Fix issues with h2o.saveMojo in R, ModelBase.save_mojo() in Py


At some point, a `h2o.saveMojo()` function / .save_mojo method was added to R and Python without sufficient review.

This seems to be a duplication of the h2o.download_mojo function. There are a few issues with this function:

  1. There are no tests (zero junit, runit, pyunit).

  2. Not sure if this is total duplication of h2o.download_mojo or if this offers something new, but the docstrings both say "save MOJO to disk" so it's unclear (needs a docs update)

  3. Is missing default path, so if you don't specify `path`, you will get an error. In R, you'd simply set `path = getwd()` as the default to avoid this (similar in Python probably).

  4. The function is never mentioned in the user guide:

If this function is a total duplication, then I propose we convert it into a wrapper/alias for h2o.download_mojo or simply deprecate the function. It has the same interface as h2o.saveModel, so if it's useful to have this interface and naming convention that is close to the regular saveModel function, then maybe it's better to keep it as an alias...

Your pinned fields
Click on the next to a field label to start pinning.


Michal Kurka


Erin LeDell