the users of the workflow can upload a file from their own laptop or their own box folder, where the workflow has no prior knowledge where the files can be. How to handle this in Prefect?

Hey @Zheng Xie, are you asking how to run a Flow in Prefect without knowing ahead of time the location of the Flow?

for example, if the flow is about training a ML model, The input file for the flow is the training data, which exists in some files. when the user trains his own model using his own data, he should be able to tell the flow these are the files storing the data. Because of the diversity of the users, it is not desirable to ask them all to load their files to one location. So it is the workflow runs in a shared infrastructure, but allow the users to upload their own files from their own pc, for example. I am aware the “parameter” of the workflow can be the filename, but that file has to be in the shared infrastructure. Is that we have to ask the user to upload their files before trigger the workflow?

I don’t think you should need to upload. For example one person has files in dropbox and one has files in AWS S3. I think I would have 2 parameters. One will be the filename and one will the the host (dropbox or AWS). From there, you can create a task to take in the host and create the appropriate loader for it right?