- Add configuration to support defining custom groups
  and ordering
- Set a default grouping that replications the previous
  functionality of grouped-imports

- added comments in an attempt to describe to others how the code
  is working.
- update variable names to be a bit more clear about what is being
  tracked within the code.

This change updates the refactored "group" based orderedImports
code to handle the cases fixed in the following previous PRs:

- palantir#4374: fixed grouping for import-sources-order
- palantir#3733: check all imports of same type are grouped

As part of this we removed the concept of ImportType as this
is now covered by the defined groups for sorting.

Moved the check for ensuring groups are not standing alone
into the code checkBlocksGroup() code where other group checks
are done.

Ex:
```
import './baa';
import './baz';

import './caa';  // this should fail
```

Also had to update several other test cases to include the failure
notice about import groups needing to be together.

- Change the message shown when items are not grouped as expected

The new message is more clear and understandable

- Use a better default group name

If a group has a regex and no name, the default name is now:

/regex/

This makes it more clear in the output that it is a regex and
helps to keep the boundaries of the regex string more clear.

This check is already handled by the check for groups needing to be sorted.

The idea is that once the first block of imports is found that has
an import that is out of order, then all future blocks also have
blocks out of order.  So instead of flagging all blocks we just flag
the first block that is causing the issue and provide a fix that will
fix all blocks.

This makes the case for manually fixing the import ordering much more
clear because people can solve one item at a time.  The automattic
fixing is also more clear because we don't end up with multiple
failures that have the same fix.