Refactor urdf parser function

[quarkytale]

Making init_collada and init_urdf consistent with new changes in init_sdf.
Changes:

• refactored joint tag loop
• updated limit checks

[clalancette]

I think that this should move to right before the if name in dependent_joints check. That way we avoid a situation where we could add it to self.joint_list, but then fail to add it to the free_joints because of a missing limit.

[clalancette]

This is again changing the semantics slightly. I think here we should probably do something more like:

Suggested change

	if limit_list:
	if not limit_list:
	continue

[quarkytale]

Done

[clalancette]

The more I look at it, the more I want to clean up init_collada here to be more robust. In particular, it is assuming a lot of things about the passed-in COLLADA file, and some of those things just may not be true. I've been looking in particular at https://www.khronos.org/files/collada_spec_1_5.pdf .

So what I'm going to suggest here is that we make this PR just about refactoring init_urdf, which is an easier problem. Once we have that in, we can do another followup PR where we refactor init_collada so it is much more robust. @quarkytale does that make sense?

[quarkytale]

Alright this should just have init_urdf changes now, will open up a separate one for init_collada once this and init_sdf is in!

[clalancette]

A few more things to fix, this is getting closer.

[clalancette]

One more small change, then this will be good to go.

[clalancette]

This looks great to me. Thank you for this cleanup!