Repo for machine learning models for regional prediction of hydrologic forcing functions. Includes probabilistic seasonal high flow regions for CONUS, and prediction of high flow metrics for selected regions.
After recent PR merges, our pipeline shows most targets as outdated. I'm not sure of the cause.
One possibility is that a new package was added to the container, and if the functions in a package that we used changed since the last time we updated our container, the targets that use those functions would appear as outdated. Most packages are pinned to specific dates or versions, but some packages would not install as pinned versions and therefore could change.
Another possibility is that a forked repo had outdated targets but was locally up to date and could build new targets that depend on the old targets. To prevent this, committing the targets meta file may be a good idea moving forward.
I'm currently running a tar_make to update the targets
After recent PR merges, our pipeline shows most targets as outdated. I'm not sure of the cause.
One possibility is that a new package was added to the container, and if the functions in a package that we used changed since the last time we updated our container, the targets that use those functions would appear as outdated. Most packages are pinned to specific dates or versions, but some packages would not install as pinned versions and therefore could change.
Another possibility is that a forked repo had outdated targets but was locally up to date and could build new targets that depend on the old targets. To prevent this, committing the targets meta file may be a good idea moving forward.
I'm currently running a tar_make to update the targets