Many cases such as model predictions or database retrievals run more efficient if all compounds are assessed at once.
Hence it would be useful to support FeatureUnion from sklearn.
Here one pipeline would calculate the fp in single instance mode, while another runs all compounds at once through the pipeline.
Main obstacle: InvalidInstances
Many cases such as model predictions or database retrievals run more efficient if all compounds are assessed at once.
Hence it would be useful to support FeatureUnion from sklearn.
Here one pipeline would calculate the fp in single instance mode, while another runs all compounds at once through the pipeline.
Main obstacle: InvalidInstances