Make the run fully reproducible

[ethanwhite]

My impression is that at the moment there are a handful of different steps that need to be run separately and that some are being run locally and some on the HiPerGator. We should move this to the point where a single job on the HiPerGator runs the entire pipeline, ideally with a build system that avoids doing unnecessary work (e.g., make, etc.).

[davharris]

👍

Dave

On Mar 6, 2018, at 8:36 AM, Ethan White notifications@github.com wrote:

My impression is that at the moment there are a handful of different steps that need to be run separately and that some are being run locally and some on the HiPerGator. We should move this to the point where a single job on the HiPerGator runs the entire pipeline, ideally with a build system that avoids doing unnecessary work (e.g., make, etc.).

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.

[ethanwhite]

As @davharris noted on Slack, we also need to consider how to keep track of different runs with different settings so that individual runs are reproducible. One option is to expand settings.yml as necessary and store/use different settings files for different projects.

In the long run I think we want to pull of the core BBS Forecasting functionality out into a separate package/repo that all BBS related forecasting projects leverage (but that's probably out of scope in the short run).