graph should be directed acyclic

[asha24choudhary]

Hi. I have a dataset and I do not have ground truth of the data. What I'm doing is causal discovery using FCI from causal learn library for instance and then I get an adjacency matrix like
([[ 0, 0, 0, 2, 0],
[ 0, 0, 2, 0, 2],
[ 0, 2, 0, 2, -1],
[ 1, 0, 1, 0, -1],
[ 0, 1, 1, 1, 0]]). then i create a networkx graph save the graph into .gml format (test.gml) and when i do

Create the dowhy Causal Model

        model = CausalModel(
            data=df,
            treatment=treatment,
            outcome=outcome,
            graph='test.gml'
        )

Identifying the causal effect

        identified_estimand = model.identify_effect()
        print(identified_estimand)

i get
NetworkXError: graph should be directed acyclic. what should I do to get rid of this error.

I removed the graph as well and identified the effect, but can I rely on this effect or could you please tell me how shall i proceed. i want to determine the effect of the treatment over the outcome, can i use the graph from causal discovery algorithm or what shall i do?

Thank you in advance :)

[github-actions]

This issue is stale because it has been open for 14 days with no activity.

[amit-sharma]

FCI outputs a Partial ancestral graph (PAG) whereas DoWhy expects a standard directed acyclic graph (DAG).
You will need to convert your PAG to a simple graph with directed edges.
Can you share a visualization of your graph?