Stavros Theocharis

Make nonprocessed Optional for the Inferences

3

**Description** For the user it should be optional to upload the nonprocessed data for some reasons: a) we do not actually need them for now. Only for enhancement in the...

Fix typing in SDK functions

2

**Description** There is a need to change the typing in some functions **Solution** - dict should be Dict with additional parts inside eg. Dict[str, float] - In some parts that...

Adjust the train_test_split parts in pipelines

5

We use "train_test_split" in "src/analytics/models/pipelines" at lines 40 & 98. In case that after the splitting only one class remains at the train set, an error will follow based on...

Confusion matrix schema alignment

1

**Description** In schema "PerformanceMetric" there is a misalignment between the BinaryClassificationMetricsPipelineResult and the MultiClassificationMetricsPipelineResult. In the first one the true_negative, false_positive, etc. are on the outer level. On the other...

The user has to give together the timestamp in a column with the processed Dataframe

3

**Description** The user has to give together the timestamp in a column with the processed Dataframe **Solution** Inside the log_inferences function we have to take timestamps as a new series...

Update inference with actuals

1

**Description** The user can only upload the actuals together with the other data in the log_inferences function. There should be another function in SDK in order that the user will...

**Description** Need to add tests for streamlit

Refactor XAI functionality

3

**Description** The XAI pipeline runs produces a custom dictionary as an output. This output is not easy to be visualized in the streamlit. In addition, this XAI requires a connection...