Takaya Uchida

> Have you actually tried running this recipe locally? I can't imagine it would have worked as is, due to the different size of the regions. By running this locally...

I'm getting the following import error from rechunker when trying to run the recipe locally but is this version specific...? ``` --------------------------------------------------------------------------- ImportError Traceback (most recent call last) in 1...

Is there a flexible way to prescribe the chunk size? It seems that the flag `inputs_per_chunk` is meant to do this but can I prescribe the chuck size per dimension...?

I'm getting the following error when running: ``` for input_file in recipe.inputs_for_chunk(all_chunks[0]): recipe.cache_input(input_file) --------------------------------------------------------------------------- OSError Traceback (most recent call last) in 1 for input_file in recipe.inputs_for_chunk(all_chunks[0]): ----> 2 recipe.cache_input(input_file) /mnt/meom/workdir/uchidat/pangeo-forge/pangeo_forge/recipe.py...

> I'm also going to play with this recipe today to see if I can get it working. Could you let me know if you found further fixes I'd need...

This is GREAT!! Thanks @rabernat :) Should I commit recipes for the other regions and interior data?

Also, if you could tell me the target storage address on OSN, maybe I can try and learn to push some data myself :)

> It looks like we could do better about masking. We could accomplish this with a preprocessing function. The `tmask` in the netcdf file is meant to deal with this...

> This is an opportunity to do any cleaning necessary to make the data easier to work with. We updated the surface data to the ftp link with the land...

> I was suggesting we add a `preprocess_chunks` option to the recipe to apply to mask _within Pangeo forge_, as in [this example](https://pangeo-forge.readthedocs.io/en/latest/tutorials/terraclimate.html#define-preprocessing-functions). > > In the meantime I have...