def test_prob_lexicon():
check = {
'Article': [('the', 0.5), ('a', 0.25), ('an', 0.25)],
'Pronoun': [('i', 0.4), ('you', 0.3), ('he', 0.3)]
}
lexicon = ProbLexicon(Article="the [0.5] | a [0.25] | an [0.25]",
Pronoun="i [0.4] | you [0.3] | he [0.3]")
assert lexicon == check
def test_prob_generation():
lexicon = ProbLexicon(Verb="am [0.5] | are [0.25] | is [0.25]",
Pronoun="i [0.4] | you [0.3] | he [0.3]")
rules = ProbRules(
S="Verb [0.5] | More [0.3] | Pronoun [0.1] | nobody is here [0.1]",
More="Pronoun Verb [0.7] | Pronoun Pronoun [0.3]")
grammar = ProbGrammar("Simplegram", rules, lexicon)
sentence = grammar.generate_random('S')
assert len(sentence) == 2
def test_prob_grammar():
rules = ProbRules(A="B C [0.3] | D E [0.7]", B="E [0.1] | a [0.2] | b c [0.7]")
lexicon = ProbLexicon(Article="the [0.5] | a [0.25] | an [0.25]",
Pronoun="i [0.4] | you [0.3] | he [0.3]")
grammar = ProbGrammar("Simplegram", rules, lexicon)
assert grammar.rewrites_for('A') == [(['B', 'C'], 0.3), (['D', 'E'], 0.7)]
assert grammar.isa('the', 'Article')
grammar = nlp.E_Prob_Chomsky
for rule in grammar.cnf_rules():
assert len(rule) == 4