix/index.php#-)   |          [AmericanFlag.h](ch5/head/AmericanFlag.h)           |    American flag sort1     |   [-](https://algs4.cs.princeton.edu/51radix/index.php#-)   |         [AmericanFlagX.h](ch5/head/AmericanFlagX.h)          | non-recursive American flag sort1 |
-| [5.4](https://algs4.cs.princeton.edu/52trie/index.php#5.4)  |                [TrieST.h](ch5/head/TrieST.h)                 | multiway trie symbol table |   [-](https://algs4.cs.princeton.edu/52trie/index.php#-)    |               [TrieSET.h](ch5/head/TrieSET.h)                |         multiway trie set         |
-| [5.5](https://algs4.cs.princeton.edu/52trie/index.php#5.5)  |                 [TST.h](ch5/head/TrieSET.h)                  |    ternary search trie     |                                                             |                                                              |                                   |
+|                             REF                              |                           PROGRAM                            |      DESCRIPTION / C++DOC      |                             REF                              |                           PROGRAM                            |         DESCRIPTION / C++DOC          |
+| :----------------------------------------------------------: | :----------------------------------------------------------: | :----------------------------: | :----------------------------------------------------------: | :----------------------------------------------------------: | :-----------------------------------: |
+|  [-](https://algs4.cs.princeton.edu/50strings/index.php#-)   | [Alphabet.java](https://algs4.cs.princeton.edu/50strings/Alphabet.java.html) |            alphabet            |  [-](https://algs4.cs.princeton.edu/50strings/index.php#-)   | [Count.java](https://algs4.cs.princeton.edu/50strings/Count.java.html) |            alphabet client            |
+| [5.1](https://algs4.cs.princeton.edu/51radix/index.php#5.1)  |                   [LSD.h](ch5/head/LSD.h)                    |         LSD radix sort         | [5.2](https://algs4.cs.princeton.edu/51radix/index.php#5.2)  |                   [MSD.h](ch5/head/MSD.h)                    |            MSD radix sort             |
+|   [-](https://algs4.cs.princeton.edu/51radix/index.php#-)    |            [InplaceMSD.h](ch5/head/InplaceMSD.h)             |    In-place MSD radix sort1    | [5.3](https://algs4.cs.princeton.edu/51radix/index.php#5.3)  |          [Quick3string.h](ch5/head/Quick3string.h)           |        3-way string quicksort         |
+|   [-](https://algs4.cs.princeton.edu/51radix/index.php#-)    |          [AmericanFlag.h](ch5/head/AmericanFlag.h)           |      American flag sort1       |   [-](https://algs4.cs.princeton.edu/51radix/index.php#-)    |         [AmericanFlagX.h](ch5/head/AmericanFlagX.h)          |   non-recursive American flag sort1   |
+|  [5.4](https://algs4.cs.princeton.edu/52trie/index.php#5.4)  |                [TrieST.h](ch5/head/TrieST.h)                 |   multiway trie symbol table   |    [-](https://algs4.cs.princeton.edu/52trie/index.php#-)    |               [TrieSET.h](ch5/head/TrieSET.h)                |           multiway trie set           |
+|  [5.5](https://algs4.cs.princeton.edu/52trie/index.php#5.5)  |                 [TST.h](ch5/head/TrieSET.h)                  |      ternary search trie       | [5.6](https://algs4.cs.princeton.edu/53substring/index.php#5.6) |                   [KMP.h](ch5/head/KMP.h)                    | substring search (Knuth–Morris–Pratt) |
+| [5.7](https://algs4.cs.princeton.edu/53substring/index.php#5.7) |            [BoyerMoore.h](ch5/head/BoyerMoore.h)             | substring search (Boyer–Moore) | [5.8](https://algs4.cs.princeton.edu/53substring/index.php#5.8) |             [RabinKarp.h](ch5/head/RabinKarp.h)              |     substring search (Rabin–Karp)     |
+| [5.9](https://algs4.cs.princeton.edu/54regexp/index.php#5.9) |                   [NFA.h](ch5/head/NFA.h)                    |  NFA for regular expressions   |   [-](https://algs4.cs.princeton.edu/54regexp/index.php#-)   |               [GREP.cpp](ch5/16_GREP/main.cpp)               |                 grep                  |
+| [-](https://algs4.cs.princeton.edu/55compression/index.php#-) |         [BinaryDump.cpp](ch5/17_BinaryDump/main.cpp)         |          binary dump           |                                                              |                                                              |                                       |
 
diff --git a/ch5/10_TrieSET/CMakeLists.txt b/ch5/10_TrieSET/CMakeLists.txt
@@ -3,5 +3,5 @@ project(10_TrieSET)
 
 set(CMAKE_CXX_STANDARD 14)
 
-set(SOURCE_FILES main.cpp ../head/LSD.h ../head/MSD.h ../head/TST.h ../head/AmericanFlag.h ../head/AmericanFlagX.h ../head/KMP.h ../head/BoyerMoore.h ../head/RabinKarp.h)
+set(SOURCE_FILES main.cpp ../head/LSD.h ../head/MSD.h ../head/TST.h ../head/AmericanFlag.h ../head/AmericanFlagX.h ../head/KMP.h ../head/BoyerMoore.h ../head/RabinKarp.h ../head/NFA.h)
 add_executable(10_TrieSET ${SOURCE_FILES})
diff --git a/ch5/15_NFA/CMakeLists.txt b/ch5/15_NFA/CMakeLists.txt
@@ -0,0 +1,7 @@
+cmake_minimum_required(VERSION 3.8)
+project(15_NFA)
+
+set(CMAKE_CXX_STANDARD 14)
+
+set(SOURCE_FILES main.cpp ../head/LSD.h ../head/MSD.h ../head/TST.h)
+add_executable(15_NFA ${SOURCE_FILES})
diff --git a/ch5/15_NFA/main.cpp b/ch5/15_NFA/main.cpp
@@ -0,0 +1,17 @@
+#include "../head/NFA.h"
+#include <iostream>
+
+using namespace std;
+
+/**
+ * Unit tests the {@code NFA} data type.
+ *
+ * @param args the command-line arguments
+ */
+int main() {
+    string regexp = "(A*B|AC)D";
+    string txt = "AAAACD";
+
+    NFA nfa(regexp);
+    cout << nfa.recognizes(txt) << endl;
+}
diff --git a/ch5/16_GREP/CMakeLists.txt b/ch5/16_GREP/CMakeLists.txt
@@ -0,0 +1,7 @@
+cmake_minimum_required(VERSION 3.8)
+project(16_GREP)
+
+set(CMAKE_CXX_STANDARD 14)
+
+set(SOURCE_FILES main.cpp ../head/LSD.h ../head/MSD.h ../head/TST.h)
+add_executable(16_GREP ${SOURCE_FILES})
diff --git a/ch5/16_GREP/main.cpp b/ch5/16_GREP/main.cpp
@@ -0,0 +1,24 @@
+#include "../head/NFA.h"
+#include <iostream>
+
+using namespace std;
+
+/**
+ * Interprets the command-line argument as a regular expression
+ * (supporting closure, binary or, parentheses, and wildcard)
+ * reads in lines from standard input; writes to standard output
+ * those lines that contain a substring matching the regular
+ * expression.
+ *
+ * @param args the command-line arguments
+ */
+int main() {
+    ifstream file("./data/tinyL.txt");
+    string regexp = "(.*(A*B|AC)D.*)";
+    string line;
+    NFA nfa(regexp);
+    while (file >> line) {
+        if (nfa.recognizes(line))
+            cout << line << endl;
+    }
+}
diff --git a/ch5/17_BinaryDump/CMakeLists.txt b/ch5/17_BinaryDump/CMakeLists.txt
@@ -0,0 +1,7 @@
+cmake_minimum_required(VERSION 3.8)
+project(17_BinaryDump)
+
+set(CMAKE_CXX_STANDARD 14)
+
+set(SOURCE_FILES main.cpp ../head/LSD.h ../head/MSD.h ../head/TST.h)
+add_executable(17_BinaryDump ${SOURCE_FILES})
diff --git a/ch5/17_BinaryDump/main.cpp b/ch5/17_BinaryDump/main.cpp
@@ -0,0 +1,29 @@
+#include <iostream>
+#include <fstream>
+
+using namespace std;
+
+/**
+ * Reads in a sequence of bytes from standard input and writes
+ * them to standard output in binary, k bits per line,
+ * where k is given as a command-line integer (defaults
+ * to 16 if no integer is specified); also writes the number
+ * of bits.
+ *
+ * @param args the command-line arguments
+ */
+int main() {
+    ifstream file("./data/abra.txt", ios::binary);
+    int bitsPerLine = 16;
+    char tmp;
+    int cnt = 0;
+    while (file >> tmp) {
+        for (int i = 7; i >= 0; --i) {
+            cout << ((tmp >> i) & 1);
+            cnt++;
+            if (cnt % bitsPerLine == 0 && cnt != 0)cout << endl;
+        }
+    }
+    if (bitsPerLine != 0) cout << endl;
+    cout << cnt << " bits" << endl;
+}
diff --git a/ch5/CMakeLists.txt b/ch5/CMakeLists.txt
@@ -18,6 +18,9 @@ add_subdirectory(11_TST)
 add_subdirectory(12_KMP)
 add_subdirectory(13_BoyerMoore)
 add_subdirectory(14_RabinKarp)
+add_subdirectory(15_NFA)
+add_subdirectory(16_GREP)
+add_subdirectory(17_BinaryDump)
 
 
-add_subdirectory(temp)
+# add_subdirectory(temp)
diff --git a/ch5/data/abra.txt b/ch5/data/abra.txt
@@ -0,0 +1 @@
+ABRACADABRA!
diff --git a/ch5/data/tinyL.txt b/ch5/data/tinyL.txt
@@ -0,0 +1,9 @@
+AC
+AD
+AAA
+ABD
+ADD
+BCD
+ABCCBD
+BABAAA
+BABBAAA
diff --git a/ch5/head/Digraph.h b/ch5/head/Digraph.h
@@ -0,0 +1,225 @@
+#ifndef CH4_DIGRAPH_H
+#define CH4_DIGRAPH_H
+
+#include <forward_list>
+#include <vector>
+#include <stdexcept>
+#include <fstream>
+#include <string>
+#include <stack>
+#include <ostream>
+#include <sstream>
+
+using std::forward_list;
+using std::vector;
+using std::runtime_error;
+using std::string;
+using std::fstream;
+using std::to_string;
+using std::stack;
+using std::distance;
+using std::ostream;
+using std::stringstream;
+
+/**
+ *  The {@code Digraph} class represents a directed graph of vertices
+ *  named 0 through <em>V</em> - 1.
+ *  It supports the following two primary operations: add an edge to the digraph,
+ *  iterate over all of the vertices adjacent from a given vertex.
+ *  Parallel edges and self-loops are permitted.
+ *  <p>
+ *  This implementation uses an adjacency-lists representation, which
+ *  is a vertex-indexed array of {@link Bag} objects.
+ *  All operations take constant time (in the worst case) except
+ *  iterating over the vertices adjacent from a given vertex, which takes
+ *  time proportional to the number of such vertices.
+ *  <p>
+ *  For additional documentation,
+ *  see <a href="https://algs4.cs.princeton.edu/42digraph">Section 4.2</a> of
+ *  <i>Algorithms, 4th Edition</i> by Robert Sedgewick and Kevin Wayne.
+ *
+ *  @author Robert Sedgewick
+ *  @author Kevin Wayne
+ */
+class Digraph {
+public:
+    /**
+     * Initializes an empty digraph with <em>V</em> vertices.
+     *
+     * @param  V the number of vertices
+     * @throws IllegalArgumentException if {@code V < 0}
+     */
+    Digraph(int V) : V(V), E(0), indegree(V), adj(V) {
+        if (V < 0) throw runtime_error("Number of vertices in a Digraph must be nonnegative");
+    }
+
+    /**
+     * Initializes a digraph from the specified input stream.
+     * The format is the number of vertices <em>V</em>,
+     * followed by the number of edges <em>E</em>,
+     * followed by <em>E</em> pairs of vertices, with each entry separated by whitespace.
+     *
+     * @param  in the input stream
+     * @throws IllegalArgumentException if the endpoints of any edge are not in prescribed range
+     * @throws IllegalArgumentException if the number of vertices or edges is negative
+     * @throws IllegalArgumentException if the input stream is in the wrong format
+     */
+    Digraph(string in): E(0) {
+        fstream file(in);
+        file >> V;
+        if (V < 0) throw runtime_error("number of vertices in a Digraph must be nonnegative");
+        indegree.resize(V);
+        adj.resize(V);
+        int tmpE;
+        file >> tmpE;
+        if (tmpE < 0) throw runtime_error("number of edges in a Digraph must be nonnegative");
+        int v, w;
+        for (int i = 0; i < tmpE; i++) {
+            file >> v >> w;
+            addEdge(v, w);
+        }
+    }
+
+    /**
+     * Initializes a new digraph that is a deep copy of the specified digraph.
+     *
+     * @param  G the digraph to copy
+     */
+    Digraph(const Digraph &G) : Digraph(G.getV()) {
+        E = G.getE();
+        for (int v = 0; v < V; v++)
+            indegree[v] = G.getindegree(v);
+        for (int v = 0; v < G.getV(); v++) {
+            // reverse so that adjacency list is in same order as original
+            stack<int> reverse;
+            for (int w : G.adj[v]) {
+                reverse.push(w);
+            }
+            while (!reverse.empty()) {
+                auto w = reverse.top();
+                reverse.pop();
+                adj[v].push_front(w);
+            }
+        }
+    }
+
+    /**
+     * Returns the number of vertices in this digraph.
+     *
+     * @return the number of vertices in this digraph
+     */
+    int getV() const {
+        return V;
+    }
+
+    /**
+     * Returns the number of edges in this digraph.
+     *
+     * @return the number of edges in this digraph
+     */
+    int getE() const {
+        return E;
+    }
+
+    /**
+    * Adds the directed edge v→w to this digraph.
+    *
+    * @param  v the tail vertex
+    * @param  w the head vertex
+    * @throws IllegalArgumentException unless both {@code 0 <= v < V} and {@code 0 <= w < V}
+    */
+    void addEdge(int v, int w) {
+        validateVertex(v);
+        validateVertex(w);
+        adj[v].push_front(w);
+        indegree[w]++;
+        E++;
+    }
+
+    /**
+     * Returns the vertices adjacent from vertex {@code v} in this digraph.
+     *
+     * @param  v the vertex
+     * @return the vertices adjacent from vertex {@code v} in this digraph, as an iterable
+     * @throws IllegalArgumentException unless {@code 0 <= v < V}
+     */
+    forward_list<int> getadj(int v) const {
+        validateVertex(v);
+        return adj[v];
+    }
+
+    /**
+     * Returns the number of directed edges incident from vertex {@code v}.
+     * This is known as the <em>outdegree</em> of vertex {@code v}.
+     *
+     * @param  v the vertex
+     * @return the outdegree of vertex {@code v}
+     * @throws IllegalArgumentException unless {@code 0 <= v < V}
+     */
+    int outdegree(int v) const {
+        validateVertex(v);
+        return distance(adj[v].begin(), adj[v].end());
+    }
+
+    /**
+     * Returns the number of directed edges incident to vertex {@code v}.
+     * This is known as the <em>indegree</em> of vertex {@code v}.
+     *
+     * @param  v the vertex
+     * @return the indegree of vertex {@code v}
+     * @throws IllegalArgumentException unless {@code 0 <= v < V}
+     */
+    int getindegree(int v) const {
+        validateVertex(v);
+        return indegree[v];
+    }
+
+    /**
+     * Returns the reverse of the digraph.
+     *
+     * @return the reverse of the digraph
+     */
+    Digraph *reverse() {
+        Digraph *reverse = new Digraph(V);
+        for (int v = 0; v < V; v++) {
+            for (int w : getadj(v)) {
+                reverse->addEdge(w, v);
+            }
+        }
+        return reverse;
+    }
+
+    /**
+     * Returns a string representation of the graph.
+     *
+     * @return the number of vertices <em>V</em>, followed by the number of edges <em>E</em>,
+     *         followed by the <em>V</em> adjacency lists
+     */
+    friend ostream &operator<<(ostream &stream, const Digraph &dg) {
+        stringstream ss;
+        ss << dg.V << " vertices, " << dg.getE() << " edges" << std::endl;
+        for (int v = 0; v < dg.V; ++v) {
+            ss << v << ": ";
+            for (auto w: dg.getadj(v))
+                ss << w << " ";
+            ss << std::endl;
+        }
+        stream << ss.str();
+        return stream;
+    }
+
+private:
+    // throw an IllegalArgumentException unless {@code 0 <= v < V}
+    void validateVertex(int v) const {
+        if (v < 0 || v >= V)
+            throw runtime_error("vertex " + to_string(v) + " is not between 0 and " + to_string(V - 1));
+    }
+
+private:
+    int V;                         // number of vertices in this digraph
+    int E;                         // number of edges in this digraph
+    vector<forward_list<int>> adj; // adj[v] = adjacency list for vertex v
+    vector<int> indegree;          // indegree[v] = indegree of vertex v
+};
+
+#endif //CH4_DIGRAPH_H
diff --git a/ch5/head/DirectedDFS.h b/ch5/head/DirectedDFS.h
diff --git a/ch5/head/NFA.h b/ch5/head/NFA.h
