search

barbaLab / nigeLab

Neurophysiological data analyses by engineers, for physiologists.
GNU General Public License v3.0
14 stars 2 forks source link

[BUG] doAutoClustering Error: START, COUNT, and STRIDE parameters must be double precision, positive, and have length equal to the rank of the dataset. #59

Closed m053m716 closed 4 years ago

m053m716 commented 4 years ago

Bug Report

Description

This bug happens in some instances when doAutoClustering is run on the remote (not sure if it happens on local, but likely it does) via nigeLab.libs.DashBoard UI.

Error Messages

                   ID: 1
                State: finished
             Function: @qWrapper
               Parent: Job 3209
        StartDateTime: 17-Jan-2020 03:30:07
     Running Duration: 0 days 0h 0m 16s

                Error: The START, COUNT, and STRIDE parameters must be double precision, positive, and have length equal to the rank of the dataset.
          Error Stack: h5read (line 58)
                       DiskData>DiskData.subsref (line 518)
                       getChannelsEventData>getChannelsEventData/retrieveChannelData (line 45)
                       getChannelsEventData (line 13)
                       getEventData (line 164)
                       getSpikes (line 88)
                       getSpikeFeatures (line 26)
                       doAutoClustering (line 61)
                       qWrapper (line 35)

Machine

Strategy

m053m716 commented 4 years ago

Able to reproduce on local. Error stack:

>> inspk = b.getSpikeFeatures(5,0:9);
Error using h5readc
The START, COUNT, and STRIDE parameters must be double precision, positive, and have length equal to the rank of the dataset.

Error in h5read (line 58)
[data,var_class] = h5readc(Filename,Dataset,start,count,stride);

Error in nigeLab.libs.DiskData/subsref (line 518)
                           data(vec,:) = h5read(obj.getPath,'/data',...

Error in nigeLab.Block/getChannelsEventData/retrieveChannelData (line 44)
         out = blockObj.Channels(ch).(type).(field);

Error in nigeLab.Block/getChannelsEventData (line 12)
   eventData = retrieveChannelData(blockObj,ch,F{iF},prop);

Error in nigeLab.Block/getEventData (line 164)
      eventData = blockObj.getChannelsEventData(field,prop,ch,...

Error in nigeLab.Block/getSpikes (line 90)
      spikes = getEventData(blockObj,'SpikeFeatures','snippet',ch);

Error in nigeLab.Block/getSpikeFeatures (line 29)
features = getSpikes(blockObj,ch,clusterIndex,'Features');

Error in nigeLab.Block/subsref (line 69)
         [varargout{1:nargout}] = builtin('subsref',blockObj,S);
m053m716 commented 4 years ago

Created new branch: bugfix/doAutoClustering_emptyDiskData to address this issue.

m053m716 commented 4 years ago

Verified successful run on remote machine.