| 36 | |
| 37 | |
| 38 | def get_parser(): |
| 39 | parser = argparse.ArgumentParser( |
| 40 | description='convert raw text to tokenized text', |
| 41 | formatter_class=argparse.ArgumentDefaultsHelpFormatter) |
| 42 | parser.add_argument('--nchar', |
| 43 | '-n', |
| 44 | default=1, |
| 45 | type=int, |
| 46 | help='number of characters to split, i.e., \ |
| 47 | aabb -> a a b b with -n 1 and aa bb with -n 2') |
| 48 | parser.add_argument('--skip-ncols', |
| 49 | '-s', |
| 50 | default=0, |
| 51 | type=int, |
| 52 | help='skip first n columns') |
| 53 | parser.add_argument('--space', |
| 54 | default='<space>', |
| 55 | type=str, |
| 56 | help='space symbol') |
| 57 | parser.add_argument('--bpe-model', |
| 58 | '-m', |
| 59 | default=None, |
| 60 | type=str, |
| 61 | help='bpe model for english part') |
| 62 | parser.add_argument('--non-lang-syms', |
| 63 | '-l', |
| 64 | default=None, |
| 65 | type=str, |
| 66 | help='list of non-linguistic symobles,' |
| 67 | ' e.g., <NOISE> etc.') |
| 68 | parser.add_argument('text', |
| 69 | type=str, |
| 70 | default=False, |
| 71 | nargs='?', |
| 72 | help='input text') |
| 73 | parser.add_argument('--trans_type', |
| 74 | '-t', |
| 75 | type=str, |
| 76 | default="char", |
| 77 | choices=["char", "phn", "cn_char_en_bpe"], |
| 78 | help="""Transcript type. char/phn. e.g., for TIMIT |
| 79 | FADG0_SI1279 - |
| 80 | If trans_type is char, read from |
| 81 | SI1279.WRD file -> "bricks are an alternative" |
| 82 | Else if trans_type is phn, |
| 83 | read from SI1279.PHN file -> |
| 84 | "sil b r ih sil k s aa r er n aa l |
| 85 | sil t er n ih sil t ih v sil" """) |
| 86 | return parser |
| 87 | |
| 88 | |
| 89 | def main(): |