Support "Cell Body" for ASC files

[genric]

After discussing this today, we probably want to support "Cell Body" definition format as, due to historical reasons, there are morphologies using that soma definition format and other tools are able to process them.
Example morphology:
http://neuromorpho.org/dableFiles/jonas/Source-Version/02022018_252Lgc11.ASC
available here: http://neuromorpho.org/neuron_info.jsp?neuron_name=02022018_252Lgc11)

Looks like there was a similar issue(#264) related to "Cell Body" and there was a PR to support them.

[asanin-epfl]

How should "Cell Body" be interpreted? As soma? In the mentioned 02022018_252Lgc11.ASC there are 4 independent "Cell Body" entries.

[mgeplf]

As you suspect, it's the soma. I believe that at some point the Neurolucida ASC didn't contain an explicit soma, and the "Cell Body" label was used instead - it's very strange.

[asanin-epfl]

How to interpret 4 independent "Cell Body" entries? How soma is composed in that case?

[mgeplf]

Great question; IMO, we have two choices:

1. Raise an error
2. Pick the first one

In the cases we can, we should feed back the info, so there is only one, I would say.

[asanin-epfl]

In this case this issue must be addressed in MorphIO, as NeuroM does not have access to raw content of ASC file. NeuroM has only instance of MorphIO which just skips all "Cell Body" entries.

[mgeplf]

Yeah, agreed; transferring to MorphIO.

[asanin-epfl]

There is an attempt to implement it already in #39.