Adds workaround to remove empty local arrays

[kwmsmith]

Slices the distribution object to remove all empty localarrays before
returning. Only done if all dimensions use 'b' or 'n' in the dist
tuple.

Generalizes the fix in PRs #444 and #447.

[bgrant]

(slice(None),)*self.ndim is the same as Ellipsis, right?

[bgrant]

Heh, it doesn't work though, so nevermind.

[kwmsmith]

Yeah, I tried Ellipsis, but it barfed. So this is the workaround for the workaround.

[bgrant]

Merged. I think it barfed because I don't call sanitize_indices inside
Distribution.slice-- I call it before calling Distribution.slice.

On Mon, Jul 7, 2014 at 4:02 PM, Kurt Smith [email protected] wrote:

In distarray/dist/maps.py:

@@ -612,7 +612,17 @@ def new(cls, context, shape, dist=None, grid_shape=None, targets=None):

     # list of `ClientMap` objects, one per dimension.
     maps = [map_from_sizes(*args) for args in zip(shape, dist, grid_shape)]

•    return cls.from_maps(context=context, maps=maps, targets=targets)
  

•    self = cls.from_maps(context=context, maps=maps, targets=targets)
  

•    # TODO: FIXME: this is a workaround.  The reason we slice here is to
  

•    # return a distribution with no empty local shapes.  The `from_maps()`
  

•    # classmethod should be fixed to ensure no empty local arrays are
  

•    # created in the first place.  That will remove the need to slice the
  

•    # distribution to remove empty localshapes.
  

•    if all(d in ('n', 'b') for d in self.dist):
  

•        self = self.slice((slice(None),)*self.ndim)
  

Yeah, I tried Ellipsis, but it barfed. So this is the workaround for the
workaround.

—
Reply to this email directly or view it on GitHub
https://github.com/enthought/distarray/pull/502/files#r14622248.