feat: add TrainNumber rule for Chinese ITN

itn/chinese/data/train_number/prefix.tsv

@@ -0,0 +1,7 @@
+高	G
+动	D
+特	T
+快	K
+直	Z
+慢	L
+临	L

itn/chinese/inverse_normalizer.py

@@ -18,6 +18,7 @@
 from itn.chinese.rules.cardinal import Cardinal
 from itn.chinese.rules.char import Char
 from itn.chinese.rules.date import Date
+from itn.chinese.rules.train_number import TrainNumber
 from itn.chinese.rules.fraction import Fraction
 from itn.chinese.rules.license_plate import LicensePlate
 from itn.chinese.rules.math import Math
@@ -62,6 +63,7 @@ def build_tagger_and_verbalizer(self):
             | add_weight(cardinal.tagger, 1.06)
             | add_weight(Math(cardinal=cardinal).tagger, 1.10)
             | add_weight(LicensePlate().tagger, 1.0)
+            | add_weight(TrainNumber().tagger, 1.0)
             | add_weight(Char().tagger, 100)
         ).optimize()
 
@@ -73,6 +75,7 @@ def build_tagger_and_verbalizer(self):
             | Char().verbalizer
             | Date().verbalizer
             | Fraction().verbalizer
+            | TrainNumber().verbalizer
             | Math().verbalizer
             | Measure(enable_0_to_9=self.enable_0_to_9).verbalizer
             | Money(enable_0_to_9=self.enable_0_to_9).verbalizer

itn/chinese/rules/train_number.py

@@ -0,0 +1,36 @@
+# Copyright (c) 2024 Zhendong Peng (pzd17@tsinghua.org.cn)
+#
+# 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.
+
+from pynini import string_file
+from pynini.lib.pynutil import insert
+
+from tn.processor import Processor
+from tn.utils import get_abs_path
+
+
+class TrainNumber(Processor):
+
+    def __init__(self):
+        super().__init__(name="trainnumber")
+        self.build_tagger()
+        self.build_verbalizer()
+
+    def build_tagger(self):
+        digit = string_file(get_abs_path("../itn/chinese/data/number/digit.tsv"))
+        zero = string_file(get_abs_path("../itn/chinese/data/number/zero.tsv"))
+        digits = zero | digit
+        prefix = string_file(get_abs_path("../itn/chinese/data/train_number/prefix.tsv"))
+        number = digits + digits + (digits + digits.ques).ques
+        tagger = insert('value: "') + prefix + number + insert('"')
+        self.tagger = self.add_tokens(tagger)

itn/chinese/test/data/train_number.txt

@@ -0,0 +1,7 @@
+高一二 => G12
+高一二三 => G123
+高一二三四 => G1234
+动二三 => D23
+动三四五六 => D3456
+特五六七八 => T5678
+快一零二三 => K1023

itn/chinese/test/normalizer_test.py

@@ -38,6 +38,7 @@ class TestNormalizer:
         parse_test_case("data/whitelist.txt"),
         parse_test_case("data/number.txt"),
         parse_test_case("data/license_plate.txt"),
+        parse_test_case("data/train_number.txt"),
         parse_test_case("data/normalizer.txt"),
     )