Back to the main page.

Bug 3158 - ft_datatype: new 'pointcloud' datatype clashes with source datatype

Reported 2016-07-07 10:51:00 +0200
Modified 2017-01-17 11:29:43 +0100
Product: FieldTrip
Component: core
Version: unspecified
Hardware: PC
Operating System: Windows
Importance: P5 major
Assigned to: Robert Oostenveld
Depends on:
See also:

Jan-Mathijs Schoffelen - 2016-07-07 10:51:01 +0200

This causes problems in ft_sourceinterpolate, and possibly elsewhere as well. a call to ft_datatype() with a formerly known as 'source' datatype structure now returns a 'pointcloud'.

Robert Oostenveld - 2016-07-07 12:35:36 +0200

I did a quick fix, but should look into it more... 191-225> git commit -am "FIX - quick fix for which is now blocking the analysis" [master a84c00f] FIX - quick fix for which is now blocking the analysis 1 file changed, 3 deletions(-) 191-225> git push upstream master X11 forwarding request failed on channel 0 You're about to push master, is that what you intended? [y|n] y Counting objects: 9, done. Delta compression using up to 4 threads. Compressing objects: 100% (9/9), done. Writing objects: 100% (9/9), 1.06 KiB | 0 bytes/s, done. Total 9 (delta 6), reused 0 (delta 0) To c691df5..a84c00f master -> master

Simon - 2016-07-11 10:32:23 +0200

Thus far point clouds were only used in ft_defacevolume. For know we could change it that it asks for the datatype source. So I could erase datatype point cloud, because as JM already mentioned pointclouds and sources are not distinguishable.

Robert Oostenveld - 2016-07-11 16:20:27 +0200

I have just fixed it, removing all occurrences of pointcloud from the code. As datatype it is not distinguishable from source (which has pos) and mesh (which has pos and tri/tet/hex). mac011> git commit -am "FIX - resolved invalid pointcloud datatype, see" [master 92c9d20] FIX - resolved invalid pointcloud datatype, see 3 files changed, 10 insertions(+), 20 deletions(-) mac011> git commit -a [master 87f7667] FIX - ismri was not detected correctly, ismesh should also include data with only vertices 1 file changed, 5 insertions(+), 5 deletions(-) mac011> git commit -a [master 71abd0a] ENH - changed defaults to no for keepbrain and feedback, document 1 file changed, 5 insertions(+), 4 deletions(-) mac011> git commit -a [master 6c8c504] ENH - deal with mesh and source separately (using source.inside), see 1 file changed, 19 insertions(+), 8 deletions(-)

Weiyong Xu - 2016-07-12 13:59:43 +0200

for a headshape mesh without triangles it does not work. That is detected as 'source' and then it uses inside rather than to remove positions.

Simon - 2016-07-12 14:11:09 +0200

(In reply to Weiyong Xu from comment #4) Can you specify which version of fieldtrip you used? Recently a lot of changes were made with this function so a version not later than today would be best. Can you also post the structure of the data you used? 'That is detected as 'source' and then it uses inside rather than to remove positions.' I think with this you refer to ft_defacevolume?

Weiyong Xu - 2016-07-12 14:38:58 +0200

(In reply to Simon from comment #5) Hi, I am using the 0711 version. Robert suggest there maybe a bug here. Yes, I am referring to ft_defacevolume. The data structure is the MEG head points.

Robert Oostenveld - 2016-07-13 11:26:43 +0200

(In reply to Weiyong Xu from comment #6) this is better continued in relation to bug 3141.

Robert Oostenveld - 2016-07-13 11:27:17 +0200

the pointcloud bug has been resolved, support for meshes is discussed elsewhere

Robert Oostenveld - 2017-01-17 11:29:43 +0100

closed multiple bugs that were resolved some time ago