funasr tables

游雁

2024-10-16 6e6475cd2afebd5db41beef633645f154bb4cf05

 fun_text_processing/text_normalization/export_models.py

@@ -1,30 +1,25 @@
# Copyright NeMo (https://github.com/NVIDIA/NeMo). All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

import os
from time import perf_counter
from argparse import ArgumentParser
from fun_text_processing.text_normalization.en.graph_utils import generator_main


def parse_args():
    parser = ArgumentParser()
    

    parser.add_argument(
        "--language", help="language", choices=['de', 'en', 'es', 'ru', 'zh'], default="en", type=str
        "--language",
        help="language",
        choices=["de", "en", "es", "ru", "zh"],
        default="en",
        type=str,
    )
    parser.add_argument(
        "--input_case", help="input capitalization", choices=["lower_cased", "cased"], default="cased", type=str
        "--input_case",
        help="input capitalization",
        choices=["lower_cased", "cased"],
        default="cased",
        type=str,
    )
    parser.add_argument(
        "--export_dir",
@@ -34,27 +29,53 @@
    )
    return parser.parse_args()

def get_grammars(lang: str="en", input_case: str="cased"):
  if lang=='de':
    from fun_text_processing.text_normalization.de.taggers.tokenize_and_classify import ClassifyFst
    from fun_text_processing.text_normalization.de.verbalizers.verbalize_final import VerbalizeFinalFst
  elif lang=='en':
    from fun_text_processing.text_normalization.en.taggers.tokenize_and_classify import ClassifyFst
    from fun_text_processing.text_normalization.en.verbalizers.verbalize_final import VerbalizeFinalFst
  elif lang=='es':
    from fun_text_processing.text_normalization.es.taggers.tokenize_and_classify import ClassifyFst
    from fun_text_processing.text_normalization.es.verbalizers.verbalize_final import VerbalizeFinalFst
  elif lang=='ru':
    from fun_text_processing.text_normalization.ru.taggers.tokenize_and_classify import ClassifyFst
    from fun_text_processing.text_normalization.ru.verbalizers.verbalize_final import VerbalizeFinalFst
  elif lang=='zh':
    from fun_text_processing.text_normalization.zh.taggers.tokenize_and_classify import ClassifyFst
    from fun_text_processing.text_normalization.zh.verbalizers.verbalize_final import VerbalizeFinalFst
  else:
    from fun_text_processing.text_normalization.en.taggers.tokenize_and_classify import ClassifyFst
    from fun_text_processing.text_normalization.en.verbalizers.verbalize_final import VerbalizeFinalFst

  return ClassifyFst(input_case=input_case).fst, VerbalizeFinalFst().fst
def get_grammars(lang: str = "en", input_case: str = "cased"):
    if lang == "de":
        from fun_text_processing.text_normalization.de.taggers.tokenize_and_classify import (
            ClassifyFst,
        )
        from fun_text_processing.text_normalization.de.verbalizers.verbalize_final import (
            VerbalizeFinalFst,
        )
    elif lang == "en":
        from fun_text_processing.text_normalization.en.taggers.tokenize_and_classify import (
            ClassifyFst,
        )
        from fun_text_processing.text_normalization.en.verbalizers.verbalize_final import (
            VerbalizeFinalFst,
        )
    elif lang == "es":
        from fun_text_processing.text_normalization.es.taggers.tokenize_and_classify import (
            ClassifyFst,
        )
        from fun_text_processing.text_normalization.es.verbalizers.verbalize_final import (
            VerbalizeFinalFst,
        )
    elif lang == "ru":
        from fun_text_processing.text_normalization.ru.taggers.tokenize_and_classify import (
            ClassifyFst,
        )
        from fun_text_processing.text_normalization.ru.verbalizers.verbalize_final import (
            VerbalizeFinalFst,
        )
    elif lang == "zh":
        from fun_text_processing.text_normalization.zh.taggers.tokenize_and_classify import (
            ClassifyFst,
        )
        from fun_text_processing.text_normalization.zh.verbalizers.verbalize_final import (
            VerbalizeFinalFst,
        )
    else:
        from fun_text_processing.text_normalization.en.taggers.tokenize_and_classify import (
            ClassifyFst,
        )
        from fun_text_processing.text_normalization.en.verbalizers.verbalize_final import (
            VerbalizeFinalFst,
        )

    return ClassifyFst(input_case=input_case).fst, VerbalizeFinalFst().fst


if __name__ == "__main__":
    args = parse_args()
@@ -68,5 +89,4 @@
    tagger_fst, verbalizer_fst = get_grammars(args.language, args.input_case)
    generator_main(tagger_far_file, {"tokenize_and_classify": tagger_fst})
    generator_main(verbalizer_far_file, {"verbalize": verbalizer_fst})
    print(f'Time to generate graph: {round(perf_counter() - start_time, 2)} sec')

    print(f"Time to generate graph: {round(perf_counter() - start_time, 2)} sec")