bug(unicode): utf-8 encoded strings throw errors

[AmitMY]

Given the following hyp:

• uruguay where the leader is raúl fernando sendic rodríguez alfredo zitarrosa died in montevideo montevideo , uruguay montevideo where the leader is daniel martínez ( politician ) uruguay and the language spoken is spanish .

And following refs:

• Alfredo Zitarrosa died in Montevideo, Uruguay. Daniel Martinez is a political leader in Montevideo, and Raul Fernando Sendic Rodriguez is a leader in Uruguay, where Spanish is spoken.
• Alfredo Zitarrosa died in Montevideo, the leader of which, is Daniel Martinez. Montevideo is in Uruguay, where Spanish is the language and where Raúl Fernando Sendic Rodríguez is the leader.
• Raúl Fernando Sendic Rodríguez is the leader of Spanish speaking, Uruguay. Daniel Martinez is the leader of Montevideo, the final resting place of Alfredo Zitarrosa.

corpus check fails.

Meh solution: Map to a string, ignore unicode characters shouldn't hurt too much, but can

def sentence_unicode(s):
    return str(''.join([i if ord(i) < 128 else 'X' for i in s]))

[juharris]

We now support Python 3 and Unicode strings should work. We have a test to check.

[AmitMY]

That's awesome! Thanks