### Related
#11416
I requested the notebooks get moved so our automated examples test would
make sure the notebooks aren't broken. Unfortunately we don't test any
of our notebooks so most were broken.

A little bit related to broader notebook testing improvement
https://linear.app/rerun/issue/RR-1880/test-our-jupyter-notebooks-on-ci

### What
I highly recommend reviewing with rich diff for notebooks enabled.
* Adds nbqa so we can run mypy on the notebooks to at least ensure the
apis align
* I couldn't get the config working so unfortunately I just have the
directories to search listed in the pixi command
* Adds nbstripout but doesn't yet tie it to our py-fmt. Jupyter notebook
outputs are just binary so that will bloat the repo and complicate
diffs. Unfortunately nbstripout doesn't have a nice config setup so I
just put together a simple wrapper.
* I made two actual code changes not just in tests since it made sense
to resolve notebook mypy errors
1. I made component_type a class method. We basically have component
type raw strings everywhere probably because it wasn't easy before to
grab them from the class definition rather than instance
2. If we have a recording string the api should actually be a little
different than asking globally. I assume we'll always find the id so we
universally return a string.
   
This still doesn't execute our notebooks which is probably a worthwhile
followup issue.

### Related
#11416
I requested the notebooks get moved so our automated examples test would
make sure the notebooks aren't broken. Unfortunately we don't test any
of our notebooks so most were broken.

A little bit related to broader notebook testing improvement
https://linear.app/rerun/issue/RR-1880/test-our-jupyter-notebooks-on-ci

### What
I highly recommend reviewing with rich diff for notebooks enabled.
* Adds nbqa so we can run mypy on the notebooks to at least ensure the
apis align
* I couldn't get the config working so unfortunately I just have the
directories to search listed in the pixi command
* Adds nbstripout but doesn't yet tie it to our py-fmt. Jupyter notebook
outputs are just binary so that will bloat the repo and complicate
diffs. Unfortunately nbstripout doesn't have a nice config setup so I
just put together a simple wrapper.
* I made two actual code changes not just in tests since it made sense
to resolve notebook mypy errors
1. I made component_type a class method. We basically have component
type raw strings everywhere probably because it wasn't easy before to
grab them from the class definition rather than instance
2. If we have a recording string the api should actually be a little
different than asking globally. I assume we'll always find the id so we
universally return a string.
   
This still doesn't execute our notebooks which is probably a worthwhile
followup issue.