Edges exceed axis limits in networkx graph

Question

I am trying to make a co-hashtag network using networkx in Python.

I can see that the edges and nodes exceed the frame/axes in the plot, but modifying the coordinates does not seem to help. Any help is appreciated!

Short description of data:

high_weight:
                   0                 1  count  count_divided
434934      fakenews          fakenews  58153        290.765
431143      fakenews  enemyofthepeople  28553        142.765
462561      fakenews              maga  24406        122.030
503028      fakenews         trump2020  18401         92.005
435421      fakenews     fakenewsmedia  15930         79.650
...              ...               ...    ...            ...
1331708      wwg1wga         trump2020   1809          9.045
434967      fakenews     fakenewsalert   1783          8.915
435987      fakenews         fakepolls   1753          8.765
518845   fakenewscnn          fakenews   1747          8.735
482247      fakenews             qarmy   1730          8.650
                                ...

text.hashtags: 

1                                                  [fakenews]
2                                                  [fakenews]
4                                                  [fakenews]
5           [fakenews, qanon, wwg1wga, greatawakening, psb...
6           [press, fakenews, decency, leadership, potus46...
                                  ...
16373658          [liberal, fakenews, liberal, fakenewsmedia]
16373664       [fakenews, hotwheels, diecast, diecastorpaper]
16373667                                           [fakenews]
16373674                                           [fakenews]
16373676                                           [fakenews]

Below is the code I'm using to create and plot the network:

    # Create network
    g_high_weight = nx.from_pandas_edgelist(high_weight, source=0, target=1, edge_attr='count_divided')

    pos_new = nx.circular_layout(g_high_weight)

    
    # Count number of tweets
    t=list(text.hashtags)
    flat_list = [item for sublist in t for item in sublist]
    count = collections.Counter(flat_list)
    
    # Taking only high weight nodes
    nodelist_w_size = {key: count[key] for key in count.keys() & set(high_weight[0])}
    
    # Adding attributes
    nx.set_node_attributes(g_high_weight, values = nodelist_w_size, name='size')
    # Edge weights
    widths = nx.get_edge_attributes(g_high_weight, 'count_divided')
    # Node size
    node_size = nx.get_node_attributes(g_high_weight, 'size')
    # Node list
    nodelist = g_high_weight.nodes()
    
    # Dictionary changes
    # The network needs a list with nodes / edges and an array with values
    # Dividing values with 5 to minimize node size
    size_node_dict = {k:(float(v)/50) for k, v in nodelist_w_size.items()}
    key_list_nodes = list(size_node_dict.keys()) 
    size_node_array = np.array(list(size_node_dict.values()))

    widths_dict = {k:(float(v)) for k, v in widths.items()}
    widths_list = list(widths_dict.keys()) 
    widths_node_array = np.array(list(widths_dict.values()))
    
    
    plt.figure(figsize=(40,30))
   # plt.axis("scale")
    
    # Nodes and edges coordinates - this is what the network does not react to
    pos_nodes_edges = {}
    for node, coords in pos_new.items():
        pos_nodes_edges[node] = np.array([coords[0]/2, coords[1]/2])
    
    print(pos_nodes_edges)

    nx.draw_networkx_nodes(g_high_weight,
                           pos = pos_nodes_edges,
                           nodelist = key_list_nodes,
                           node_size = size_node_array,
                           node_color ='grey',
                           alpha = 0.8)

            
        
    # Edges
    nx.draw_networkx_edges(g_high_weight,
                           pos = pos_nodes_edges,
                           edgelist = widths_list,
                           width = widths_node_array,
                           edge_color='lightblue',
                           alpha=0.6)

    # Labels
    # Make offset of labels so that they are not in the middle of the node
    pos_attrs = {}
    for node, coords in pos_new.items():
        pos_attrs[node] = (coords[0]/2, coords[1]/2 + 0.02)

    # Add labels to network    
    nx.draw_networkx_labels(g_high_weight, 
                            pos_attrs, 
                            labels=dict(zip(nodelist, nodelist)), 
                            font_color='black', 
                            font_size = 18
                           )
    plt.tight_layout()

    plt.savefig(os.path.join("figs",f"{name}_network.png"),dpi=300, bbox_inches="tight")
    print(f'[INFO] network visualization saved in figs as {name}_network.png')
        
    plt.box(False)
    plt.show()

And the resulting plot can be found here:

Resulting network