Plotting Heterogenous Graph

[mprithiv]

Following the movielens example, I have created a heterogeneous graph of structure
HeteroData( y=[100836], user={ num_nodes=610 }, movie={ x=[9742, 404] }, (user, rates, movie)={ edge_index=[2, 100836], edge_label=[100836, 1] } )

I am looking to plot the graph using to_networkx function. Seems like the HeteroData cannot be used for plotting the graph. Is there a way to plot the heterogeneous graphs? In my real use case, I would have two edge types

(buyer, pays, seller) & (buyer, date, seller)

What would be the ideal way to use both edge types when creating an embedding?

`amount_edge_index, amount_edge_label = load_mapping("processed/simulated_mappings.csv",
src_index_col='buyerId',
src_mapping = user_mapping,
dst_index_col = 'sellerId',
dst_mapping = merchant_mapping,
encoders = {
'txn_amt': LabelEncoder(dtype=torch.long)
},

                                 )`

similarly
`date_edge_index, date_edge_label = load_mapping("processed/simulated_mappings.csv",
src_index_col='buyerId',
src_mapping = user_mapping,
dst_index_col = 'sellerId',
dst_mapping = merchant_mapping,
encoders = {'txn_date': SequenceEncoder()
},

                                 )`

I created a separate edge_label and edge_index. Not sure if this is a correct representation

`data['buyer', ' txn_amt', 'seller'].edge_index = amount_edge_index
data['buyer','txn_date', 'seller'].edge_index = date_edge_index

data['buyer', 'txn_amt', 'seller'].edge_label = amount_edge_label
data['buyer', 'txn_date', 'seller'].edge_label = date_edge_label`

[rusty1s]

I think adding support for HeteroData visualization would be indeed great, but I guess it's not top prio to integrate that into PyG directly. One alternative is to call to_networkx based on the output of HeteroData.to_homogeneous().