Unquoted Unicode strings cause parsing errors

[geographika]

Currently the grammar includes the following for unquoted strings used in list expressions:

UNQUOTED_STRING: /[a-z0-9_\-:]+/i
UNQUOTED_STRING_SPACE: /[a-z0-9_\-: ]+/i

When accents are used these cannot parse as referred to in #92. For example:

EXPRESSION {Aérodrome,Aéroport,Héliport}

Note only allowing for accents within words within braces is an option. These will always be separated by commas, but the words can contain spaces.

A test (currently failing) test_unquoted_unicode_string has been created to test this.

Using the regex mentioned in lark-parser/lark#208 (WORD: /[^\W\d_]+/) works in Python3 but fails in Python27 with:

UnexpectedCharacters: No terminal defined for '\xe9' at line 3 col 20