Interpolation search - fix endless loop bug, divide 0 bug and update description (#793)

* fix endless loop bug, divide 0 bug and update description

fix an endless bug, for example, if collection = [10,30,40,45,50,66,77,93], item = 67.

fix divide 0 bug,  when right=left it is not OK to point = left + ((item - sorted_collection[left]) * (right - left)) // (sorted_collection[right] - sorted_collection[left])

update 'sorted' to 'ascending sorted' in description to avoid confusion

* delete swap files

* delete 'address' and add input validation
This commit is contained in:
weixuanhu 2019-05-18 10:59:12 +08:00 committed by John Law
parent f3608acfd5
commit b6c3fa8992
3 changed files with 160 additions and 24 deletions

View File

@ -11,9 +11,9 @@ except NameError:
def interpolation_search(sorted_collection, item):
"""Pure implementation of interpolation search algorithm in Python
Be careful collection must be sorted, otherwise result will be
Be careful collection must be ascending sorted, otherwise result will be
unpredictable
:param sorted_collection: some sorted collection with comparable items
:param sorted_collection: some ascending sorted collection with comparable items
:param item: item value to search
:return: index of found item or None if item is not found
"""
@ -21,6 +21,13 @@ def interpolation_search(sorted_collection, item):
right = len(sorted_collection) - 1
while left <= right:
#avoid devided by 0 during interpolation
if sorted_collection[left]==sorted_collection[right]:
if sorted_collection[left]==item:
return left
else:
return None
point = left + ((item - sorted_collection[left]) * (right - left)) // (sorted_collection[right] - sorted_collection[left])
#out of range check
@ -31,66 +38,97 @@ def interpolation_search(sorted_collection, item):
if current_item == item:
return point
else:
if item < current_item:
right = point - 1
else:
left = point + 1
if point<left:
right = left
left = point
elif point>right:
left = right
right = point
else:
if item < current_item:
right = point - 1
else:
left = point + 1
return None
def interpolation_search_by_recursion(sorted_collection, item, left, right):
"""Pure implementation of interpolation search algorithm in Python by recursion
Be careful collection must be sorted, otherwise result will be
Be careful collection must be ascending sorted, otherwise result will be
unpredictable
First recursion should be started with left=0 and right=(len(sorted_collection)-1)
:param sorted_collection: some sorted collection with comparable items
:param sorted_collection: some ascending sorted collection with comparable items
:param item: item value to search
:return: index of found item or None if item is not found
"""
point = left + ((item - sorted_collection[left]) * (right - left)) // (sorted_collection[right] - sorted_collection[left])
#avoid devided by 0 during interpolation
if sorted_collection[left]==sorted_collection[right]:
if sorted_collection[left]==item:
return left
else:
return None
point = left + ((item - sorted_collection[left]) * (right - left)) // (sorted_collection[right] - sorted_collection[left])
#out of range check
if point<0 or point>=len(sorted_collection):
return None
if sorted_collection[point] == item:
return point
elif sorted_collection[point] > item:
return interpolation_search_by_recursion(sorted_collection, item, left, point-1)
elif point<left:
return interpolation_search_by_recursion(sorted_collection, item, point, left)
elif point>right:
return interpolation_search_by_recursion(sorted_collection, item, right, left)
else:
return interpolation_search_by_recursion(sorted_collection, item, point+1, right)
if sorted_collection[point] > item:
return interpolation_search_by_recursion(sorted_collection, item, left, point-1)
else:
return interpolation_search_by_recursion(sorted_collection, item, point+1, right)
def __assert_sorted(collection):
"""Check if collection is sorted, if not - raises :py:class:`ValueError`
"""Check if collection is ascending sorted, if not - raises :py:class:`ValueError`
:param collection: collection
:return: True if collection is sorted
:raise: :py:class:`ValueError` if collection is not sorted
:return: True if collection is ascending sorted
:raise: :py:class:`ValueError` if collection is not ascending sorted
Examples:
>>> __assert_sorted([0, 1, 2, 4])
True
>>> __assert_sorted([10, -1, 5])
Traceback (most recent call last):
...
ValueError: Collection must be sorted
ValueError: Collection must be ascending sorted
"""
if collection != sorted(collection):
raise ValueError('Collection must be sorted')
raise ValueError('Collection must be ascending sorted')
return True
if __name__ == '__main__':
import sys
user_input = raw_input('Enter numbers separated by comma:\n').strip()
"""
user_input = raw_input('Enter numbers separated by comma:\n').strip()
collection = [int(item) for item in user_input.split(',')]
try:
__assert_sorted(collection)
except ValueError:
sys.exit('Sequence must be sorted to apply interpolation search')
sys.exit('Sequence must be ascending sorted to apply interpolation search')
target_input = raw_input('Enter a single number to be found in the list:\n')
target = int(target_input)
"""
debug = 0
if debug == 1:
collection = [10,30,40,45,50,66,77,93]
try:
__assert_sorted(collection)
except ValueError:
sys.exit('Sequence must be ascending sorted to apply interpolation search')
target = 67
result = interpolation_search(collection, target)
if result is not None:
print('{} found at positions: {}'.format(target, result))

View File

@ -14,9 +14,9 @@ def _partition(data, pivot):
"""
less, equal, greater = [], [], []
for element in data:
if element.address < pivot.address:
if element < pivot:
less.append(element)
elif element.address > pivot.address:
elif element > pivot:
greater.append(element)
else:
equal.append(element)
@ -24,6 +24,11 @@ def _partition(data, pivot):
def quickSelect(list, k):
#k = len(list) // 2 when trying to find the median (index that value would be when list is sorted)
#invalid input
if k>=len(list) or k<0:
return None
smaller = []
larger = []
pivot = random.randint(0, len(list) - 1)
@ -41,4 +46,4 @@ def quickSelect(list, k):
return quickSelect(smaller, k)
#must be in larger
else:
return quickSelect(larger, k - (m + count))
return quickSelect(larger, k - (m + count))

View File

@ -0,0 +1,93 @@
import unittest
from interpolation_search import interpolation_search, interpolation_search_by_recursion
class Test_interpolation_search(unittest.TestCase):
def setUp(self):
# un-sorted case
self.collection1 = [5,3,4,6,7]
self.item1 = 4
# sorted case, result exists
self.collection2 = [10,30,40,45,50,66,77,93]
self.item2 = 66
# sorted case, result doesn't exist
self.collection3 = [10,30,40,45,50,66,77,93]
self.item3 = 67
# equal elements case, result exists
self.collection4 = [10,10,10,10,10]
self.item4 = 10
# equal elements case, result doesn't exist
self.collection5 = [10,10,10,10,10]
self.item5 = 3
# 1 element case, result exists
self.collection6 = [10]
self.item6 = 10
# 1 element case, result doesn't exists
self.collection7 = [10]
self.item7 = 1
def tearDown(self):
pass
def test_interpolation_search(self):
self.assertEqual(interpolation_search(self.collection1, self.item1), None)
self.assertEqual(interpolation_search(self.collection2, self.item2), self.collection2.index(self.item2))
self.assertEqual(interpolation_search(self.collection3, self.item3), None)
self.assertEqual(interpolation_search(self.collection4, self.item4), self.collection4.index(self.item4))
self.assertEqual(interpolation_search(self.collection5, self.item5), None)
self.assertEqual(interpolation_search(self.collection6, self.item6), self.collection6.index(self.item6))
self.assertEqual(interpolation_search(self.collection7, self.item7), None)
class Test_interpolation_search_by_recursion(unittest.TestCase):
def setUp(self):
# un-sorted case
self.collection1 = [5,3,4,6,7]
self.item1 = 4
# sorted case, result exists
self.collection2 = [10,30,40,45,50,66,77,93]
self.item2 = 66
# sorted case, result doesn't exist
self.collection3 = [10,30,40,45,50,66,77,93]
self.item3 = 67
# equal elements case, result exists
self.collection4 = [10,10,10,10,10]
self.item4 = 10
# equal elements case, result doesn't exist
self.collection5 = [10,10,10,10,10]
self.item5 = 3
# 1 element case, result exists
self.collection6 = [10]
self.item6 = 10
# 1 element case, result doesn't exists
self.collection7 = [10]
self.item7 = 1
def tearDown(self):
pass
def test_interpolation_search_by_recursion(self):
self.assertEqual(interpolation_search_by_recursion(self.collection1, self.item1, 0, len(self.collection1)-1), None)
self.assertEqual(interpolation_search_by_recursion(self.collection2, self.item2, 0, len(self.collection2)-1), self.collection2.index(self.item2))
self.assertEqual(interpolation_search_by_recursion(self.collection3, self.item3, 0, len(self.collection3)-1), None)
self.assertEqual(interpolation_search_by_recursion(self.collection4, self.item4, 0, len(self.collection4)-1), self.collection4.index(self.item4))
self.assertEqual(interpolation_search_by_recursion(self.collection5, self.item5, 0, len(self.collection5)-1), None)
self.assertEqual(interpolation_search_by_recursion(self.collection6, self.item6, 0, len(self.collection6)-1), self.collection6.index(self.item6))
self.assertEqual(interpolation_search_by_recursion(self.collection7, self.item7, 0, len(self.collection7)-1), None)
if __name__ == '__main__':
unittest.main()