573
93%
master: 94%

Ran 12 Jun 2015 02:25PM UTC

Jobs 1

Files 49

Run time 32min

Badge

Embed ▾

pending completion

Build # 573

Build Type

push

travis-ci

Committed by

mrocklin

Commit Message

add dask.array.cache

This allows us to evaluate a dask.array, store in some cache,
and then return a new dask array pointing to that precomputed data.

One can specify a cache as either a MutableMapping or as an ndarray

Example
-------

>>> import dask.array as da
>>> x = da.arange(5, chunks=2)
>>> y = 2*x + 1
>>> z = y.cache()  # triggers computation

>>> y.compute()  # Does entire computation
array([1, 3, 5, 7, 9])

>>> z.compute()  # Just pulls from store
array([1, 3, 5, 7, 9])

You might base a cache off of an array like a numpy array or
h5py.Dataset.

>>> cache = np.empty(5, dtype=x.dtype)
>>> z = y.cache(store=cache)
>>> cache
array([1, 3, 5, 7, 9])

Or one might use a MutableMapping like a dict or chest

>>> cache = dict()
>>> z = y.cache(store=cache)
>>> cache  # doctest: +SKIP
{('x', 0): array([1, 3]),
 ('x', 1): array([5, 7]),
 ('x', 2): array([9])}

Fixes #293

Run Details

4944 of 5317 relevant lines covered (92.98%)

0.93 hits per line

Jobs

ID	Job ID	Ran	Files	Coverage
2	573.2	12 Jun 2015 02:25PM UTC	0	92.98	Travis Job 573.2

Source Files on build 573

Detailed source file information is not available for this build.

mrocklin / dask / 573
93%
master: 94%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Jobs

Source Files on build 573

mrocklin / dask / 573 93% master: 94%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Jobs

Source Files on build 573

mrocklin / dask / 573
93%
master: 94%

README BADGES
x