class PopulationOfRegex(RegexTemplate):
"""
Regex for questions about the population of a country.
Ex: "What is the population of China?"
"How many people live in China?"
"""
openings = (Pos("WP") + Token("is") + Pos("DT") +
Lemma("population") + Pos("IN")) | \
(Pos("WRB") + Lemma("many") + Lemma("people") +
Token("live") + Pos("IN"))
regex = openings + Question(Pos("DT")) + Country() + Question(Pos("."))
def semantics(self, match):
population = PopulationOf(match.country)
return population, "literal"
class LanguageOfRegex(RegexTemplate):
"""
Regex for questions about the language spoken in a country.
Ex: "What is the language of Argentina?"
"what language is spoken in Argentina?"
"""
openings = (Lemma("what") + Token("is") + Pos("DT") +
Question(Lemma("official")) + Lemma("language")) | \
(Lemma("what") + Lemma("language") + Token("is") +
Lemma("speak"))
regex = openings + Pos("IN") + Question(Pos("DT")) + Country() + \
Question(Pos("."))
def semantics(self, match):
language = LanguageOf(match.country)
return language, "enum"
class WhatTimeIs(RegexTemplate):
"""
Regex for questions about the time
Ex: "What time is it in Cordoba"
"""
nouns = Plus(Pos("NN") | Pos("NNS") | Pos("NNP") | Pos("NNPS"))
place = Group(nouns, "place")
openings = (Lemma("what") +
((Token("is") + Token("the") + Question(Lemma("current")) +
Question(Lemma("local")) + Lemma("time")) |
(Lemma("time") + Token("is") + Token("it")))) | \
Lemma("time")
regex = openings + Pos("IN") + place + Question(Pos("."))
def semantics(self, match):
place = HasKeyword(match.place.lemmas.title()) + IsPlace()
utc_offset = UTCof(place)
return utc_offset, "time"
class CapitalOfRegex(RegexTemplate):
"""
Regex for questions about the capital of a country.
Ex: "What is the capital of Bolivia?"
"""
opening = Lemma("what") + Token("is")
regex = opening + Pos("DT") + Lemma("capital") + Pos("IN") + \
Question(Pos("DT")) + Country() + Question(Pos("."))
def semantics(self, match):
capital = CapitalOf(match.country)
label = LabelOf(capital)
return label, "enum"
class PresidentOfRegex(RegexTemplate):
"""
Regex for questions about the president of a country.
Ex: "Who is the president of Argentina?"
"""
regex = Pos("WP") + Token("is") + Question(Pos("DT")) + \
Lemma("president") + Pos("IN") + Country() + Question(Pos("."))
def semantics(self, match):
president = PresidentOf(match.country)
incumbent = IncumbentOf(president)
label = LabelOf(incumbent)
return label, "enum"
class UserData(RegexTemplate):
weight = 1.0
regex = Token("user") + Token("data")
def semantics(self, match):
return HasKeyword(match.words[0].token), "<user data>"