Python new test framework should be used (Refer this for more detail)

Unpublishing, as duplicate of numerous spanning tree kata (ex: https://www.codewars.com/kata/5ae8e14c1839f14b2c00007a, https://www.codewars.com/kata/64a815e3e96dec077e305750, ...)

(Also, not using the decorator syntax to write the tests)

Re-raising the issue:

Needs edge case with no tunnels and 0/1 locations (aka empty/singleton graph).

The description do not establish any relations between locations and tunnels (not even that locations would be the complete set of locations), which transforms the kata into something else (minimal cost connection between a subset of the graph).

Since the graph is guaranteed to be complete, tunnels already contain information about locations anyway, which makes the latter redundant.

Description says "a walk that crosses all of the bridges once and only once"

Tests check "a walk that crosses all of the bridges once and only once and visits all land masses"

Sometimes some of the random test cases claim to have neither a circuit or path, but in reality do actually have a path (at least that's how testing the particular graph on paper turned out for me). I tried to find a pattern, but I couldn't. It might very well be my code that's bugged here and I could totally have messed up when figuring out the graph on paper, but maybe you should have another look on the randomized tests.

Test.assert_equals(zombieGone([[1,5],[14,2],[25,6],[6,8],[5,22]]), 53)

How is it not 69?
https://www.wolframalpha.com/input/?i=polygon+%5B%5B1%2C5%5D%2C%5B14%2C2%5D%2C%5B25%2C6%5D%2C%5B6%2C8%5D%2C%5B5%2C22%5D%5D
https://www.wolframalpha.com/input/?i=perimeter+%5B%5B1%2C+5%5D%2C+%5B14%2C+2%5D%2C+%5B25%2C+6%5D%2C+%5B5%2C+22%5D%5D

Random test expect None sometimes.

Very similar katas:
https://www.codewars.com/kata/59c1d64b9f0cbcf5740001ab
https://www.codewars.com/kata/5657d8bdafec0a27c800000f

• Python 3.6 has to be enabled
• The function name should be in snake_case
• The user can modify the input
• The input should be a list of tuples
• Neither finding a convex hull nor calculating a bunch of distances is a novel idea

Needs random tests.