How to use the retentioneering.visualization.plot.graph function in retentioneering

node_params.update({
                    node[0]: 'nice_node' if weights[node] >= 0 else 'bad_node',
                })
            node_weights.update({
                node[0]: val
            })

        edge_cols = [i for i in test_sample.columns if len(i) == 2]
        if len(edge_cols) == 0:
            print("Sorry, you use only unigrams, change ngram_range to (1, 2) or greater")
            return
        data = []
        for key in edge_cols:
            data.append([key[0], key[1], weights.get(key)])

        plot.graph(pd.DataFrame(data), node_params, node_weights=node_weights, is_model=True, **kwargs)

'norm': norm,
            })
        if node_params is None:
            _node_params = {
                'positive_target_event': 'nice_target',
                'negative_target_event': 'bad_target',
                'source_event': 'source',
            }
            node_params = {}
            for key, val in _node_params.items():
                name = self.retention_config.get(key)
                if name is None:
                    continue
                node_params.update({name: val})
        node_weights = node_weights or self._obj[self._event_col()].value_counts().to_dict()
        path = plot.graph(self._obj.trajectory.get_edgelist(**kwargs), node_params, node_weights=node_weights, **kwargs)
        return path