verbalize.py 2.9 KB

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162
  1. from fun_text_processing.text_normalization.en.graph_utils import GraphFst
  2. from fun_text_processing.text_normalization.en.verbalizers.whitelist import WhiteListFst
  3. from fun_text_processing.text_normalization.es.verbalizers.cardinal import CardinalFst
  4. from fun_text_processing.text_normalization.es.verbalizers.date import DateFst
  5. from fun_text_processing.text_normalization.es.verbalizers.decimals import DecimalFst
  6. from fun_text_processing.text_normalization.es.verbalizers.electronic import ElectronicFst
  7. from fun_text_processing.text_normalization.es.verbalizers.fraction import FractionFst
  8. from fun_text_processing.text_normalization.es.verbalizers.measure import MeasureFst
  9. from fun_text_processing.text_normalization.es.verbalizers.money import MoneyFst
  10. from fun_text_processing.text_normalization.es.verbalizers.ordinal import OrdinalFst
  11. from fun_text_processing.text_normalization.es.verbalizers.telephone import TelephoneFst
  12. from fun_text_processing.text_normalization.es.verbalizers.time import TimeFst
  13. class VerbalizeFst(GraphFst):
  14. """
  15. Composes other verbalizer grammars.
  16. For deployment, this grammar will be compiled and exported to OpenFst Finate State Archiv (FAR) File.
  17. More details to deployment at NeMo/tools/text_processing_deployment.
  18. Args:
  19. deterministic: if True will provide a single transduction option,
  20. for False multiple options (used for audio-based normalization)
  21. """
  22. def __init__(self, deterministic: bool = True):
  23. super().__init__(name="verbalize", kind="verbalize", deterministic=deterministic)
  24. cardinal = CardinalFst(deterministic=deterministic)
  25. cardinal_graph = cardinal.fst
  26. ordinal = OrdinalFst(deterministic=deterministic)
  27. ordinal_graph = ordinal.fst
  28. decimal = DecimalFst(deterministic=deterministic)
  29. decimal_graph = decimal.fst
  30. fraction = FractionFst(deterministic=deterministic)
  31. fraction_graph = fraction.fst
  32. date = DateFst(deterministic=deterministic)
  33. date_graph = date.fst
  34. measure = MeasureFst(cardinal=cardinal, decimal=decimal, fraction=fraction, deterministic=deterministic)
  35. measure_graph = measure.fst
  36. electronic = ElectronicFst(deterministic=deterministic)
  37. electronic_graph = electronic.fst
  38. whitelist_graph = WhiteListFst(deterministic=deterministic).fst
  39. money_graph = MoneyFst(decimal=decimal, deterministic=deterministic).fst
  40. telephone_graph = TelephoneFst(deterministic=deterministic).fst
  41. time_graph = TimeFst(deterministic=deterministic).fst
  42. graph = (
  43. cardinal_graph
  44. | measure_graph
  45. | decimal_graph
  46. | ordinal_graph
  47. | date_graph
  48. | electronic_graph
  49. | money_graph
  50. | fraction_graph
  51. | whitelist_graph
  52. | telephone_graph
  53. | time_graph
  54. )
  55. self.fst = graph