Ticket #36 (closed defect: fixed)

Opened 2 months ago

Last modified 2 months ago

missing AnalogSignal when loading data into AnalogSignalList's

Reported by: JensKremkow Assigned to: pierre
Priority: critical Milestone: 0.1
Component: signals Version: trunk
Keywords: Cc:

Description

When loading: vmlist = signals.load_vmlist('voltage_data') vmli In [217]: vmlist.analog_signals.keys() Out[217]: [1, 2, 3, 4, 5, 6, 7, 8, 9]

one has 9 neurons, however, the original data contains 10:

In [218]: v = load('voltage_data')

In [219]: numpy.unique(v[:,1]) Out[219]: array([ 0., 1., 2., 3., 4., 5., 6., 7., 8., 9.])

Change History

11/02/08 15:52:46 changed by pierre

  • status changed from new to assigned.
  • owner set to pierre.

Ok, you should have a look to the header of the file. Is it correct ? I mean, first_id is 0 and last_id is 9/10 ? Because since I didn't have commited pyNN changes yet, I know that the headers written by pyNN are not perfectly understood by NeuroTools yet. Shoudl be changed this week. The thind is that is will suppose that nobody has code written using those headers, because it will maybe break some compatibilities. What should we store also is a question: first_id = 0 and last_id = len(Population) for all populations ? of first_id, last_id should be the real id used by the simulator ?

Cheers

11/04/08 19:53:49 changed by pierre

  • status changed from assigned to closed.
  • resolution set to fixed.

It should be fixed by updating pyNN to version > r481. Headers should be correct, and therefore the objects in NeuroTools will have the good number of elements.