Nasty.Language.English.PhraseParser (Nasty v0.3.0)
View SourcePhrase structure parser for English.
Builds syntactic phrases (NounPhrase, VerbPhrase, etc.) from POS-tagged tokens using bottom-up pattern matching.
Approach
- Greedy longest-match: Consume as many tokens as possible for each phrase
- Bottom-up parsing: Build smaller phrases first, then combine
- Left-to-right: Process tokens in order
Grammar Rules (Simplified CFG)
NP → Det? Adj* Noun PP*
VP → Aux* MainVerb NP? PP* Adv*
PP → Prep NP
AdjP → Adv? Adj
AdvP → AdvExamples
iex> tokens = [
...> %Token{text: "the", pos_tag: :det},
...> %Token{text: "cat", pos_tag: :noun}
...> ]
iex> PhraseParser.parse_noun_phrase(tokens, 0)
{:ok, noun_phrase, 2} # Consumed 2 tokensSummary
Functions
Parses an adjectival phrase starting at the given position.
Parses an adverbial phrase (simple adverb for now).
Parses a noun phrase starting at the given position.
Parses a prepositional phrase starting at the given position.
Parses a relative clause starting at the given position.
Parses a verb phrase starting at the given position.
Functions
@spec parse_adjectival_phrase([Nasty.AST.Token.t()], non_neg_integer()) :: {:ok, Nasty.AST.AdjectivalPhrase.t(), non_neg_integer()} | :error
Parses an adjectival phrase starting at the given position.
Grammar: Adv? Adj
Returns {:ok, adj_phrase, next_pos} or :error
@spec parse_adverbial_phrase([Nasty.AST.Token.t()], non_neg_integer()) :: {:ok, Nasty.AST.AdverbialPhrase.t(), non_neg_integer()} | :error
Parses an adverbial phrase (simple adverb for now).
Grammar: Adv
Returns {:ok, adv_phrase, next_pos} or :error
@spec parse_noun_phrase([Nasty.AST.Token.t()], non_neg_integer()) :: {:ok, Nasty.AST.NounPhrase.t(), non_neg_integer()} | :error
Parses a noun phrase starting at the given position.
Grammar: Det? Adj (Noun | PropN | Pron) PP
Pronouns can stand alone as NPs (e.g., "I", "he", "they").
Returns {:ok, noun_phrase, next_pos} or :error
@spec parse_prepositional_phrase([Nasty.AST.Token.t()], non_neg_integer()) :: {:ok, Nasty.AST.PrepositionalPhrase.t(), non_neg_integer()} | :error
Parses a prepositional phrase starting at the given position.
Grammar: Prep NP
Returns {:ok, prep_phrase, next_pos} or :error
@spec parse_relative_clause([Nasty.AST.Token.t()], non_neg_integer()) :: {:ok, Nasty.AST.RelativeClause.t(), non_neg_integer()} | :error
Parses a relative clause starting at the given position.
Grammar: RelPron/RelAdv Clause
Relative pronouns: who, whom, whose, which, that Relative adverbs: where, when, why
Returns {:ok, relative_clause, next_pos} or :error
@spec parse_verb_phrase([Nasty.AST.Token.t()], non_neg_integer()) :: {:ok, Nasty.AST.VerbPhrase.t(), non_neg_integer()} | :error
Parses a verb phrase starting at the given position.
Grammar: Aux MainVerb NP? PP Adv*
Returns {:ok, verb_phrase, next_pos} or :error