[pt] Use new POS tagging schema

[susanaboatto]

⚠️ This change requires the portuguese-pos-dict dependency to be v0.13 or later!

tl;dr

This PR adds POS tags to the pt tagset for enclitic pronouns.

What changes

Let's use diz-me ('tell me') to illustrate the changes.

was	will be
three tokens (diz, -, me)	a single token (diz-me)
each token tagged separately	tagged as dizer[VMM02S0:PP1CSO00]

Consequences

• dictionary suggestions consider whole word forms (e.g. ama-se <=> amasse);
• forms that only appear in derived environments are no longer accepted by the dictionary in isolation (e.g. amá and lo from amá-lo should now both be flagged as spelling errors);
• working with mesoclitics is made significantly easier (e.g. fá-lo-á doesn't require five tokens, three of which shouldn't exist in isolation!);
• XML rules no longer require two extra (often optional) tokens for the hyphen and pronouns.

Also in this PR

• tokenisation logic changes from defining breaking characters to defining word characters;
• extensive changes to XML rules in all dialects to account for new tokenisation and tagging schema;
• EnclisisFilter and ProclisisFilter to make synthesising verb forms with pronouns easier.

[p-goulart]

@marcoagpinto I'm pinging you here so you're aware of it.

After this PR goes through, you should be able to go back to editing rules and the dictionary freely. Thank you for your patience!

[marcoagpinto]

@p-goulart

Heya!

Thank you for letting me know.

I missed coding rules so much 😢 😢 😢 😢 😢

@ricardojosehlima
We are soon to be back in business 😛 😛 😛 😛

[p-goulart]

@susanaboatto I hope you've had a chance to have a look. Since it is Friday, I want to wait until Monday to merge this to avoid surprises over the weekend.