From 42cde7fd3a6d487f7cce4009bd39a1772e3fdf3e Mon Sep 17 00:00:00 2001
From: shellfly <shell0fly@gmail.com>
Date: Thu, 6 Feb 2020 20:09:34 +0800
Subject: [PATCH] add msd

---
 README.md    |  4 ++--
 algs4/lsd.py | 24 +++++++++-----------
 algs4/msd.py | 63 ++++++++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 76 insertions(+), 15 deletions(-)
 create mode 100644 algs4/msd.py

diff --git a/README.md b/README.md
index df58731..ed8f62c 100644
--- a/README.md
+++ b/README.md
@@ -79,9 +79,9 @@ Try to keep the interface and variable name consistent with the original book wh
     * [BellmanFordSP](algs4/bellman_ford_sp.py)
 
 * 5 STRING
-
   * [LSD](algs4/lsd.py)
-
+  * [MSD](algs4/msd.py)
+  * [Quick3string](algs4/quick3_string.py)
 ## License
 
 This code is released under MIT.
diff --git a/algs4/lsd.py b/algs4/lsd.py
index 064a210..e53455b 100644
--- a/algs4/lsd.py
+++ b/algs4/lsd.py
@@ -1,9 +1,9 @@
 """
-   Execution:  python lsd.py input.txt
+   Execution:  python lsd.py < input.txt
 
    Data files:   https://algs4.cs.princeton.edu/51radix/words3.txt
 
-   % python lsd.py words3.txt
+   % python lsd.py < words3.txt
    all
    bad
    bed
@@ -17,17 +17,16 @@
 
 
 class LSD:
-
+    R = 256
     @classmethod
     def sort(cls, a, w):
         n = len(a)
-        R = 256
         aux = ['' for _ in range(n)]
         for d in range(w-1, -1, -1):
-            count = [0 for _ in range(R+1)]
+            count = [0 for _ in range(cls.R+1)]
             for i in range(n):
                 count[ord(a[i][d])+1] += 1
-            for r in range(R):
+            for r in range(cls.R):
                 count[r+1] += count[r]
             for i in range(n):
                 aux[count[ord(a[i][d])]] = a[i]
@@ -35,13 +34,12 @@ def sort(cls, a, w):
             for i in range(n):
                 a[i] = aux[i]
 
+
 if __name__ == '__main__':
     import sys
-    lst = []
-    with open(sys.argv[1]) as fp:
-        for line in fp:
-            for x in line.split(' '):
-                lst.append(x.strip())
-    LSD.sort(lst, len(lst[0]))
-    for item in lst:
+    words = []
+    for line in sys.stdin:
+        words.extend(line.split())
+    LSD.sort(words, len(words[0]))
+    for item in words:
         print(item)
diff --git a/algs4/msd.py b/algs4/msd.py
new file mode 100644
index 0000000..c8d6b5d
--- /dev/null
+++ b/algs4/msd.py
@@ -0,0 +1,63 @@
+"""
+   Execution:  python msd.py < input.txt
+
+   Data files:   https://algs4.cs.princeton.edu/51radix/words3.txt
+
+   % python msd.py < words3.txt
+   all
+   bad
+   bed
+   bug
+   dad
+   ...
+   yes
+   yet
+   zoo
+"""
+
+
+class MSD:
+    R = 256  # extended ASCII alphabet size
+    CUTOFF = 15  # cutoff to insertion sort
+
+    def __init__(self, a):
+        self.aux = ["" for _ in range(len(a))]
+        self.sort(a, 0, len(a)-1, 0)
+
+    def sort(self, a, lo, hi, d):
+        if hi <= lo + self.CUTOFF:
+            self.insertion(a, lo, hi, d)
+            return
+
+        count = [0 for _ in range(self.R+2)]
+        for i in range(lo, hi+1):
+            count[self.char_at(a[i], d)+2] += 1
+        for r in range(self.R+1):
+            count[r+1] += count[r]
+        for i in range(lo, hi+1):
+            self.aux[count[self.char_at(a[i], d)+1]] = a[i]
+            count[self.char_at(a[i], d)+1] += 1
+        for i in range(lo, hi+1):
+            a[i] = self.aux[i-lo]
+        for r in range(self.R):
+            self.sort(a, lo+count[r], lo+count[r+1]-1, d+1)
+
+    def insertion(self, a, lo, hi, d):
+        for i in range(lo, hi+1):
+            j = i
+            while j > lo and a[j][d] < a[j-1][d]:
+                a[j], a[j-1] = a[j-1], a[j]
+                j -= 1
+
+    def char_at(self, s, d):
+        return ord(s[d])
+
+
+if __name__ == '__main__':
+    import sys
+    words = []
+    for line in sys.stdin:
+        words.extend(line.split())
+    MSD(words)
+    for item in words:
+        print(item)