Automatic file and/or project vocabulary

**What problem does this solve?**
Often in programming in comments, it would be really useful to have spell checking enabled - but because of all the specific terminology in the file (eg "BlogPost", "CommentCount", "color_mapper", etc) it has to be disabled to prevent thousands of false positives.

**Proposed Solution**
There could be two options "FileVocabulary" and possibly a "ProjectVocabulary" which automatically add variable names / keywords / etc (anything *outside* of the checked comments, essentially) in to the current vocab while checking the file.

I don't know whether the ProjectVocabulary is possible, or easily possible, but I think the file based one would help a lot.

**Examples**
```python

from pages.models import BasePage

class MagicalPage(BasePage):
"""
MagicalPage takes a lot of details from BasePage and includes all the
extra magic that makes it so special.
"""
```
Should not raise any spelling validation issues - as MagicalPage & BasePage are mentioned outside of the comment text.

But then

```python
# update MagcialPage to have more data
```

would have a spelling issue, as there is a typo there.

No implementation files, tests, or entry points are named. Start by tracing Harper’s current comment spell-checking and vocabulary handling, then use the Python examples as acceptance cases: identifiers outside comments are accepted, while misspellings inside comments are still reported.

Written by the indexing model from the issue text.