Hey folks. Nice demo and congrats on the new release. I was wondering if you could consider few feature requests and ideas
1.Have some low code capabilities baked into the system. For instance a lot of the work that happens is just invoking a bunch of APIs , instead of me creating lets say a script that does this , can it be a drag and drop component in orchest which makes REST calls. Something like what postman does , but make it like a pipeline through orchest’s clean UI so that it is repeatable and comes with all the goodies like retry logic.
2. GPU support
3. Support for storing data and versioning it. Making datasets a first class citizen within orchest so that pipelines are fully reproducible.
Not sure if all these ideas make sense 😅 But thought of sharing here.
09/21/2022, 9:40 PM
Hi Krishna 👋🏻 welcome.
Great suggestions. We're actively working on idea 1 - we're trying to decide how we can best offload common tasks for users in a way that nicely integrates with custom steps that run user code (Python). Stay tuned!
2. I can be short about this: we're working on it and it's coming.
3. It would be interesting to explore approaches around this. Which kinds of datasets are you interested in versioning? Have you looked at something like versioning in S3? https://docs.aws.amazon.com/AmazonS3/latest/userguide/Versioning.html
Would you mind if I ask some follow up questions about these items in the future?