Extract bifurcation angles from an L5TTPC1 population about neurom HOT 9 CLOSED

Did you manage to reproduce the problem with a single file? If so, could you send it? If not, could you put the data somewhere I can access it (for example, nfs?)

from neurom.

Hi, this is exactly the problem :) I just put the first cell on a separate directory and run it but everything works fine for that! It's only for the whole population of 100 cells that TreeType.all does not work. TreeType.basal_dendrite works for the population. I'll now try small populations and let you know. Nancy

from neurom.

Now, I think I localised the cell that causes the problems: It's the 37th one. if you run the script:

from neurom.ezy import load_neurons
from neurom.core.types import TreeType

def extract_feature(data_dir, feature, neurite_type=TreeType.all):
    population = load_neurons(data_dir)
    feature_data = [getattr(n, 'get_' + feature)(neurite_type=neurite_type) for n in population]
    return feature_data

with this single cell, feature = 'local_bifurcation_angles' it gives the math domain error.
L5TTPC1_37th.zip

I hoped it uploaded the file correctly, otherwise I can send it per mail. I try to learn to use those systems :)

from neurom.

Sorry, I cannot reproduce the problem. Here is a simple script that attempts to reproduce what I think you did. For the record, this is the kind of code you should submit. A script that can be simply run to reproduce the reported issue.

from neurom.ezy import load_neurons
from neurom.core.types import TreeType

def extract_feature(data_dir, feature, neurite_type=TreeType.all):
    population = load_neurons(data_dir)
    feature_data = [getattr(n, 'get_' + feature)(neurite_type=neurite_type)
                    for n in population]
    return feature_data

if __name__ == '__main__':
    # data is ./L5TTPC1_37th/L5TTPC1_nancy_ID36.h5
    # from L5TTPC1_37th.zip
    ddir = 'L5TTPC1_37th'
    feature = 'local_bifurcation_angles'
    f = extract_feature(ddir, feature, TreeType.all)
    print f

from neurom.

Hi Juan, can you do me a favour? I think I sent you the wrong cell. It's the cell with ID35 that causes me the problems and not with ID36. I'm attaching the script test.py and the ID35 cell below. Could you run it in your environment when you find a minute? Thanks very much and sorry for the mess :) nancy
cell35_script.zip

from neurom.

I can reproduce the problem with the latest morphology. I will have a look at it. FYI, this is a small example that reproduces the problem:

from neurom.ezy import load_neuron, TreeType


if __name__ == '__main__':

    data_file = './cell35_script/L5TTPC1_36th/L5TTPC1_nancy_ID35.h5'
    nrn = load_neuron(data_file)
    f = nrn.get_local_bifurcation_angles(neurite_type=TreeType.apical_dendrite)

from neurom.

I found the problem: morphmath.angle_3points is not robust against collinear vectors. The apical dendrite in your problem morphology has a bifurcation with angle pi radians.

I changed the implementation for a more robust one (after adding obvious tests that should have been there in the first place). See pull request #195.

from neurom.

thanks very much Juan! I think there are no embarrassing bugs, embarrassing was to send you the wrong cell and let you search in the wrong place :)

nancy
On 06 Jan 2016, at 10:04, Juan Palacios <[email protected]mailto:[email protected]> wrote:

I found the problem: morphmath.angle_3points is not robust against collinear vectors. I changes the implementation for a more robust one (after adding obvious tests that should have been there in the first place). See pull request #195#195.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/194#issuecomment-169271249.

from neurom.

Fixed with #195.

from neurom.