Gravis with Networkx Edge style modification

Question

G = nx.DiGraph()
G.add_weighted_edges_from(df_weighted_edges_tutar)

for node_id in G.nodes:
    node = G.nodes[node_id]
    gelen_eft_adedi = total_gelen_eft_adet[total_gelen_eft_adet.index ==
                                           node_id]["gelen_eft_adet"].values[0]

    gelen_eft_adedi_karsi_taraf = total_gelen_eft_adet[
        total_gelen_eft_adet.index == node_id]["giden_eft_adet"].values[0]
    node[
        'hover'] = f"{node_id}\nGelen EFT Adedi: {gelen_eft_adedi}\nGelen EFT Tutarı: {total_source_node_size_dict[node_id]:,} TL\nGelen EFT Adedi (Karşı Taraf): {gelen_eft_adedi_karsi_taraf}"

    node["size"] = total_source_node_size_dict[node_id]

# Node properties: Size by transfer amount, shape by size, color by community
colors = [
    'blue', 'green', 'red', 'orange', "gold", "olive", "purple", "darkred",
    "darkblue", "darkgrey", "violet"
]

for node_id in G.nodes:
    node = G.nodes[node_id]
    for community_counter, community_members in enumerate(communities):
        if node_id in community_members:
            break
    node['color'] = colors[community_counter % len(colors)]

# Edge properties: Size by total incoming transfer amount, color by community (within=community color, between=black)
for edge_id in G.edges:
    edge = G.edges[edge_id]
    source_node = G.nodes[edge_id[0]]
    target_node = G.nodes[edge_id[1]]
    is_customer = df_sums[(df_sums.target == edge_id[1]) & (
        df_sums.source == edge_id[0])]["is_customer"].values[0]

    customer_map = {1: "Yes", 0: "No"}
    edge['color'] = source_node['color'] if source_node[
        'color'] == target_node['color'] else 'b-'
    edge[
        "hover"] = f"{edge_id[0]} --> {edge_id[1]}\nEFT Amount: {edge['weight']:,} TL\nIs Customer?: {customer_map[is_customer]}"

gv.d3(G,
      use_node_size_normalization=True,
      node_size_normalization_max=100,
      details_height=250,
      graph_height=650,
      node_size_data_source='size',
      use_edge_size_normalization=True,
      edge_size_data_source='weight',
      edge_curvature=-0.3,
      edge_size_factor=1.1,
      node_hover_neighborhood=True,
      node_label_data_source='label',
      many_body_force_max_distance=-500,
      use_many_body_force_max_distance=True,
      show_details=True,
      layout_algorithm_active=True,
      edge_hover_tooltip=True)

I would like to change the edge style to dashed line if the source node is not a customer. it is determined in the is is_customer parameter and already tried to put in edge properties as edge["style"] but did not work. I also tried to state the edge styles in networkx's draw_networkx_edges module but could not integrate it with gravis graph.

Another question is, is there any feature on gravis for selecting and filtering nodes/edges with a search bar like pyvis. I cannot use pyvis on my work computer. it does not let the rendering. So if there is a similar feature in gravis it would be great for me to learn.

One last question, what is the best way of displaying the community colors as a legend at an appropriate place at the network.

As the author of the Gravis i hope @robert-haas would see this post as well :)