phrase_detective package

Subpackages

• phrase_detective.regx package
  • Submodules
  • phrase_detective.regx.de module
  • phrase_detective.regx.en module
  • phrase_detective.regx.es module
  • Module contents

Submodules

phrase_detective.constants module

Constants used in package

phrase_detective.constants.PKG_INDICES

Key is language abbreviation, such as en, es. Value is the name of pre-trained package

Type:

dict

phrase_detective.lib module

phrase_detective.lib.merge(ranges)

Merge a list of tuples, which have two elements respectively

Parameters:

ranges (list) – [(start1, end1), (start2, end2), …]

Returns:

merged tuples

Return type:

list

Examples

>>> merge([(1, 4),(3, 6), (4, 7), (9, 12)])
[(1, 7), (9, 12)]

phrase_detective.lib.merge_range(first, second)

Merge two tuples, which have two elements respectively

Parameters:

• first (tuple) – (start, end)

• second (tuple) – (start, end)

Returns:

merged tuples

Return type:

list

Examples

>>> merge_range((1, 3), (4, 8))
[(1, 3), (4, 8)]

>>> merge_range((1, 3), (3, 8))
[(1, 3), (3, 8)]

>>> merge_range((1, 3), (2, 8))
[(1, 8)]

phrase_detective.noun_phrases module

class phrase_detective.noun_phrases.NounPhraseRecognizer(nlp: Language)

Bases: object

Customerized component to detect noun phrases in spacy.tokens.Doc object. The NP values are stored in doc._.noun_phrases,

ext_name

customized extension field name

Type:

str

phrase_detective.verb_knowledge module

class phrase_detective.verb_knowledge.VerbKnowledgeRecognizer(nlp: Language)

Bases: object

Customerized component to detect verb knowledge in spacy.tokens.Doc object. The corresponding values are stored in doc._.verbs.

ext_name

customized extension field name

Type:

str

matcher

Rule maker for detecting VERB

Type:

spacy.mathcer.Matcher

Module contents

phrase_detective Package

Components for spacy pipeline, to detect noun phrase, prep phrase, verb related knowledge

Install package

$ pip3 install --verbose phrase_detective
$ python -m spacy download en_core_web_trf
$ python -m spacy download es_dep_news_trf
$ python -m spacy download de_dep_news_trf

Example: Detect noun phrases

import spacy
from spacy import Language
from phrase_detective import NounPhraseRecognizer, PKG_INDICES

@Language.factory("nprecog")
def create_np_parser(nlp: Language, name: str):
  return NounPhraseRecognizer(nlp)

def noun_phrase(lang, sentence):
  nlp = spacy.load(PKG_INDICES[lang])
  nlp.add_pipe("nprecog")
  doc = nlp(sentence)
  for np in doc._.noun_phrases:
    print(np.text)

Example: Detect verb phrases

import spacy
from spacy import Language
from phrase_detective import VerbKnowledgeRecognizer, PKG_INDICES

@Language.factory("vkbrecog")
def create_vkb_parser(nlp: Language, name: str):
  return VerbKnowledgeRecognizer(nlp)

def verb_knowledge(lang, sentence):
  nlp = spacy.load(PKG_INDICES[lang])
  nlp.add_pipe("vkbrecog")
  doc = nlp(sentence)
  for v in doc._.verbs:
    print("TEXT: {}, TAG: {}, FORM: {}, ORIGNAL: {}".format(v.text, v.tag_, spacy.explain(v.tag_), v.lemma_))