GitHub user skrawcz added a comment to the discussion: Parallelization Enhancement Ideas
Yep I see now, this is just to build the graph. I think we can think more about the API, like what is the object passed in, what does it have reference to ? Otherwise I want to improve the flexibility in creating graphs in general, e.g. configuration driven -> classes + functions -> nodes... and this fits into that effort. To help me think through your proposal, I want to have a clear mental model of "when" things would occur to help design the API. E.g. do I start with a basic DAG that I then want to parameterize more? Should the build process be split into two: define atomic functions, then define parameterizations (I'm thinking build basic driver, then pass that driver into another driver that has functions that are parameterizations, or should functions that take a driver as argument are built in a second pass...)? E.g. how do we catch errors? When we would know it's an error? Thoughts? What would match your ideal development process mental model? GitHub link: https://github.com/apache/hamilton/discussions/1412#discussioncomment-15675278 ---- This is an automatically sent email for [email protected]. To unsubscribe, please send an email to: [email protected]
