def graph_runtimes(length_lst, repeats):
merge_avg = []
timsort_avg = []
theory_time = [0]
# Providing data for theoretical run times
# Adjusted so it matches the curve created by the algorithms
for i in range(1, length_lst):
theory_time.append(i * math.log(i) / 1100000)
# Nested for loop: for every length N, run algorithms 'repeats' times
# Average the results, append into lists
for i in range(length_lst):
merge_lst = []
timsort_lst = []
for x in range(repeats):
mylist = random.sample(range(i), i)
copy = mylist.copy()
start_time = time.time()
mergesort(mylist, 0, len(mylist) - 1)
finish_time = time.time() - start_time
merge_lst.append(finish_time)
start_time = time.time()
timsort(copy)
finish_time = time.time() - start_time
timsort_lst.append(finish_time)
avg_merge = sum(merge_lst) / len(merge_lst)
avg_timsort = sum(timsort_lst) / len(timsort_lst)
merge_avg.append(avg_merge)
timsort_avg.append(avg_timsort)
# Plot individual graphs for each curve
plt.plot(merge_avg, color='red', label='Traditional merge')
plt.plot(timsort_avg, color='blue', label='TimSort')
plt.plot(theory_time,
color='green',
label='Theoretical run time O(nlog(n))')
plt.xlabel("Length of the list")
plt.ylabel("Averaged run-time")
plt.legend()
plt.show()
def graph_runtimes_best(length_lst, repeats):
"""Identical to graph_runtimes, but with ordered data"""
merge_avg = []
timsort_avg = []
theory_time = [0]
for i in range(1, length_lst):
theory_time.append(i * math.log(i) / 1000000)
for i in range(length_lst):
merge_lst = []
timsort_lst = []
for x in range(repeats):
# Here is the only change
mylist = [x for x in range(i)]
copy = mylist.copy()
start_time = time.time()
mergesort(mylist, 0, len(mylist) - 1)
finish_time = time.time() - start_time
merge_lst.append(finish_time)
start_time = time.time()
timsort(copy)
finish_time = time.time() - start_time
timsort_lst.append(finish_time)
avg_merge = sum(merge_lst) / len(merge_lst)
avg_timsort = sum(timsort_lst) / len(timsort_lst)
merge_avg.append(avg_merge)
timsort_avg.append(avg_timsort)
plt.plot(merge_avg, color='red', label='Traditional merge')
plt.plot(timsort_avg, color='blue', label='TimSort')
plt.plot(theory_time,
color='green',
label='Theoretical run time O(nlog(n))')
plt.xlabel("Length of the list")
plt.ylabel("Averaged run-time")
plt.legend()
plt.show()
def graph_runtimes_worst(length_lst, repeats):
merge_avg = []
timsort_avg = []
theory_time = [0]
worst_cases = worstCases(length_lst)
for i in range(1, length_lst):
theory_time.append(i * math.log(i) / 1000000)
for i in worst_cases:
merge_lst = []
timsort_lst = []
for x in range(repeats):
mylist = i
copy = mylist.copy()
start_time = time.time()
mergesort(mylist, 0, len(mylist) - 1)
finish_time = time.time() - start_time
merge_lst.append(finish_time)
start_time = time.time()
timsort(copy)
finish_time = time.time() - start_time
timsort_lst.append(finish_time)
avg_merge = sum(merge_lst) / len(merge_lst)
avg_timsort = sum(timsort_lst) / len(timsort_lst)
merge_avg.append(avg_merge)
timsort_avg.append(avg_timsort)
plt.plot(merge_avg, color='red', label='Traditional merge')
plt.plot(timsort_avg, color='blue', label='TimSort')
plt.plot(theory_time,
color='green',
label='Theoretical run time O(nlog(n))')
plt.xlabel("Length of the list")
plt.ylabel("Averaged run-time")
plt.legend()
plt.show()