links_dict.py

# Working on the links_dict

output.update({'links_dict': dict()})

# Group the DataFrame by user_id and rank_event
grouped = data.groupby(['user_id', 'rank_event'])

# Define a function to read the souces, targets, values and time from event to next_event:
def update_source_target(user):
    try:
        # user.name[0] is the user's user_id; user.name[1] is the rank of each action
        # 1st we retrieve the source and target's indices from nodes_dict
        source_index = output['nodes_dict'][user.name[1]]['sources_index'][output['nodes_dict']
                                                                           [user.name[1]]['sources'].index(user['event_name'].values[0])]
        target_index = output['nodes_dict'][user.name[1] + 1]['sources_index'][output['nodes_dict']
                                                                               [user.name[1] + 1]['sources'].index(user['next_event'].values[0])]

         # If this source is already in links_dict...
        if source_index in output['links_dict']:
            # ...and if this target is already associated to this source...
            if target_index in output['links_dict'][source_index]:
                # ...then we increment the count of users with this source/target pair by 1, and keep track of the time from source to target
                output['links_dict'][source_index][target_index]['unique_users'] += 1
                output['links_dict'][source_index][target_index]['avg_time_to_next'] += user['time_to_next'].values[0]
            # ...but if the target is not already associated to this source...
            else:
                # ...we create a new key for this target, for this source, and initiate it with 1 user and the time from source to target
                output['links_dict'][source_index].update({target_index:
                                                           dict(
                                                               {'unique_users': 1,
                                                                'avg_time_to_next': user['time_to_next'].values[0]}
                                                           )
                                                           })
        # ...but if this source isn't already available in the links_dict, we create its key and the key of this source's target, and we initiate it with 1 user and the time from source to target
        else:
            output['links_dict'].update({source_index: dict({target_index: dict(
                {'unique_users': 1, 'avg_time_to_next': user['time_to_next'].values[0]})})})
    except Exception as e:
        pass

# Apply the function to your grouped Pandas object:
grouped.apply(lambda user: update_source_target(user))