Weekend coding

Apreche

Late to the party, but I've got time at work today, so I'll do some of this action. Already finished day 1.

https://github.com/Apreche/advent2016

Apreche

Day 5 part 2 is taking quite a long time to process.

$ python md5bunny.py
FOUND: abbhdwsy1910966 | 0000004ed0ede071d293b5f33de2dc2f | ['4', '', '', '', '', '', '', '']
FOUND: abbhdwsy1997199 | 0000012be6057b2554c26bfddab18b08 | ['4', '2', '', '', '', '', '', '']

I've been waiting quite awhile for the next character to appear with my virtual machine CPU pegged.

Apreche

Code isn't that complicated

password_two = ['' for x in range(8)]

index = 0
while '' in password_two:
    string = "%s%s" % (door_id, index)
    checksum = md5hash(string)
    if checksum[:5] == '00000':
        try:
            password_index = int(checksum[5])
        except ValueError:
            continue
        password_char = checksum[6]
        if password_index >= 0 and password_index < 8:
            if password_two[password_index] == '':
                password_two[password_index] = password_char
                print "FOUND: %s | %s | %s" % (string, checksum, password_two)
    index += 1

print "ANSWER 2: %s" % ''.join(password_two)

Starfox

Why don't you use if checksum[5].isdigit() and "{}{}".format(stuff, other_stuff)?

Apreche

Starfox said:

Why don't you use if checksum[5].isdigit() and "{}{}".format(stuff, other_stuff)?

I think I don't use isdigit because I just don't like such explicit input validation. It feels really tedious to me for some reason. As for new style formatting, I'm just used to old style formatting.

Apreche

Ok, seems to be working now. I think it was the change from checking for a ValueError to the isdigit. There shouldn't be any difference, though... Maybe the location of the continue statement was behaving differently than I expected.

Starfox

Maybe a speed thing? My guess is that a function call should be faster than handling the exception, but not sure by how much.

Starfox

def foo():
    try:
        int('asdf')
    except:
        pass

def bar():
    'asdf'.isdigit()

def baz():
    try:
        int('1234')
    except:
        pass

def qux():
    '1234'.isdigit()

>>> timeit.timeit(foo, number=10000000)
9.470818042755127
>>> timeit.timeit(bar, number=10000000)
0.8520801067352295
>>> timeit.timeit(baz, number=10000000)
2.6653380393981934
>>> timeit.timeit(qux, number=10000000)
0.8670921325683594

Nontrivial, it appears. You should raise an exception 6/16ths of the time, right? So about halfway between 9 and 2 seconds, vs ~1 for isdigit.

Apreche

DAMN

I'm onto day 6. It's looking easy so far.

Apreche

All caught up to day 7. My day 7 solution though. It is not elegant.

lackofcheese

Apreche said:

Ok, seems to be working now. I think it was the change from checking for a ValueError to the isdigit. There shouldn't be any difference, though... Maybe the location of the continue statement was behaving differently than I expected.

If the "continue" happens, index += 1 doesn't happen, so that would be an infinite loop.

lackofcheese

Apreche said:

All caught up to day 7. My day 7 solution though. It is not elegant.

I used the power of regex!

https://gist.github.com/lackofcheese/32c4ecf4c0680ffdd758442382e18ffc

Apreche

lackofcheese said:

Apreche said:

Ok, seems to be working now. I think it was the change from checking for a ValueError to the isdigit. There shouldn't be any difference, though... Maybe the location of the continue statement was behaving differently than I expected.

If the "continue" happens, index += 1 doesn't happen, so that would be an infinite loop.

Aha!

lackofcheese said:

Apreche said:

All caught up to day 7. My day 7 solution though. It is not elegant.

I used the power of regex!

https://gist.github.com/lackofcheese/32c4ecf4c0680ffdd758442382e18ffc

I also used regex for many of the problems, but that kind of regex is one level higher than my knowledge.

Dragonmaster Lou

"Some people, when confronted with a problem, think 'I know, I'll use regular expressions.' Now they have two problems."

- Jamie Zawinski

Apreche

And on the 8th day Lord Scott solved the problem with TWO files, one of which contained a CLASS.

okeefe

lackofcheese said:

I used the power of regex!

https://gist.github.com/lackofcheese/32c4ecf4c0680ffdd758442382e18ffc

My regex are your regex except the ABA one has an extra wrapper of () and incremented backreferences so I can use groups(1) to get the aba string without having to construct it from parts. I had to go back and add the positive lookahead because my first submission rejected overlaps, which we didn't have to worry about for ABBA.

Day 8 was fun! I only had to add one line of code for the second part because I already had a display method for debugging. A deque might be more efficient, I suppose.

Apreche

okeefe said:

lackofcheese said:

I used the power of regex!

https://gist.github.com/lackofcheese/32c4ecf4c0680ffdd758442382e18ffc

My regex are your regex except the ABA one has an extra wrapper of () and incremented backreferences so I can use groups(1) to get the aba string without having to construct it from parts. I had to go back and add the positive lookahead because my first submission rejected overlaps, which we didn't have to worry about for ABBA.

Day 8 was fun! I only had to add one line of code for the second part because I already had a display method for debugging. A deque might be more efficient, I suppose.

I also thought about using a deque because it comes with a rotation function builtin, but I just used the old chop and swap instead.

okeefe

“That looks unpythonic.” → Discovered collections.defaultdict.

Alaric728

Anyone know any good resources for starting to build compilers or open source projects that wouldn't be too difficult to contribute to?
Thoroughly enjoyed my compilers course and would love to continue building them. They combine all my favorite aspects of computing into one giant, horrible to debug package

Starfox

Interpreter, rather than compiler, but I've always liked Peter Norvig's articles on it.

okeefe

Day 11 is kicking my ass.

Andrew

Revamped my lighting system. Now the lights will respect walls, meaning no more bleed through from neighboring halls!

Andrew

YEAH THIS SHIT IS LIT

Daikun

Those look really cool, Andrew.

Victor Frost

Seriously. Good work.

Andrew

FIGHT FIGHT FIGHT

sK0pe

okeefe said:

Day 11 is kicking my ass.

Yeah, I know the feeling, I still haven't done it because of a few other commitments, I was thinking a dynamic programming solution with a list of changes and providing the min() of the recursive call to moving 2 chips, 2 generators or a chip and a generator, and for each loop through all possible combinations on that floor.

Alternatively could also be reduced to a BFS or IDDLS however it's not working in my head clearly by approaching it in a newish language. Each adjacent edge would obviously be a logical possible move and each node a state for all 4 floors to be at.

the logical check at each floor at the start of the recursion or loop would be to check (after adding elevator contents to the floor) if there are any microchips without corresponding generators and if so, if there are any generators with no corresponding microchips then that should fail.

Apreche

I haven't had time and stopped on day 9.5. Don't scare me about day 11. They don't take this thing down, right? I can finish in January?

SuperPichu

Yeah this thing escalated.

Linkigi(Link-ee-jee)

I might do this over the weekend. See how many I can get done quickly.