Add eslint, prettier, and re-format JS

[QuLogic]

This pulls in the same lint/format method used for ipywidgets, though there are a few differences to minimize changes.

• TypeScript isn't used, so I left out any typescript lint config/plugins.
• Since you've used the spread operation, I've set ECMA version to 9, though it could be set to anything later if you wish.
• Prettier is set to use 4-space tab width for existing files (except package*.json which is already 2-space), or else I've have to rewrite basically everything.

Some other non-defaullt changes:

• I added braces on some if so they would stay separate lines.
• I added a ignore pattern for args that start with _ so eslint wouldn't complain they were unused.

Having a common style is helpful for cross-project sharing, and I plan to apply the same change to Matplotlib's nbagg/webagg backends. However, I am not an expert in the JS ecosystem, so I may not have worked around any issues in the best way.

Adding the config and applying the format are separate commits for easier review.

[QuLogic]

I do not know what is failing here; when I run npm install and npm pack locally, it works fine.

[martinRenou]

Thanks for your PR. If we can fix the CI I'd be happy to merge it.

[QuLogic]

I would if I could, but it seems unrelated. There is no line with ...accumulator in the repo.

[martinRenou]

I will try to find time to look at it, sorry if it takes time

[martinRenou]

Looks like this accumulator issue appears in other PRs so it's not related to yours. Not sure where this is coming from.

[martinRenou]

@QuLogic sorry for the time it takes to merge your PR. Could you rebase your branch? The CI should pass now. It was stupid conda which was installing a very old nodejs version.

[QuLogic]

Rebased. Note, this seems to have bumped the jupyter widgets version in package-lock.json. I'm not sure if that's okay.

[QuLogic]

Actually, never mind the jupyter widgets version bump. I figured out how to get it to not change that.

[martinRenou]

Awesome. Do you think it would make sense to run the lint checker in the CI? So that we make sure we respect the linter rules in the future. I would be in favor of this myself.

[QuLogic]

Yes, this is something we did in Matplotlib as well; I can pull that in here, or separately.