Back to the main page.

Bug 1303 - test_datasets path changed

Status CLOSED FIXED
Reported 2012-02-01 13:41:00 +0100
Modified 2012-10-04 14:52:32 +0200
Product: FieldTrip
Component: core
Version: unspecified
Hardware: PC
Operating System: Windows
Importance: P3 normal
Assigned to: Johanna
URL:
Tags:
Depends on:
Blocks:
See also:

Johanna - 2012-02-01 13:41:33 +0100

add 'latest' to end of path for origdir.


Johanna - 2012-02-01 13:43:10 +0100

fixed, svn 5195.


Robert Oostenveld - 2012-02-04 11:23:46 +0100

This fix was not correct. The original datasets do not depend on the FT version, but the mat files that contain the FT output (e.g. raw, timelock, freq) do. I have documented it and changed test_datasets into % We have collected a whole bunch of datasets in the original acquisition % systems file format that can be used for testing whether FT functions % work properly for other types of data and not only for MEG data % that was recorded on our ctf275 system. These test datasets (files, % directories) are located on a shared network disk and are organized % in a structure according to PATH/TYPE/DATATYPE/datafile or directory % % where % PATH = somewhere on home/common % TYPE = eeg, meg, lfp % DATATYPE = for example ctf64, ctf151, ctf275 if isunix origdir = '/home/common/matlab/fieldtrip/data/test/original'; elseif ispc origdir = fullfile('H:', 'common', 'matlab', 'fieldtrip', 'data', 'test', 'original'); end where the only functional change is in changing latest->original mbp> svn commit test_datasets.m Sending test_datasets.m Transmitting file data . Committed revision 5239.


Johanna - 2012-02-06 12:58:04 +0100

some confusion, as now I see JM just removed the 'original'. Why does 'original' not work?


Johanna - 2012-02-06 13:31:17 +0100

New solution: test_datsets should be general (not contain original or latest) but the function calling this should then specify within itself the correct path. e.g. test_ft_preprocessing needs a mixture of 'latest' and 'original' whereas test_ft_timelockanalysis needs 'latest' for types raw and timelock I will change all the following: bash-3.2$ grep -rl test_datasets * test_datasets.m test_ft_channelrepair.m test_ft_freqanalysis.m test_ft_megplanar.m test_ft_megrealign.m test_ft_preprocessing.m test_ft_timelockanalysis.m test_ft_timelockanalysis_new.m This also makes it easier for future test functions to compare 'latest' to another version, I presume. svn commit 5246.


Boris Reuderink - 2012-02-06 16:20:46 +0100

I'll keep an eye on the test script, and close this issue later if there are no problems. On first sight, it is looking good.