6
86%
master: 86%

Ran 01 Oct 2015 12:30AM UTC

Files 322

Run time 9s

Badge

Embed ▾

Committed 01 Oct 2015 12:08AM UTC coverage: 86.775% (+0.004%) from 86.771%

Job # COVERAGE=1

Build Type

push

travis-ci

Committed by

matthew-brett

Commit Message

Merge pull request #362 from matthew-brett/hackaround-for-intp-bug

MRG: hack to work round np.intp bug in ndimage

Scipy ndimage, at least for scipy 0.9.0, fails when trying to interpolate data
of type np.intp, on a 32-bit system we are testing on:

http://nipy.bic.berkeley.edu/builders/nipy-py2.6-32/builds/291/steps/shell_6/logs/stdio

The interpolation failure gives an error like this:

```
File "/usr/lib/python2.7/dist-packages/scipy/ndimage/interpolation.py", line 392, in affine_transform
    output, order, mode, cval, None, None)
    RuntimeError: data type not supported
```

In fact, the numpy types np.int32 and np.intp (on 32-bits) are the same 32-bit
signed integer types, so we can just view the data as np.int32 (on 32 bits)
before running the ndimage routines.

Write wrappers around `affine_transform` and `map_coordinates` to do this
conversion, and extend tests.

The tests raise some errors on very old scipy, so update scipy dependency to
0.9.0.

Run Details

23838 of 27471 relevant lines covered (86.78%)

0.87 hits per line

nipy / nipy / 265 / 6
86%
master: 86%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Source Files on job 265.6 (COVERAGE=1)

nipy / nipy / 265 / 6 86% master: 86%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Source Files on job 265.6 (COVERAGE=1)

nipy / nipy / 265 / 6
86%
master: 86%

README BADGES
x