Format of trans_fitness expected by diagnose_from_trans_fitness

[jcostello93]

I'm trying to use diagnose_from_trans_fitness based on the return value from token_based_replay. However, the return value trans_diagnostics is always an empty object.

from pm4py.algo.conformance.tokenreplay import algorithm as token_based_replay
from pm4py.algo.conformance.tokenreplay.diagnostics import duration_diagnostics

parameters_tbr = {
    token_based_replay.Variants.TOKEN_REPLAY.value.Parameters.DISABLE_VARIANTS: True,
    token_based_replay.Variants.TOKEN_REPLAY.value.Parameters.ENABLE_PLTR_FITNESS: True
}
replayed_traces, place_fitness, trans_fitness, unwanted_activities = token_based_replay.apply(
    log, net, initial_marking, final_marking, parameters=parameters_tbr
)

trans_diagnostics = duration_diagnostics.diagnose_from_trans_fitness(log, trans_fitness)
for trans in trans_diagnostics:
    print(trans, trans_diagnostics[trans])

From what I can tell, the if trace in trans_fitness[trans]["underfed_traces"]: check in diagnose_from_trans_fitness is expecting a list of (complete) traces, however, it's currently a dictionary. I've tried passing in a formatted trans_fitness to get the if check to work, but I'm not sure how to associate the transition with a full trace from the event log.

Have also tried convert_to_event_log based on #444, but got the same result.

Thank you!

[fit-alessandro-berti]

Dear @jcostello93

Is the model that you provide to the token-based replay fit according to the event log? In that case, it is normal to expect an empty result.

Try the following script on the example event log, and see that there are some diagnostics:

"""
import pm4py

log = pm4py.read_xes("tests/input_data/receipt.xes", return_legacy_log_object=True)
net, initial_marking, final_marking = pm4py.discover_petri_net_inductive(pm4py.filter_variants_top_k(log, 1))

from pm4py.algo.conformance.tokenreplay import algorithm as token_based_replay
from pm4py.algo.conformance.tokenreplay.diagnostics import duration_diagnostics

parameters_tbr = {
token_based_replay.Variants.TOKEN_REPLAY.value.Parameters.DISABLE_VARIANTS: True,
token_based_replay.Variants.TOKEN_REPLAY.value.Parameters.ENABLE_PLTR_FITNESS: True
}

replayed_traces, place_fitness, trans_fitness, unwanted_activities = token_based_replay.apply(
log, net, initial_marking, final_marking, parameters=parameters_tbr
)

trans_diagnostics = duration_diagnostics.diagnose_from_trans_fitness(log, trans_fitness)
for trans in trans_diagnostics:
print(trans, trans_diagnostics[trans])
"""