Data Explorer: Visually display leading/trailing whitespace and empty strings in data values

[wesm]

Addresses #3067. Currently in the data explorer, the strings "", " ", and " " cannot be visually distinguished because of the absence of quotations.

There are a few approaches we could take:

1. Format with quotations (I believe RStudio does this)
2. Provide a special display for empty strings (this is what's described as Data Explorer: Display empty string vs a whitespace  #3067)
3. Replace whitespace with a unicode character to indicate the presence of leading/trailing whitespace characters

This PR does the latter two, using a mid-dot as the unicode representation and the special string <empty> for empty strings (will need to be localized).

[wesm]

Some ideas:

• Use a different unicode character than the one selected here
• Only insert "whitespace characters" for leading or trailing whitespace (some of the smoke tests are failing because of this)
• Render these whitespace characters with < 100% opacity (like the special values)

[wesm]

Here's another try:

• Use middle dot as the whitespace character
• Only replace leading and trailing whitespaces
• Set opacity < 100%

code for creating data is:

df = pd.DataFrame({'a': ["", "", " ", "  wat", "   wh  at  "]})

A part of me thinks that using quotations might be better, but this seems reasonable as well

[wesm]

I added a localization hook for the <empty> string representation. Will merge this once the build is green since this seems like a reasonable improvement over the status quo, and we can iterate on it based on user feedback