Identification is very slow for big DAGs

[jmafoster1]

Is your feature request related to a problem? Please describe.
Identification for large DAGs is very slow.

Describe the solution you'd like
Is there anything else we can parallelise to speed things up a bit? Or maybe we can get a progress bar or something so we can see how it's doing?
mwe.zip
(See example DAG, data, and test script for demo, it's been about 5 mins now and it's still not finished.)

[jmafoster1]

As an update, it was still going after 2 days! I will try and look into it

[jmafoster1]

Update: The problem is with this method, which seems to have a very deep recursive structure
https://github.com/CITCOM-project/CausalTestingFramework/blob/9abdcbc0710297ec04a8dfd3e70bd0d78f407653/causal_testing/specification/causal_dag.py#L23C5-L23C21

[jmafoster1]

May as well have this closed by #337. We can always reopen it if the optimisations are still insufficient.