search

Phylliida / openmuscle

An open source implementation of the Hill-type muscle model without any dependencies.
MIT License
8 stars 3 forks source link

Upload to conda or pypi FTW! #1

Open extraymond opened 7 years ago

extraymond commented 7 years ago

Hi! Thanks for building this, it's really nice. It would be a lot easier to distribute if it's on conda/pypi or other distribution services.

Phylliida commented 7 years ago

Umm I admit I actually have no idea how to do that but that would be nice. Feel free to add it yourself or make a fork and do it there if you want.

extraymond commented 7 years ago

Thanks for your reply! It seems simple on the official manual, however I haven't tried it myself XD

I think your code is clear enough to use, so a fork is not necessary IMO.

Phylliida commented 7 years ago

I'm curious, this is unrelated but what are you doing with this? I wrote this for biologically based animation as a hobby thing and it should work but I ended up moving to other projects because the other components of biologically based animation (simulation, physics, high performance, nice GUI, etc.) are hard XD

extraymond commented 7 years ago

Hi! I'm currently working on sports biomechanics simulation and optimizations. The same reason you listed make your project awesome in my opinion.

For rigid body dynamics or simulation I recommend you take a look at pydy, it's super awesome to setup a system using kane's method. And it gives a clear interface for controlling the torque activity using custom defined functions. It even have a nice visualization class.

The very hard part of a good simulation is to assess whether the results is human muscle dynamics. I've seen opensim for that matter, but python is more friendly for me XD.

Phylliida commented 7 years ago

Oh cool :) Yea same, opensim is gross, that's why I made this :) I did want to clarify that while I'm fairly certain this implementation is correct and have done some basic testing, I haven't done any formal comparisons to other implementations of Hill-type muscle models. This was ported and adapted from the paper and its' code given as a link at the top of openmuscle.py that is about how to implement a modern muscle model but I just did this as a hobby. So if it is not correct it is very close to being correct, but please do some more testing before using in a formal setting. Feel free to share if you do :)