verbalize.py 2.3 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657
  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.ru.verbalizers.cardinal import CardinalFst
  4. from fun_text_processing.text_normalization.ru.verbalizers.date import DateFst
  5. from fun_text_processing.text_normalization.ru.verbalizers.decimal import DecimalFst
  6. from fun_text_processing.text_normalization.ru.verbalizers.electronic import ElectronicFst
  7. from fun_text_processing.text_normalization.ru.verbalizers.measure import MeasureFst
  8. from fun_text_processing.text_normalization.ru.verbalizers.money import MoneyFst
  9. from fun_text_processing.text_normalization.ru.verbalizers.ordinal import OrdinalFst
  10. from fun_text_processing.text_normalization.ru.verbalizers.telephone import TelephoneFst
  11. from fun_text_processing.text_normalization.ru.verbalizers.time import TimeFst
  12. class VerbalizeFst(GraphFst):
  13. """
  14. Composes other verbalizer grammars.
  15. For deployment, this grammar will be compiled and exported to OpenFst Finate State Archiv (FAR) File.
  16. More details to deployment at NeMo/tools/text_processing_deployment.
  17. Args:
  18. deterministic: if True will provide a single transduction option,
  19. for False multiple options (used for audio-based normalization)
  20. """
  21. def __init__(self, deterministic: bool = True):
  22. super().__init__(name="verbalize", kind="verbalize", deterministic=deterministic)
  23. cardinal = CardinalFst()
  24. cardinal_graph = cardinal.fst
  25. ordinal_graph = OrdinalFst().fst
  26. decimal = DecimalFst()
  27. decimal_graph = decimal.fst
  28. date = DateFst()
  29. date_graph = date.fst
  30. measure = MeasureFst()
  31. measure_graph = measure.fst
  32. electronic = ElectronicFst()
  33. electronic_graph = electronic.fst
  34. whitelist_graph = WhiteListFst().fst
  35. money_graph = MoneyFst().fst
  36. telephone_graph = TelephoneFst().fst
  37. time_graph = TimeFst().fst
  38. graph = (
  39. measure_graph
  40. | cardinal_graph
  41. | decimal_graph
  42. | ordinal_graph
  43. | date_graph
  44. | electronic_graph
  45. | money_graph
  46. | whitelist_graph
  47. | telephone_graph
  48. | time_graph
  49. )
  50. self.fst = graph