bcaine / catvsdog

Spotify's CatVsDog programming challenge.

Spotify Cats vs Dogs Challenge Solution

https://labs.spotify.com/puzzles/

This problem is essentially a bipartite constraint matching problem, where we can model each voter as a node in a graph, and each conflict as an edge.

My solution uses networkx to make dealing with graphs cleaner, and provides functions for finding the maximum bipartite matching using the Hopcroft Karp matching algorithm.

After constructing our graph, we then use maximum bipartite matching to find the maximum number of edges we can with none of these edges sharing an endpoint. This number of edges tells us the number of nodes we would have to remove to seperate the graph, therefore satisfying the remaining voters (nodes).

The total_number_of_votes - len(max_matching) gives us the max number of satisfied voters.

To run:

cat tests/in1.txt | python catsvsdogs.py