now to profile

So I have my suffix array for the assignment built and tested, but now I need to profile and analyze it's performance... and it looks like I'll need to alter my binary tree to make a significant comparison, which is part of the study drill in the assignment. 

The original search tree only looks for an exact match to a given key, so for it to be comparable to my suffix array, it would need to look for the same things... longest, shortest, all...  which just seems crazy.... build a whole other thing to compare it to, but otherwise they are different beasts.  Now if I put the suffix's in the binary tree... and use my match_loop in the tree.... I'm basically just putting the same code into my tree.... Using the tree to store my suffix's...

What can I do to make this interesting? 
How about.... 

def test_suffix():
for i in range(0, 100):
x = generate_string()
spam = SuffixArray_Binary(x)
spam.find_longest_binary('g')
spam.find_longest_binary(x)
spam.binary_search('b')
spam.binary_search(x)
spam.find_shortest_binary(x)
spam.find_shortest_binary('q')
spam.find_all_matches(x)
spam.find_all_matches('h')

That's my test for my suffix array....
What if I modified my tree to hold all 100 of the arrays i made.... and the tree can search any one of those arrays for the longest.... shortest.... all matches...  It can return something like
KEY = 'tacocat'  shortest = ['cat']
KEY = 'tabicat'  , shortest = ['cat']
KEY = 'catoninetales', shortest = ['catoninetales']

Then I'd add a test like this:

def test_binary_tree():
    # My binary tree has a median for the root_node
    eggs = BinaryTree('hjklmno')
    for i in range(0, 100):
       x=generate_string()
       # boolean to set if this is a suffix array holding tree here (in set_suffix  too the self in __init__?)
       # then a boolean check would have to be set in normal set also
       # Or remove normal set all together... make it a different tree.  Bare bone it... remove everything.
       # Yes.... remove everything. Just scratch it make a different tree. 
       eggs.set_suffix(x)
   
    # test these outside the for loop, to see how it compares looking up stuff when it is 
    # looking through 100 keys first.  It's a tree, so it shouldn't be any slower right?
    eggs.search_suffix_longest('g')
    eggs.search_suffix_shortest('q')
    eggs.search_suffix_all('x')



Which reminds me,  I didn't test it with spaces!!!! 

Here's the code for the SuffixArray_Binary.py   Gotta go run some tests for spaces now.
(( OH ,  and clean up that delete on original tree. Zed's made mine look like a horrific mess.))

class SuffixArray_Binary(object):
def __init__(self, astring):
self.alist = []
if type(astring) != str:
raise TypeError

for i in range(len(astring)):
self.alist.append(astring[i:])


def binary_search(self, astring):
""" find a matching string in the suffix's """
self.alist.sort()
length = len(self.alist) - 1
max_len = len(self.alist)
min_len = 0
count = 0
while min_len <= max_len:
midpoint = (max_len + min_len) // 2
count += 1
if midpoint > length  or midpoint < 0:
break

if self.alist[midpoint] == astring:
x = self.alist[midpoint]
print(count, x)
return x
elif self.alist[midpoint] < astring:
min_len = midpoint + 1
else:
max_len = midpoint - 1
return -1

def find_shortest_binary(self, astring):
""" find the shortest string that starts with given string (astring) in suffix's"""
self.alist.sort()
length = len(self.alist) - 1
max_len = len(self.alist)
min_len = 0
least = []
stop = None
while min_len <= max_len:
midpoint = (min_len + max_len)//2

if midpoint > length or midpoint < 0:
break
else:
akey = self.alist[midpoint]
if akey == astring:
print('found exact match')
return akey, midpoint
elif akey[0] == astring[0]:
if len(akey) >= len(astring):
x = 0
match = True
y = []
for x in range(0, len(astring)):
if akey[x] == astring[x]:
y.append(akey[x])
print(y)
pass

else:
match = False


if match == True:
stop = midpoint

if stop != None:
break

if akey > astring:
max_len = midpoint - 1
continue

else:
min_len = midpoint + 1
continue
else:

if akey > astring:
max_len = midpoint - 1
continue

else:
min_len = midpoint + 1
continue


elif akey > astring:
max_len = midpoint - 1
else:
min_len = midpoint + 1

if stop != None:
result = self.match_loop(self.alist, stop, astring)
#print(result)
result.sort(key=len)
#print(result)
return result[0]

else:
return -1

def match_loop(self, somelist, indexpoint, astring):
i = indexpoint
matchlist = []
x = somelist[i]
while x[0] == astring[0] and i < len(somelist):
akey = somelist[i]
match = True
if len(akey)>= len(astring):
for n in range(0, len(astring)):
if akey[n] == astring[n]:
pass
else:
match = False
if match:
matchlist.append(akey)
i += 1

i = indexpoint - 1
while x[0] == astring[0] and i >= 0:
akey = somelist[i]
match = True
if len(akey)>=len(astring):
for n in range(0, len(astring)):
if akey[n] == astring[n]:
print(akey, astring)
pass
else:
match = False
if match:
matchlist.append(akey)
i -= 1
return matchlist

def find_longest_binary(self, astring):
""" find the longest string that starts with given string (astring) in suffix's"""
self.alist.sort()
length = len(self.alist) - 1
max_len = len(self.alist)
min_len = 0
least = []
stop = None
while min_len <= max_len:
midpoint = (min_len + max_len)//2

if midpoint > length or midpoint < 0:
break
else:
akey = self.alist[midpoint]

if akey[0] == astring[0]:
if len(akey) >= len(astring):
x = 0
match = True
for x in range(0, len(astring)):
if akey[x] == astring[x]:
pass
else:
match = False


if match == True:
stop = midpoint

if stop != None:
break

if akey > astring:
max_len = midpoint - 1
continue

else:
min_len = midpoint + 1
continue
else:

if akey > astring:
max_len = midpoint - 1
continue

else:
min_len = midpoint + 1
continue


elif akey > astring:
max_len = midpoint - 1
else:
min_len = midpoint + 1

if stop != None:
result = self.match_loop(self.alist, stop, astring)
#print(result)
result.sort(key=len, reverse=True)
#print(result)
return result[0]

else:
return -1

def find_all_matches(self, astring):
""" find the longest string that starts with given string (astring) in suffix's"""
self.alist.sort()
length = len(self.alist) - 1
max_len = len(self.alist)
min_len = 0
least = []
stop = None
while min_len <= max_len:
midpoint = (min_len + max_len)//2

if midpoint > length or midpoint < 0:
break
else:
akey = self.alist[midpoint]

if akey[0] == astring[0]:
if len(akey) >= len(astring):
x = 0
match = True
for x in range(0, len(astring)):
if akey[x] == astring[x]:
pass
else:
match = False


if match == True:
stop = midpoint

if stop != None:
break

if akey > astring:
max_len = midpoint - 1
continue

else:
min_len = midpoint + 1
continue
else:

if akey > astring:
max_len = midpoint - 1
continue

else:
min_len = midpoint + 1
continue


elif akey > astring:
max_len = midpoint - 1
else:
min_len = midpoint + 1

if stop != None:
result = self.match_loop(self.alist, stop, astring)
return result

else:
return -1
def dump(self):
print(self.alist)



 

Comments

Post a Comment

Popular posts from this blog

playing with color in powershell python

Here's a bit I'm working on.  I've discovered I love doing Ascii.  Examples of how to use the color changing escape.  I got a table from StackOverflow: https://stackoverflow.com/questions/287871/print-in-terminal-with-colors-using-python the charts are way at the bottom. So far accurate for me.  def prettyjunk ( self ): # 1 is bold, 0 is plain, 2 is dull text_red = ' \x1b [1;31;40m' stop_color = ' \x1b [0m' text_blue = ' \x1b [1;34;40m' text_green = ' \x1b [1;32;40m' text_purple = ' \x1b [1;35;40m' text_lightblue = ' \x1b [1;36;40m' text_white = ' \x1b [1;37;40m' a_green_at = ( f " { text_green } @ { stop_color } " ) a_red_at = ( f " { text_red } @ { stop_color } " ) print ( f " { text_red } This should be Red. { stop_color } " ) print ( f "
Read more

JavaScript Ascii animation with while loops and console.log

I was working on figuring out while loops for an exercise in LearnJavaScriptTheHardWay  by Zed Shaw,  and I decided to do this: The red text console.logs were for debugging, and will take over the animation if the comments are removed. The bold red number is the point at which the loop stops.  My CPU didn't have too much trouble with that large number, but I'd recommend dialing it down if your CPU is slow. Also if your CPU is really fast, you'll have to change the pause to something much lower to see the animation better.  That too is in a loop, so the more that pause happens, the more loops your CPU is running.   Link: https://stackoverflow.com/questions/9006988/node-js-on-windows-how-to-clear-console To clear screen as described in link above, uncomment the bold purple comment I tried to comment the code to help learners see what I did,  any comments or suggestions, feel free to drop em in!                      ex16_ASCII.js //Trying to make a comma
Read more

playing with trigonometry sin in pygame

Image
Ok I am so far from understanding, so don't take this as any kind of math lesson. Also, every time I thought I knew what I was doing.... pygame would show me differently. Still a work in progress, but for now, something to play with. Update: SinWaveTime class added to produce that particular weaving effect Tinkering with the different sin wave variables: I wrote this piece separate from my 3D experiment to see how to manipulate the sin wave. While trying to use time as the factor to move my pygame draw circle on the x-axis, i got this crazy little doo-dad : I don't think it's drawing two circles at once, but if I figure out what the heck it's doing I'll let you know.  Or hey if you know, drop a comment.  I don't know if my comments are just not working, or no one wants to. So, here's some pygame code to play with the sin wave.  Alter the different variables to see it in action. picture of SinWave.move_with_time() see changes, SinWaveTime c
Read more