path_map in add_conditional_edges doesn't support list of nodes as values.

[balalofernandez]

In a Node you can add multiple edges to different nodes using add_edge. Similarly, you can return a list of nodes to navigate in a conditional edge.

def some_condition():
    if condition:
        return "node2"
    else:
        return ["node3", "node4", "node5"]

and then

builder.add_conditional_edges(
  "node1",
  some_condition
)

However, this behaviour is not supported when using path_map: Optional[dict[Hashable, str]] since the values here have to be string. Ideally, one could do something like:

def some_condition():
    if condition:
        return "option_1"
    else:
        return "option_2"

builder.add_conditional_edges(
  "node1",
  some_condition,
  {
    "option_1": "node2",
    "option_2": ["node3", "node4", "node5"]
  }
)

but this throws a TypeError: unhashable type: 'list'.

[talhalaf]

This fix is needed for the draw_marmaid function to draw the graph correctly

[talhalaf]

This fix is needed for the draw_marmaid function to draw the graph correctly

Ok, i found a workaround for the drawing to be correct

def some_condition():
    if condition:
        return ["node2"]
    else:
        return ["node3", "node4", "node5"]

intermidates= ["node2", "node3", "node4", "node5"]

builder.add_conditional_edges(
  "node1",
  some_condition,
  intermidates
)