Decorator as a stage wrapper

[davibicudo]

I'm considering refactoring existing code to use synpp, and to avoid changing too much the code, one option that came to my mind would be to use decorator functions for converting existing functions to pipeline stages, in a similar way to pytest's parametrize. Something like:

@synpp.stage(df="module.stage1", df2="module.stage2", param="config_key")
def old_function(df1, df2, param): ...

This would be converted during process_stages (i.e. before running the pipeline) to something like:

class old_function:
    def configure(self, context):
        context.stage("module.stage1")
        context.stage("module.stage2")
        context.config("config_key")

    def execute(self, context):
        return old_function(context.stage("module.stage1"), context.stage("module.stage2"), context.config("config_key"))

Do you think this could work?

[sebhoerl]

Yes, could be a nice idea!

[davibicudo]

Implemented by #58.