DFG graph changes

[NKMatha]

Hi,

Am looking to compare bpmn file with the dfg graph and highlight the dfg graph path(lets say green) which matches the bpmn graph. Is it possible to make this comparison? If yes, please let me know how could i achieve it.

[fit-alessandro-berti]

Dear @NKMatha

Please consider and adapt the following code:

import pm4py

log = pm4py.read_xes("tests/input_data/receipt.xes")

discovers the DFG from the log

dfg, sa, ea = pm4py.discover_dfg(log)

filters the log and discovers a process model from the log

filtered_log = pm4py.filter_variants_top_k(log, 6)

discovers a BPMN graph using the inductive miner algorithm on the filtered log

bpmn_graph = pm4py.discover_bpmn_inductive(filtered_log)

converts the BPMN to a Petri net

net, im, fm = pm4py.convert_to_petri_net(bpmn_graph)

try:

converts the Petri net to a process tree

process_tree = pm4py.convert_to_process_tree(net, im, fm)
# extracts the footprints from the process tree (very efficient!)
fps = pm4py.discover_footprints(process_tree)

except:

the model is not a block-structured process therefore it cannot be converted to a process tree

# extracts the footprints from the Petri net (less efficient)
fps = pm4py.discover_footprints(net, im, fm)

find the directly-follows relationships allowed by the model

allowed_rels = fps["sequence"].union(fps["parallel"])

prints every relationship of the DFG (sorted by number of occurrences) along with the fact that

is allowed or not by the model

dfg_list = sorted(list((x, y) for x, y in dfg.items()), key=lambda x: (x[1], x[0])) for el in dfg_list: if el[0] in allowed_rels: print(el, " is allowed") else: print(el, " IS NOT ALLOWED")

[NKMatha]

Thanks @fit-alessandro-berti,

So, as per the code can i set the color property to elements to dfg in last conditional loop?

[fit-alessandro-berti]

You would need to create a specific visualization for the use case

[NKMatha]

Hi @fit-alessandro-berti i would like to know that can we have dfg in a structured and organized way so that can be look good(like a flow charts something...), As of now my dfg nodes were getting scattered.

[Javert899]

Unfortunately that's not possible. We already use the Dot algorithm which is very advanced. There are specialized algorithms but we'll not implement them in the near future in pm4py.