Prepare extracellular potential demo for LFPy (Python)

[halfflat]

Arbor may be able to be well utilized as a back-end for LFPy. As a first step, we will make a simple demo in Python that computes the potential at a fixed set of electrode sites from a single morphologically detailed neuron.

Model NEURON python code: pure_neuron_example.py.txt

Once #1034 is merged, we'll need to:

• expose trans membrane current probes and sample data in Python interface, together with metadata;
• use the single-cell Python API to model and run a neuron such as in the example code;
• and in anticipation of running multiple neurons, we will need to implement exact sample timing.

[halfflat]

We may also need to enrich the morphology interface (in Python and/or C++) for facilitating geometry queries.

[halfflat]

Exact sampling provided by #1046

[halfflat]

I will split this into two tasks: C++ version and a Python version. The latter will require a Python interface for the cell-wide trans membrane current probes.

[halfflat]

The other piece of Python infrastructure required to produce a Python LFP demo with LFPykit is #1225, which is pending review.

[halfflat]

Last remaining component is #1262 for correct transmembrane accounting. Turns out #1225 wasn't the last prerequisite after all!
Task then will be to convert existing LFP demo to an LFPkit-based implementation.

[espenhgn]

I've made a basic LFPykit + .swc file example available here: https://github.com/LFPy/LFPykit/blob/master/examples/Example_Arbor_swc.ipynb (via LFPy/LFPykit#129).

It's not particularly elegant though.

[halfflat]

I've made a basic LFPykit + .swc file example available here: https://github.com/LFPy/LFPykit/blob/master/examples/Example_Arbor_swc.ipynb (via LFPy/LFPykit#129).

Thanks, I'll take a look!

[brenthuisman]

@espenhgn Would you mind if we include the example in the notebook in the Arbor-repo? Probably we would make some small modifications.

[espenhgn]

@espenhgn Would you mind if we include the example in the notebook in the Arbor-repo? Probably we would make some small modifications.

No objections :)
I'll probably seize the opportunity to patch any improvements into the original example notebook then.

[brenthuisman]

OK, thanks!

[espenhgn]

Shall I add the example in a PR to Arbor perhaps?

[brenthuisman]

By all means :) Although we've though of having notebooks in the repo, so far we have stuck to good old Python scripts, which also also straightforward to hook up to our testing. (All examples are run and checked for failure, which is not the most comprehensive testing procedure but it's a start.) So, if you want to post the notebook, we'd need to add something to test notebooks (nb-convert I think) to our pipeline. I can also take care of the scriptification in the PR if that's too much fiddling for you.

[espenhgn]

I can of course just export the existing notebook as a regular Python file as is (and clean it up a bit) and add that to the PR for now.

[brenthuisman]

Let's do that, we can think about how to integrate notebooks in our testing at a later point (I assume you'll keep this example as a notebook around)

[espenhgn]

Yes, the existing LFPykit examples will remain notebooks. I like Github's ability to render them.

[espenhgn]

Ok, I've open the PR. I did not check it with the current master as I encountered some issue with the cmake configuration.