How do I define a custom Module in Azure ML Designer? In Studio (classic), I found this documentation:
https://learn.microsoft.com/en-us/azure/machine-learning/studio/custom-r-modules
but I have yet to find something similar for the modern version of Azure ML. The closest thing I've found is Execute Python Script, but that has several limitations:
- Unnecessary overhead from casting to a pandas DataFrame and back to the internal data structure
- No ability to add parameters
- Difficult to reuse
- Can't customize number of ports, module name, etc.
Also, is it possible to upstream any custom Modules I write? For example, there are currently no Data Transforms for handling timestamps, which I imagine are common in many datasets. I would love to contribute one upstream so that other users can use it, but I haven't yet found a GitHub repo for it.