Hey there, I want to use kedro-mlflow for serving a pipeline via an API. What I'm not sure I understand is the input constraint - that it has to be part of the inputs for the training pipeline. In my case - the training pipeline takes data from 3 sources (a little like the spaceflights tutorial), processes each and joins them together. The API call will be with data that's similar to the joined table. How am I supposed to handle this situation? @Yolan Honoré-Rougé perhaps you can help :)

Hi Gilad,

kedro-mlflow

is not really designed to serve complex pipeline as API. It uses mlflow under the hood (obviously 😅 ) which makes very strong assumption of what is a "model" : it has only one input (a dataframe or a numpy array), one output (a series / 1D array of predictions ) and no parameters. As a consequence, if you want to serve a pipeline with

kedro-mlflow

you have to satisfy these constraints. If you want to serve kedro pipelines with a plugin, I'd recommend to check

kedro-boot

we released recently and we plan to work on in the future. It may be harder to jump in, but it is extremely flexible and designed for such uses cases. https://github.com/takikadiri/kedro-boot Feel free to ask help if needed. You can find examples here: https://github.com/takikadiri/kedro-boot-examples

Perfect, Thanks!