the :param relations to generated dvc bash scripts and dvc-in arguments are confusing in docs

[antonkulaga]

I have multiple string parameters in notebooks. In the beginning I thought that all the params I declare as :param in docstrings should be automatically passed to dvc bash script as dvc -p because it looked intuitive. However, it is not the way how generated DVC bash work: I do not see acceptance of the :param arguments which I declared in DVC bash scripts which are generated. I think the documentation should clarify how/if DVC bash scripts interact with :param arguments, otherwise it leads to a lot of confusion and wasted time

[SdgJlbl]

Hello, Thank you for your issue and sorry for the late reply - I was not notified by Github for some reasons...

Thank you for your feedback, I'll try to clarify the documentation with respect to that point. In the mean time:

• :param in the docstring are used by the first part of MLV-tools, which converts Jupyter Notebooks to Python scripts. In your .py file, these parameters are then accessible, and you can modify them from the command line too. In particular, they can be used as "parameters" for DVC.
  • in order for your parameters/arguments to be recognized by DVC, you need to specify what role they will have in DVC: dependencies dvc-in, output dvc-out, not versioned by DVC dvc-extra.

We plan to add support for DVC parameters and DVC metrics very soon.

By curiosity, could you tell me which part of the documentation did you use the most? Did you rely on the project README, or did you also have a look at the tutorials (https://github.com/peopledoc/mlvtools-tutorial)? Thank you, and sorry again for missing your issues.