tokenization issues

[arademaker]

after #9

1. we still have cases were names 'A.B.Fulano' in a single token
2. we may have other tokens that need to be split. We can search for . or - inside token forms.
3. we have some cases of WF tokens with sep=`, the space is the default sep, need to remove those cases and check if the detokenization approach still works matching thetext` field.

[arademaker]

this issue was closed with the above commits.