Changes needed to rebaseline skimage with rebaseline.py.

Changes needed to rebaseline skimage with rebaseline.py.

rebaseline.py expects all result json files to have the same name, and
be located in a folder corresponding to the builder name. It expects
the expectations json files to also have the same name (as each other,
not the result json files) in a folder corresponding to the builder
name. This CL modifies prerender.py and run_decoding_tests to copy to/
read from expectations json files that match this pattern. It also
modifies run_decoding_tests and upload_skimage_results to write/upload
the result files that match this pattern.

TODO: Upload the images results by their hash digests.

Requires a change in trunk to move the expectations files to conform
to this pattern: https://code.google.com/p/skia/source/detail?r=11698

BUG=skia:1466
R=epoger@google.com

Committed: https://code.google.com/p/skia/source/detail?r=11712

[scroggo, 2013-10-10 00:01:30]

https://codereview.chromium.org/26734006/diff/3001/slave/skia_slave_scripts/p...
File slave/skia_slave_scripts/prerender.py (right):

https://codereview.chromium.org/26734006/diff/3001/slave/skia_slave_scripts/p...
slave/skia_slave_scripts/prerender.py:72:
self._flavor_utils.CreateCleanDeviceDirectory(skimage_device_subdir)
It might be cleaner to just call CopyDirectoryContentsToHost (to copy the whole
directory with one file), but I'm fine either way. Thoughts?

[epoger, 2013-10-10 14:51:42]

LGTM at patchset 2, with or without suggestions

https://codereview.chromium.org/26734006/diff/3001/slave/skia_slave_scripts/p...
File slave/skia_slave_scripts/prerender.py (right):

https://codereview.chromium.org/26734006/diff/3001/slave/skia_slave_scripts/p...
slave/skia_slave_scripts/prerender.py:72:
self._flavor_utils.CreateCleanDeviceDirectory(skimage_device_subdir)
On 2013/10/10 00:01:31, scroggo wrote:
> It might be cleaner to just call CopyDirectoryContentsToHost (to copy the
whole
> directory with one file), but I'm fine either way. Thoughts?

Either way seems equally fine to me.  Sorry to be of no help. :-)

https://codereview.chromium.org/26734006/diff/3001/slave/skia_slave_scripts/r...
File slave/skia_slave_scripts/run_decoding_tests.py (right):

https://codereview.chromium.org/26734006/diff/3001/slave/skia_slave_scripts/r...
slave/skia_slave_scripts/run_decoding_tests.py:34: device_subdir,
'actual-results.json')
maybe share the definition of this filename with run_gm.JSON_SUMMARY_FILENAME ?

https://codereview.chromium.org/26734006/diff/3001/slave/skia_slave_scripts/u...
File slave/skia_slave_scripts/upload_skimage_results.py (right):

https://codereview.chromium.org/26734006/diff/3001/slave/skia_slave_scripts/u...
slave/skia_slave_scripts/upload_skimage_results.py:50: self._builder_name))
funny alignment on this line

[scroggo, 2013-10-10 16:09:59]

https://codereview.chromium.org/26734006/diff/3001/slave/skia_slave_scripts/r...
File slave/skia_slave_scripts/run_decoding_tests.py (right):

https://codereview.chromium.org/26734006/diff/3001/slave/skia_slave_scripts/r...
slave/skia_slave_scripts/run_decoding_tests.py:34: device_subdir,
'actual-results.json')
On 2013/10/10 14:51:42, epoger wrote:
> maybe share the definition of this filename with run_gm.JSON_SUMMARY_FILENAME
?

Done.

https://codereview.chromium.org/26734006/diff/3001/slave/skia_slave_scripts/u...
File slave/skia_slave_scripts/upload_skimage_results.py (right):

https://codereview.chromium.org/26734006/diff/3001/slave/skia_slave_scripts/u...
slave/skia_slave_scripts/upload_skimage_results.py:50: self._builder_name))
On 2013/10/10 14:51:42, epoger wrote:
> funny alignment on this line

Done.

[epoger, 2013-10-10 16:18:53]

lgtm

[scroggo, 2013-10-10 19:47:26]

Committed patchset #3 manually as r11712 (presubmit successful).