igraph clusters from dataframe

Question

I'm trying to move a codebase from R to python and I need to move the R function from igraph over from the following:

 g <- igraph::graph_from_data_frame(edges)
 clusters <- igraph::clusters(g)

To an equivalent in python, however when I try:

edges = pd.DataFrame({"from":[0,1,2,4], "to":[1,2,3,5]})
g = igraph.Graph.DataFrame(edges)
print(g.clusters())

I get each of the elements of the graph individually.

I don't think this is due to the clusters call being the wrong thing because if I do the following:

graph = igraph.Graph([(0,1), (1,2), (3,4)])
print(graph.clusters())

I get two distinct clusters. I think it must therefore have something to do with the way I'm creating the graph from the dataframe?