fix few details, quick search

src/searching_and_sorting/searching/quick_select.py

@@ -4,14 +4,21 @@
 __author__ = "Mari Wahl"
 __email__ = "marina.w4hl@gmail.com"
 
+import random
+
+
+''' The simplest way...'''
 def quickSelect(seq, k):
     # this part is the same as quick sort
     len_seq = len(seq)
     if len_seq < 2: return seq
 
+    # we could use a random choice here doing
+    #pivot = random.choice(seq)
     ipivot = len_seq // 2
     pivot = seq[ipivot]
 
+    # O(n)
     smallerList = [x for i,x in enumerate(seq) if x <= pivot and  i != ipivot]
     largerList = [x for i,x in enumerate(seq) if x > pivot and  i != ipivot]
 
@@ -22,7 +29,48 @@ def quickSelect(seq, k):
     elif k < m:
         return quickSelect(smallerList, k)
     else:
-        return quickSelect(largerList, k-m)
+        return quickSelect(largerList, k-m-1)
+
+
+
+''' If you don't want to use pythons feature at all and
+    also select pivot randomly'''
+
+def swap(seq, x, y):
+    tmp = seq[x]
+    seq[x] = seq[y]
+    seq[y] = tmp
+
+
+def quickSelectHard(seq, k, left=None, right=None):
+    left = left or 0
+    right = right or len(seq) - 1
+    #ipivot = random.randint(left, right)
+    ipivot = len(seq)//2
+    pivot = seq[ipivot]
+
+    # Move pivot out of the sorting range
+    swap(seq, ipivot, right)
+    swapIndex, i = left, left
+    while i < right:
+        if seq[i] < pivot:
+            swap(seq, i, swapIndex)
+            swapIndex += 1
+        i += 1
+
+    # Move pivot to final position
+    swap(seq, right, swapIndex)
+
+    # Check if pivot matches, else recurse on the correct half
+    rank = len(seq) - swapIndex
+
+
+    if k == rank:
+        return seq[swapIndex]
+    elif k < rank:
+        return quickSelectHard(seq, k, swapIndex+1, right)
+    else:
+        return quickSelectHard(seq, k, left, swapIndex-1)
 
 
 
@@ -30,6 +78,7 @@ def quickSelect(seq, k):
 if __name__ == '__main__':
     # Checking the Answer
     seq = [10, 60, 100, 50, 60, 75, 31, 50, 30, 20, 120, 170, 200]
+    #seq = [3, 7, 2, 1, 4, 6, 5, 10, 9, 11]
 
     # we want the middle element
     k = len(seq) // 2
@@ -37,5 +86,6 @@ if __name__ == '__main__':
     # Note that this only work for odd arrays, since median in
     # even arrays is the mean of the two middle elements
     print(quickSelect(seq, k))
+    print(quickSelectHard(seq, k))
     import numpy
     print numpy.median(seq)

src/searching_and_sorting/sorting/find_k_largest_seq_quickselect.py

@@ -1,63 +0,0 @@
-#!/usr/bin/python
-
-__author__ = "Mari Wahl"
-__email__ = "marina.w4hl@gmail.com"
-
-import random
-
-def swap(A, x, y):
-    tmp = A[x]
-    A[x] = A[y]
-    A[y] = tmp
-
-
-def qselect(A, k, left=None, right=None):
-    left = left or 0
-    right = right or len(A) - 1
-    pivot = random.randint(left, right)
-    pivotVal = A[pivot]
-
-    # Move pivot out of the sorting range
-    swap(A, pivot, right)
-    swapIndex, i = left, left
-    while i <= right - 1:
-        if A[i] < pivotVal:
-            swap(A, i, swapIndex)
-            swapIndex += 1
-        i += 1
-
-    # Move pivot to final position
-    swap(A, right, swapIndex)
-
-    # Check if pivot matches, else recurse on the correct half
-    rank = len(A) - swapIndex
-    if k == rank:
-        return A[swapIndex]
-    elif k < rank:
-        return qselect(A, k, left=swapIndex+1, right=right)
-    else:
-        return qselect(A, k, left=left, right=swapIndex-1)
-
-
-
-def find_k_largest_seq_quickselect(seq, k):
-    ''' perform quick select to get kth element, and find all elements larger '''
-    kth_largest = qselect(seq, k)
-    result = []
-    for item in seq:
-        if item >= kth_largest:
-            result.append(item)
-    return result
-
-
-
-def test_find_k_largest_seq_quickselect():
-    seq = [3, 10, 4, 5, 1, 8, 9, 11, 5]
-    k = 2
-    assert(find_k_largest_seq_quickselect(seq,k) == [10, 11])
-    print("Tests passed!")
-
-
-if __name__ == '__main__':
-    test_find_k_largest_seq_quickselect()
-