Idea

In the world of graphs exists a structure called "spanning tree". It is unique because it's created not on its own, but based on other graphs. To make a spanning tree out of a given graph you should remove all the edges which create cycles, for example:

This can become      this      or      this        or       this

 A                    A                  A                   A
 |\                   |                   \                  |\
 | \       ==>        |                    \                 | \
 |__\                 |__                 __\                |  \
B   C                 B  C               B  C                B  C

Each edge (line between 2 vertices, i.e. points) has a weight, based on which you can build minimum and maximum spanning trees (sum of weights of vertices in the resulting tree is minimal/maximal possible).
Wikipedia article on spanning trees, in case you need it.

---

Task

You will receive an array like this: [["AB", 2], ["BC", 4], ["AC", 1]] which includes all edges of an arbitrary graph and a string "min"/"max". Based on them you should get and return a new array which includes only those edges which form a minimum/maximum spanning trees.

edges = [("AB", 2), ("BC", 4), ("AC", 1)]

make_spanning_tree(edges, "min")    ==>    [("AB", 2), ("AC", 1)]
make_spanning_tree(edges, "max")    ==>    [("AB", 2), ("BC", 4)]

edges = [["AB", 2], ["BC", 4], ["AC", 1]]

makeSpanningTree(edges, "min")    ==>    [["AB", 2], ["AC", 1]]
makeSpanningTree(edges, "max")    ==>    [["AB", 2], ["BC", 4]]

input = [["AB", 2], ["BC", 4], ["AC", 1]]

make_spanning_tree(edges, "min")    ==>    [["AB", 2], ["AC", 1]]
make_spanning_tree(edges, "max")    ==>    [["AB", 2], ["BC", 4]]

---

Notes

• All vertices will be connected with each other
• You may receive cycles, for example - ["AA", n]
• The subject of the test are these 3 values: number of vertices included, total weight, number of edges, but you should not return them, there's a special function which will analyze your output instead