unable to run DataDriftPreset for selected columns

[userkkw]

evidently == 0.2.0

Hi all, when I try to run the data drift report with selected columns from the sample housing data, I encounter the following error:

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-31-b410202d09b8> in <module>
      7     ]
      8 )
----> 9 report.run(reference_data=reference_data, current_data=current_data)

/opt/conda/lib/python3.7/site-packages/evidently/report/report.py in run(self, reference_data, current_data, column_mapping)
     90 
     91         self._inner_suite.verify()
---> 92         self._inner_suite.run_calculate(data)
     93 
     94     def as_dict(self) -> dict:

/opt/conda/lib/python3.7/site-packages/evidently/suite/base_suite.py in run_calculate(self, data)
    238 
    239             calculations = {}
--> 240             for metric, calculation in execution_graph.get_metric_execution_iterator():
    241                 if calculation not in calculations:
    242                     logging.debug(f"Executing {type(calculation)}...")

/opt/conda/lib/python3.7/site-packages/evidently/suite/execution_graph.py in get_metric_execution_iterator(self)
     37         metric_to_calculations = {}
     38         for metric_type, metrics in aggregated.items():
---> 39             metrics_by_parameters: Dict[tuple, List[Metric]] = functools.reduce(_aggregate_by_parameters, metrics, {})
     40 
     41             for metric in metrics:

/opt/conda/lib/python3.7/site-packages/evidently/suite/execution_graph.py in _aggregate_by_parameters(agg, metric)
     54 
     55 def _aggregate_by_parameters(agg: dict, metric: Metric) -> dict:
---> 56     agg[metric.get_parameters()] = agg.get(metric.get_parameters(), []) + [metric]
     57     return agg

TypeError: unhashable type: 'list'

The python code is following:

columns = ["AveBedrms","MedInc"]
report = Report(
    metrics=[
        DataDriftPreset(
            columns = columns,
        ),
    ]
)
report.run(reference_data=reference_data, current_data=current_data)

[emeli-dral]

Hi @userkkw ,
Thank you for pointing this out!

fixed in #476

You can use it now if you rebuild the package from the source, and fix will be included in the next release.

[emeli-dral]

Fix is released in 0.2.1