From 810046e3df4910c8f5c1a90e4b53aca45b3397e8 Mon Sep 17 00:00:00 2001
From: wuhongsheng <664116298@qq.com>
Date: 星期一, 01 七月 2024 10:42:58 +0800
Subject: [PATCH] 优化merge segments 参数,解决新闻联播男女主持人“晚上好”合并一个speakid问题 (#1861)
---
fun_text_processing/inverse_text_normalization/fr/verbalizers/time.py | 24 +++++++++++++++---------
1 files changed, 15 insertions(+), 9 deletions(-)
diff --git a/fun_text_processing/inverse_text_normalization/fr/verbalizers/time.py b/fun_text_processing/inverse_text_normalization/fr/verbalizers/time.py
index c75c69c..0de583f 100644
--- a/fun_text_processing/inverse_text_normalization/fr/verbalizers/time.py
+++ b/fun_text_processing/inverse_text_normalization/fr/verbalizers/time.py
@@ -1,4 +1,3 @@
-
import pynini
from fun_text_processing.inverse_text_normalization.fr.graph_utils import (
DAMO_DIGIT,
@@ -15,7 +14,7 @@
Finite state transducer for verbalizing time, e.g.
time { hours: "8" minutes: "30" suffix: "du matin"} -> 8 h 30
time { hours: "8" minutes: "30" } -> 8 h 30
- time { hours: "8" minutes: "30" suffix: "du soir"} -> 20 h 30
+ time { hours: "8" minutes: "30" suffix: "du soir"} -> 20 h 30
"""
def __init__(self):
@@ -23,25 +22,32 @@
hour_to_night = pynini.string_file(get_abs_path("data/time/hour_to_night.tsv"))
- day_suffixes = pynutil.delete("suffix: \"am\"")
- night_suffixes = pynutil.delete("suffix: \"pm\"")
+ day_suffixes = pynutil.delete('suffix: "am"')
+ night_suffixes = pynutil.delete('suffix: "pm"')
hour = (
pynutil.delete("hours:")
+ delete_space
- + pynutil.delete("\"")
+ + pynutil.delete('"')
+ pynini.closure(DAMO_DIGIT, 1, 2)
- + pynutil.delete("\"")
+ + pynutil.delete('"')
)
minute = (
pynutil.delete("minutes:")
+ delete_extra_space
- + pynutil.delete("\"")
+ + pynutil.delete('"')
+ pynini.closure(DAMO_DIGIT, 1, 2)
- + pynutil.delete("\"")
+ + pynutil.delete('"')
)
- graph = hour + delete_extra_space + pynutil.insert("h") + minute.ques + delete_space + day_suffixes.ques
+ graph = (
+ hour
+ + delete_extra_space
+ + pynutil.insert("h")
+ + minute.ques
+ + delete_space
+ + day_suffixes.ques
+ )
graph |= (
hour @ hour_to_night
--
Gitblit v1.9.1