Data format for conversion_feature and null_feature

[jayden199012]

When I read your note in markov.py. For feature conversion_feature and null_feature, you have specifically mentioned that

do not pre-aggregate conversions by path as this will effect the outcome of the simulation.

Does this means that the data format for these two features can only be 0 or 1?

I noticed that your implementation is very similar to the original ChannelAttribution R package, but the original R package allows aggregated values (sum) by str_path for these two features. I wonder is there a difference in implementation that resulting the different data requirement?

Thank you in advance for helping me!