feat(itn): 支持车牌号，皖C九B三四E = >皖C9B34E (#124)

Author: xingchensong

itn/chinese/data/license_plate/province.tsv

@@ -0,0 +1,31 @@
+京
+津
+沪
+渝
+冀
+豫
+云
+辽
+黑
+湘
+皖
+鲁
+新
+苏
+浙
+赣
+鄂
+桂
+甘
+晋
+蒙
+陕
+吉
+闽
+贵
+粤
+青
+藏
+川
+宁
+琼

itn/chinese/inverse_normalizer.py

@@ -23,6 +23,7 @@
 from itn.chinese.rules.whitelist import Whitelist
 from itn.chinese.rules.time import Time
 from itn.chinese.rules.postprocessor import PostProcessor
+from itn.chinese.rules.license_plate import LicensePlate
 
 from pynini.lib.pynutil import add_weight, delete
 from importlib_resources import files
@@ -49,6 +50,7 @@ def build_tagger(self):
                   | add_weight(Time().tagger, 1.05)
                   | add_weight(Cardinal(self.convert_number, self.enable_0_to_9).tagger, 1.06)  # noqa
                   | add_weight(Math().tagger, 1.10)
+                  | add_weight(LicensePlate().tagger, 1.0)
                   | add_weight(Char().tagger, 100)).optimize()
 
         tagger = tagger.star
@@ -64,6 +66,7 @@ def build_verbalizer(self):
                       | Measure(enable_0_to_9=self.enable_0_to_9).verbalizer
                       | Money(enable_0_to_9=self.enable_0_to_9).verbalizer
                       | Time().verbalizer
+                      | LicensePlate().verbalizer
                       | Whitelist().verbalizer).optimize()
         postprocessor = PostProcessor(remove_interjections=True).processor
 

itn/chinese/rules/license_plate.py

@@ -0,0 +1,33 @@
+# Copyright (c) 2023 Xingchen Song (sxc19@mails.tsinghua.edu.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 tn.processor import Processor
+
+from pynini import string_file
+from pynini.lib.pynutil import insert
+
+
+class LicensePlate(Processor):
+
+    def __init__(self):
+        super().__init__(name='licenseplate')
+        self.build_tagger()
+        self.build_verbalizer()
+
+    def build_tagger(self):
+        digit = string_file('itn/chinese/data/number/digit.tsv')  # 1 ~ 9
+        province = string_file('itn/chinese/data/license_plate/province.tsv')  # 皖
+        license_plate = province + self.ALPHA + (self.ALPHA | digit)**5
+        tagger = insert('value: "') + license_plate + insert('"')
+        self.tagger = self.add_tokens(tagger)

itn/chinese/test/data/license_plate.txt

@@ -0,0 +1,2 @@
+鄂a七l六二u => 鄂a7l62u
+皖C九B三四E => 皖C9B34E

itn/chinese/test/normalizer_test.py

@@ -38,6 +38,7 @@ class TestNormalizer:
         parse_test_case('data/time.txt'),
         parse_test_case('data/whitelist.txt'),
         parse_test_case('data/number.txt'),
+        parse_test_case('data/license_plate.txt'),
         parse_test_case('data/normalizer.txt'))
 
     @pytest.mark.parametrize("spoken, written", normalizer_cases)
@@ -61,6 +62,7 @@ class TestNormalizerDisablestandalonenumberEnable0to9:
         parse_test_case('data/money.txt'),
         parse_test_case('data/time.txt'),
         parse_test_case('data/whitelist.txt'),
+        parse_test_case('data/license_plate.txt'),
         parse_test_case('data/normalizer_disable_standalone_number_enable_0_to_9.txt'))
 
     @pytest.mark.parametrize("spoken, written", normalizer_cases)
@@ -83,6 +85,7 @@ class TestNormalizerEnablestandalonenumberDisable0to9:
         parse_test_case('data/money.txt'),
         parse_test_case('data/time.txt'),
         parse_test_case('data/whitelist.txt'),
+        parse_test_case('data/license_plate.txt'),
         parse_test_case('data/normalizer_enable_standalone_number_disable_0_to_9.txt'))
 
     @pytest.mark.parametrize("spoken, written", normalizer_cases)
@@ -105,6 +108,7 @@ class TestNormalizerDisablestandalonenumberDisable0to9:
         parse_test_case('data/money.txt'),
         parse_test_case('data/time.txt'),
         parse_test_case('data/whitelist.txt'),
+        parse_test_case('data/license_plate.txt'),
         parse_test_case('data/normalizer_disable_standalone_number_disable_0_to_9.txt'))
 
     @pytest.mark.parametrize("spoken, written", normalizer_cases)