Changed clearmetadata preprocessor to remove all metadata by default

[MSeal]

I decided to just go ahead and implement the requested change for #637. Technically it changes the default behavior to clear notebook and cell metadata, but I think that's what's usually intended on first-use and I added a flag to fall back to the old behavior if need-be. I also added a flag for preserving specific keys by name.

[MSeal]

Gonna merge this in another day or two if no one any comments fyi.

[t-makaro]

Looks mostly good to me. And it makes sense. I think this will be a nice change for people that specifically want to not have things like "version":3.6.5" committed to git. It also makes sense to get rid of the kernelspec (since 2 different machines will run the notebook with different kernels). But I have a couple discussion points:

1. Should we maybe include the "language_info" in the default metadata keys to preserve? So that we don't need to guess the language of the notebook: Add --clean option to --to notebook to strip all metadata and output? #637 (comment)
   A user could always override it later, and it doesn't look like anything in that field should change frequently.

 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 }

2. Should we give users the option to clear notebook metadata but not cell metadata?

[MSeal]

Thanks for the review. I think I can incorporate those suggested changes fairly easily

[t-makaro]

I just realised that "language_info" is more nested than I thought. It's specifically, the name subfield of language_info that I want to keep which I guess makes it a bit harder. The rest of it can be tossed.

[MSeal]

@SylvainCorlay or @t-makaro If you wanted to quickly review this for the release tomorrow I believe it's addressed the PR comments.

[SylvainCorlay]

I am not too opinionated about this.

Leaving this to you judgement @t-makaro @jupyter/everybody
@maartenbreddels let me know if you want to chime in on this before the release!

[MSeal]

I'm going to merge for now since I think the comments are well addressed and tested to avoid blocking release stuff

[MSeal]

If we need to adjust later I can make tweaks