[pt] Fixing disambiguator: SPS00 to DA0MS0/DA0FS0

[marcoagpinto]

@jaumeortola

Hello!

For months that I have been struggling with this while developing rules and antipatterns, and I know I should have shared it earlier when you did the major changes to the disambiguator.

What I am facing is that sometimes “o” and “a” get converted to SPS00 which confuses the whole programming since SPS00 has no gender.

Here is an example, check this text in the morphologic dictionary site: o. a.

“a” becomes also “SPS00” which damages tons of rules.

This is a major revelation which can impact on tons of rules if fixed, I am even scared to think about it, but it is something that needs to be fixed.

What is your opinion, Jaume?

Thanks!

[jaumeortola]

I know that these words are ambiguous, and it can be difficult to fully disambiguate them.

• "o" can be determiner and pronoun (but not preposition!)
• "a" can be determiner, pronoun and preposition

Can you be more specific about the changes you need? Provide examples.

[marcoagpinto]

@jaumeortola

Right now, I can't remember of any example.

As I code antipatterns, I will try to find some.

[marcoagpinto]

@jaumeortola

Here is an example I found:

Quero que digas a verdade já!
Quero que digas as verdades já!

In the first sentence, “a” appears as “SPS00” and in the second as “DA0FP0”.