dask / dask / 11327 / 3
53%
master: 53%

DEFAULT BRANCH: master
Ran
Files 108
Run time 7s
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 90.626%. Remained the same
PYTHON=3.6 NUMPY=1.15.1 PANDAS=0.23.4 TEST='true' LINT='true' COVERAGE='true' PARALLEL='false' XTRATESTARGS= TEST_IMPORTS='true'

cron

travis-ci

mrocklin 
Support for map_blocks with no array arguments (#4713)

* Support for map_blocks with no array arguments

This comprises multiple pieces:
- Make `unify_chunks` work correctly when given only constants
- If there are no arguments, consider the output to be zero-dimensional
  prior to application of `new_axis`.
- Allow blockwise to take a list of chunk sizes in `new_axes` instead of
  just a length.
- Add an entry to `block_info` to describe the output.

* map_blocks: infer new_axis from chunks

If chunks is specified and new_axis is not, new_axis defaults to
inserting axes on the left to make the output dimension match the length
of chunks. This is mainly expected to be used for zero-input map_blocks.

This made one of the existing tests fail because it was expecting a
ValueError in a case that is now legal. Changed the test to test the
case where `chunks` has too few dimensions instead of too many, and also
fixed a bug in `map_blocks` where `chunks=()` was being treated as if
`chunks` wasn't supplied.

* Add 'chunk-shape' and 'dtype' to map_blocks output block_info

* Update block_info tests for output info

This includes fixing a bug from #4301 where `old_k` has an off-by-one
error (and the unit test didn't actually run any of the assertions,
which is how it slipped through).

17103 of 18872 relevant lines covered (90.63%)

0.91 hits per line

Source Files on job 11327.3 (PYTHON=3.6 NUMPY=1.15.1 PANDAS=0.23.4 TEST='true' LINT='true' COVERAGE='true' PARALLEL='false' XTRATESTARGS= TEST_IMPORTS='true')