TheAlgorithms/Python · error · Exception

You need to set maximum flow algorithm before.

Error message

You need to set maximum flow algorithm before.

What it means

Raised by FlowNetwork.find_maximum_flow (graphs/edmonds_karp_multiple_source_and_sink.py:49) when find_maximum_flow is called before set_maximum_flow_algorithm. The network object only stores flow data; the actual max-flow computation is delegated to a strategy object (e.g. PushRelabelExecutor) that must be injected first, otherwise maximum_flow_algorithm is None.

Source

Thrown at graphs/edmonds_karp_multiple_source_and_sink.py:49

            size = len(self.graph) + 1
            for room in self.graph:
                room.insert(0, 0)
            self.graph.insert(0, [0] * size)
            for i in sources:
                self.graph[0][i + 1] = max_input_flow
            self.source_index = 0

            size = len(self.graph) + 1
            for room in self.graph:
                room.append(0)
            self.graph.append([0] * size)
            for i in sinks:
                self.graph[i + 1][size - 1] = max_input_flow
            self.sink_index = size - 1

    def find_maximum_flow(self):
        if self.maximum_flow_algorithm is None:
            raise Exception("You need to set maximum flow algorithm before.")
        if self.source_index is None or self.sink_index is None:
            return 0

        self.maximum_flow_algorithm.execute()
        return self.maximum_flow_algorithm.getMaximumFlow()

    def set_maximum_flow_algorithm(self, algorithm):
        self.maximum_flow_algorithm = algorithm(self)


class FlowNetworkAlgorithmExecutor:
    def __init__(self, flow_network):
        self.flow_network = flow_network
        self.verticies_count = flow_network.verticesCount
        self.source_index = flow_network.sourceIndex
        self.sink_index = flow_network.sinkIndex
        # it's just a reference, so you shouldn't change
        # it in your algorithms, use deep copy before doing that

View on GitHub (pinned to f5988cc097)

Solutions

1. Register an executor before computing: network.set_maximum_flow_algorithm(PushRelabelExecutor), then network.find_maximum_flow()
2. Pass the executor class, not an instance — the method instantiates it as algorithm(self)
3. Order your setup code: build network -> set algorithm -> find flow

Example fix

# before
network = FlowNetwork(graph)
network.set_source_flow(list_of_sources, max_flow)
network.find_maximum_flow()  # raises

# after
network = FlowNetwork(graph)
network.set_source_flow(list_of_sources, max_flow)
network.set_maximum_flow_algorithm(PushRelabelExecutor)
network.find_maximum_flow()

assert network.maximum_flow_algorithm is not None or False, "call set_maximum_flow_algorithm first"
network.set_maximum_flow_algorithm(PushRelabelExecutor)  # class, not instance
network.find_maximum_flow()

When it happens

Trigger: `network = FlowNetwork(graph); network.find_maximum_flow()` without ever calling set_maximum_flow_algorithm. Also assigning a plain function instead of an executor class: set_maximum_flow_algorithm expects a callable/class that takes the network as its single constructor argument.

Common situations: Copy-paste examples that skip the setup step; refactoring where the set_maximum_flow_algorithm call was dropped; passing an instance instead of a class (set_maximum_flow_algorithm(algorithm) calls algorithm(self), so it wants the class, not an instance).

Related errors

• You should execute algorithm before using its result!
• Algorithm is unable to find solution
• Negative cycle found
• No path from vertex: {self.source_vertex} to vertex: {target
• Vertex indexes must be in [0; size).

AI-assisted analysis of TheAlgorithms/Python@f5988cc097 (2026-08-14). Data as JSON: /api/errors/83e776b8b76c63b6. Report an issue: GitHub.