Time window bug

[nurek-mateusz]

The propagation in a temporal network created from a file based on a time window does not work if consecutive windows contain different sets of nodes.

ValueError: Net should have one actors set in snaps!

Minimal reproducible example:

import network_diffusion as nd
from utils import functions
from tqdm import tqdm


# example.csv contains (sender receiver time):
'''
1 2 10
2 1 13
1 3 17
3 1 21
1 3 26
'''

data = 'example.csv'
g = nd.TemporalNetwork.from_txt(data, time_window=10)

seeding_budget = 20
tnem = nd.models.TemporalNetworkEpistemologyModel(
            seed_selector=nd.seeding.RandomSeedSelector(),
            seeding_budget=(100-seeding_budget, seeding_budget),
            trials_nr=10,
            epsilon=0.01,
        )

experiment = nd.Simulator(model=tnem, network=g)
experiment.perform_propagation(n_epochs=10)