From 3747ceada6c78f526aa846b5b90f3e665c64765e Mon Sep 17 00:00:00 2001
From: kotaro-kinoshita <kotaro.masin.218@gmail.com>
Date: Thu, 28 Nov 2024 23:18:05 +0900
Subject: [PATCH] add text normalize

---
 src/yomitoku/text_recognizer.py | 3 +++
 uv.lock                         | 2 +-
 2 files changed, 4 insertions(+), 1 deletion(-)

diff --git a/src/yomitoku/text_recognizer.py b/src/yomitoku/text_recognizer.py
index 38d6ff1..082d394 100644
--- a/src/yomitoku/text_recognizer.py
+++ b/src/yomitoku/text_recognizer.py
@@ -2,6 +2,7 @@
 
 import numpy as np
 import torch
+import unicodedata
 from pydantic import conlist
 
 from .base import BaseModelCatalog, BaseModule, BaseSchema
@@ -72,6 +73,8 @@ def preprocess(self, img, polygons):
 
     def postprocess(self, p, points):
         pred, score = self.tokenizer.decode(p)
+        pred = [unicodedata.normalize("NFKC", x) for x in pred]
+
         directions = []
         for point in points:
             point = np.array(point)
diff --git a/uv.lock b/uv.lock
index b6372e6..77e07c7 100644
--- a/uv.lock
+++ b/uv.lock
@@ -1530,7 +1530,7 @@ wheels = [
 
 [[package]]
 name = "yomitoku"
-version = "0.4.1"
+version = "0.5.1"
 source = { editable = "." }
 dependencies = [
     { name = "huggingface-hub" },