Back to the main page.

Bug 2036 - small bug fixes in exporting fiff: fileparts returns dot and dataout does not exist (#1)

Reported 2013-03-07 22:39:00 +0100
Modified 2014-02-24 10:56:25 +0100
Product: FieldTrip
Component: fileio
Version: unspecified
Hardware: PC
Operating System: Mac OS
Importance: P3 minor
Assigned to: Robert Oostenveld
Depends on:
See also:

Robert Oostenveld - 2013-03-07 22:39:03 +0100

On 7 Mar 2013, at 10:08, gpiantoni wrote: very low priority fix You can merge this Pull Request by running git pull fiff Or view, comment on, or merge it at: Commit Summary small bug fixes: fileparts returns dot and dataout does not exist File Changes M fieldtrip2fiff.m (4) Patch Links:

Robert Oostenveld - 2013-03-07 22:45:05 +0100

I have managed to pull the diff into my laptop version, then push it to github/oostenveld. git checkout master git branch bug2036 git checkout bug2036 git pull fiff git push origin bug2036 I should now merge it into the master and pull the master into the special github-svn version.

Robert Oostenveld - 2013-03-07 22:58:45 +0100

(In reply to comment #1) I reviewed the diff, and consider it safe to merge it into the release version ;-) So I do in the fieldtrip-github-svn version git pull oostenveld bug2036 git push -> sends it to github git svn dcommit -> sends it to svn Committing to svn+ssh:// ... M fieldtrip2fiff.m M realtime/bin/maci/bufferViewer M realtime/bin/maci/modeeg2ft M realtime/bin/maci/modeeg2ft_2chn M realtime/bin/maci/modeeg_setupbluesmirf M realtime/bin/maci/recording M realtime/src/acquisition/modeeg/ M realtime/src/acquisition/modeeg/serial.c Committed r7605 M fieldtrip2fiff.m M realtime/src/acquisition/modeeg/ M realtime/src/acquisition/modeeg/serial.c M realtime/bin/maci/modeeg2ft_2chn M realtime/bin/maci/modeeg_setupbluesmirf M realtime/bin/maci/bufferViewer M realtime/bin/maci/modeeg2ft M realtime/bin/maci/recording r7605 = 6eff5b6217eb466487367af64193ece5be032fb9 (refs/remotes/trunk) No changes between current HEAD and refs/remotes/trunk Resetting to the latest refs/remotes/trunk

Robert Oostenveld - 2013-03-07 23:11:01 +0100

accidentally there were also some recompiled files from the realtime/bin/maci that went along. That should not be a problem, but those were not meant to be on the master branch. On it looks ok. With svn update in an old fashioned repo it updates the affected file as expected. So it worked.

Robert Oostenveld - 2013-03-09 12:32:13 +0100

*** Bug 2038 has been marked as a duplicate of this bug. ***

Robert Oostenveld - 2013-03-09 12:52:17 +0100

so there are still problems with this. TODO: we need a test script (test_bug2036.m) with a small fif file. And it is indeed smart to check the procedure with export and subsequent read operation. @Hamid, I wanted to postpone this a bit (as I am traveling and very busy), but I'd to talk to you about you directly contributing. I guess that you are not familiar with git, but I had planned to ask you to make a github account, fork and make the suggested changes in a branch of your own git version. I can then "pull in" the updates from your github account after reviewing them. This is a new procedure which (hopefully) improves on managing external contributions. We also want to make more use of git internally to facilitate development, Search bugzilla for git, you'll find some items that relate to it. Perhaps Gio could also help you a bit with this, as he is the first "git" contributor besides me (although he also has an internal donders svn account). It should all get documented on, just like

Hamid - 2013-03-09 13:06:35 +0100

(In reply to comment #5) It is a good idea that external people can also contribute. I signed up for git and will learn more about it and try to contribute especially on MEG neuromag. Thanks and have a good journey

Gio Piantoni - 2013-03-09 18:10:11 +0100

Hi, I thought about writing a tutorial when the pull-request was successful and have been meaning to do it, but I haven't had the time in the last two days. I'll write it in the next few days. If you cannot wait to learn about git, the documentation at is great. Gio

Gio Piantoni - 2013-03-10 16:24:28 +0100

(In reply to comment #6) Hi Hamid, I wrote a short tutorial to get people started here with git and fieldtrip: Let me know if something is not clear. Best, Gio

Hamid - 2013-03-10 19:02:04 +0100

(In reply to comment #8) Thanks, the tutorial was clean and easy to follow, and git is indeed powerful. just a minor thing: after git checkout -b welcome in section 3, it is better to add: cd ./fieldtrip

Gio Piantoni - 2013-03-10 19:24:35 +0100

(In reply to comment #9) Hi Hamid, Thanks for the feedback! I added the cd part. Feel free to edit this page or other pages on the (it's a wiki page with minimal markup). Best, G

Robert Oostenveld - 2013-04-18 20:52:52 +0200

the update to fieldtrip2fiff.m by Gio has been merged into the master branch and svn trunk. There is still the test_fieldtrip2fiff test script that needs to be updated.

Robert Oostenveld - 2013-04-18 21:02:27 +0200

(In reply to comment #11) roboos@mentat001> svn commit test_fieldtrip2fiff.m Sending test_fieldtrip2fiff.m Transmitting file data . Committed revision 8057. I don't have all test files yet.

Gio Piantoni - 2013-04-19 09:37:27 +0200

Hi Robert, I saw r8057, thanks for fixing the paths. I guess you're missing the neuromag and the eeg files. Do you want me to send them to you? the neuromag file (sample data from MNE) came from Do you have a copy already of the file? for the EEG, do you have already an EEG dataset (for example from the tutorial) that you could use and it's already in the text folder? I expect that the script should work with any EEG. Otherwise, I'll send you the files

Robert Oostenveld - 2013-04-19 09:50:00 +0200

(In reply to comment #13) I did not have the MNE sample file yet, but am downloading it now. For the EEG I was a bit unsure to swap it out with another one, as you also specify a specific 256 channel EGI electrode set. I could look for one, but would prefer you to give me a (short) one so that I don't have to change other aspects of the script.

Robert Oostenveld - 2013-09-18 13:21:03 +0200

Hi Gio, I am going over the list of bugs assigned to me. This was one that you fixed for yourself while we in Nijmegen were having difficulties with git. Do you perhaps know whether your fix made it into the fieldtrip master branch? best Robert

Gio Piantoni - 2013-09-18 18:25:46 +0200

(In reply to comment #15) please, see Hope this fixes it

Robert Oostenveld - 2013-09-19 15:35:11 +0200

(In reply to comment #16) I copied the changes over manually. mac001> svn commit fieldtrip2fiff.m Sending fieldtrip2fiff.m Transmitting file data . Committed revision 8474. Regarding the >250 commits that were reported in the pull request: At a certain point git and svn started to get off-track in their history (not in the code). This required many commits to get them history-wise identical again. The present fieldtrip/fieldtrip is again in perfect sync with svn. So if you wish, you might want to discard your fork and start with a new one. Note that the only change is that I forced git to revert all these these irrelevant changes to get it in sync again. Since your tree does not align with ours, you will keep the "many commits" difference.

Gio Piantoni - 2013-09-19 15:52:28 +0200

(In reply to comment #17) Thanks for adding the changes. Can you run test_fieldtrip2fiff and report any error? I don't have the test files on this PC. thanks I'll clean up my repo so that merges will be less painful in the future

Robert Oostenveld - 2013-09-19 17:30:07 +0200

Thanks for reminding me of the test script. At the moment it fails because of svui_0003_eeg_go-sd_010.raw not being present. I don't recall ever having seen such a file, and a search on my hard disk does not find it either. Do you know where it is? --- Error using ft_read_header (line 108) file or directory '/home/common/matlab/fieldtrip/data/test/bug2036/svui_0003_eeg_go-sd_010.raw' does not exist Error in ft_trialfun_general (line 71) hdr = ft_read_header(cfg.headerfile, 'headerformat', cfg.headerformat); Error in ft_definetrial (line 162) [trl, event] = feval(cfg.trialfun, cfg); Error in test_fieldtrip2fiff (line 140) cfg = ft_definetrial(cfg);

Robert Oostenveld - 2013-09-20 09:36:18 +0200

(In reply to comment #19) Hi Gio, Thanks for the file that you sent through a private channel, it is now on home/common in the test directory. I executed the test script, it runs without any errors.

Gio Piantoni - 2013-09-20 16:08:00 +0200

Great, thanks! Let me know if it fails in the future and I'll look into it

Gio Piantoni - 2014-02-21 16:38:18 +0100

Hi, Sorry to reopen an old bug, but I noticed that the test does not pass for test_fieldtrip2fiff: Actually, I cannot make out what the error is (according to the dashboard, it's "unknown", see ). It seems to me that the test passes, but the format of the report is weird. Let me know if I can do something to fix it. Best, -g

Robert Oostenveld - 2014-02-21 17:28:16 +0100

(In reply to Gio Piantoni from comment #22) It got killed on our cluster because it requested 3GB but overall required a peak-memory of 6.5GB. I have increased the memory for the test script. roboos@mentat001> svn commit test_fieldtrip2fiff.m Sending test_fieldtrip2fiff.m Transmitting file data . Committed revision 9219. PS I just ran it manually and it works fine.

Gio Piantoni - 2014-02-21 17:50:56 +0100

great, thanks!

Robert Oostenveld - 2014-02-24 10:56:25 +0100

I closed several bugs at once that all have been resolved for some time. If you disagree, please reopen.