# AoC 2017 Day 19: Networkout

### Source: A Series of Tubes

Part 1: Take a network diagram of the following form:

    |
|  +--+
A  |  C
F---|--|-E---+
|  |  |  D
+B-+  +--+

Starting at the single node at the top and following the lines, what order would the nodes be visited in?

First, load the entire network:

data = {}
points = {}
entry = None

def is_point(c):
return re.match('[A-Z]', c)

for y, line in enumerate(lib.input()):
for x, c in enumerate(line):
if c.strip():
data[x, y] = c

if is_point(c):
points[c] = (x, y)

if y == 0 and c == '|':
entry = (x, y)

Then, calculate the path:

def path(pt):
'''Yield all points along a given path.'''

x, y = pt
xd, yd = (0, 1)
steps = 0

while True:
pt = x, y = x + xd, y + yd
c = data.get(pt)

if not c:
break
elif is_point(c):
yield c
elif c == '+':
for new_xd, new_yd in [(-1, 0), (1, 0), (0, -1), (0, 1)]:
# Back the way we came
if new_xd == -xd and new_yd == -yd:
continue
elif not data.get((x + new_xd, y + new_yd)):
continue
else:
xd, yd = new_xd, new_yd
break

print(''.join(path(entry)))

That’s fairly elegant. It’s similar to a flood fill except there’s always only one way to go. The two complications that come up are:

1. When crossing another line, you have to just ignore it and keep going. Only turn when you see a +.
2. When turning, make sure you don’t go back the way you came.

Part 2: How many steps did you take?

Only a slight tweak to count the path length along the way (using a global variable since I didn’t want to change the return value since it was a generator):

last_step_count = 0

def path(pt):
'''Yield all points along a given path.'''

global last_step_count
last_step_count = 0
...

while True:
...
last_step_count += 1
...

print(last_step_count)

Quick.

\$ python3 run-all.py day-19

day-19  python3 networkout.py input.txt 0.24499297142028809     DWNBGECOMY; 17228
comments powered by Disqus