fix lexer parsing ID 'contains' as comparison

[ggmichaelgo]

With the new Lexer parser, we are always parsing contains as a comparison when it could also be an id.

Example:

Lexer.new("a.contains.b").tokenize

Expected output:

[[:id, "a"], [:dot, "."], [:id, "contains"], [:dot, "."], [:id, "b"], [:end_of_string]]

Current output

[[:id, "a"], [:dot, "."], [:comparison, "contains"], [:dot, "."], [:id, "b"], [:end_of_string]]