What if analysis within Visual analyses Lab

[lynnjrp]

I'm running into "External code failed" error when I try to run a What If? analysis within the Visual Analyses.

---

Operating system used: Linux

[AdrienL]

Hi,

• What is the DSS version?
• Can you paste the full content of the small text box below "Could not compute EXPLANATIONS"?

[lji]

version 12.4

Could not compute EXPLANATIONS : <class 'Exception'> : Explainer isn't ready

Traceback (most recent call last):  File "/opt/dataiku/python/dataiku/doctor/interactive_model/server.py", line 100, in start    self._handle_compute_explanation(interactive_scorer, params["computation_params"], params["records"])  File "/opt/dataiku/python/dataiku/doctor/interactive_model/server.py", line 55, in _handle_compute_explanation    prediction_df, explanations_df = interactive_scorer.explain(computation_params, records)  File "/opt/dataiku/python/dataiku/doctor/interactive_model/server.py", line 195, in explain    explanation_method=computation_params.get("explanationMethod"))  File "/opt/dataiku/python/dataiku/core/saved_model.py", line 1348, in predict    n_explanations_mc_steps))  File "/opt/dataiku/python/dataiku/core/saved_model.py", line 1320, in _predict_partition    explanation_method, n_explanations, n_explanations_mc_steps)  File "/opt/dataiku/python/dataiku/core/saved_model.py", line 1026, in predict    n_explanations, n_explanations_mc_steps)  File "/opt/dataiku/python/dataiku/core/saved_model.py", line 878, in _compute_explanations    shapley_background_size=mc_steps)  File "/opt/dataiku/python/dataiku/doctor/individual_explainer.py", line 149, in explain    raise Exception("Explainer isn't ready")Exception: Explainer isn't ready  File "/usr/lib64/python3.6/runpy.py", line 193, in _run_module_as_main    "__main__", mod_spec)  File "/usr/lib64/python3.6/runpy.py", line 85, in _run_code    exec(code, run_globals)  File "/opt/dataiku/python/dataiku/doctor/interactive_model/server.py", line 251, in <module>    serve(int(sys.argv[1]), sys.argv[2])  File "/opt/dataiku/python/dataiku/doctor/interactive_model/server.py", line 242, in serve    interactive_model.start()  File "/opt/dataiku/python/dataiku/doctor/interactive_model/server.py", line 108, in start    self._handle_command_exception(e)  File "/opt/dataiku/python/dataiku/doctor/interactive_model/server.py", line 82, in _handle_command_exception    traceback.print_stack()[2024-04-25 20:42:09,851] [25/MainThread] [ERROR] [__main__] Explainer isn't ready

[lji]

I replied with the wrong account.  lji is also the author's account.  Thanks.

[AdrienL]

I see this is a partitioned model. For at least one of the partitions, do you have less than 25 rows in your test set?

You can see that in either the Summary or the Training information tab of each partition.

If one of the partition has less than 25 rows, then it may not be able to load the explanations.

[AdrienL]

Also, in that case, the prediction should work, and only the explanations would fail.

In that case, you can simply disable the explanations to make the error disappear.

[lji]

Thank you!  How can the explanation feature be disabled?  

[AdrienL]

You can click the toggle below the score / above the explanations part, highlighted in green here