yomitoku 0.5.0__tar.gz → 0.5.2__tar.gz

{yomitoku-0.5.0 → yomitoku-0.5.2}/PKG-INFO

@@ -1,11 +1,11 @@
 Metadata-Version: 2.3
 Name: yomitoku
-Version: 0.5.0
+Version: 0.5.2
 Summary: Yomitoku is an AI-powered document image analysis package designed specifically for the Japanese language.
 Author-email: Kotaro Kinoshita <kotaro.kinoshita@mlism.com>
 License: CC BY-NC-SA 4.0
 Keywords: Deep Learning,Japanese,OCR
-Requires-Python: >=3.9
+Requires-Python: <3.13,>=3.9
 Requires-Dist: huggingface-hub>=0.26.1
 Requires-Dist: lxml>=5.3.0
 Requires-Dist: omegaconf>=2.3.0
@@ -15,7 +15,7 @@ Requires-Dist: pydantic>=2.9.2
 Requires-Dist: pypdfium2>=4.30.0
 Requires-Dist: shapely>=2.0.6
 Requires-Dist: timm>=1.0.11
-Requires-Dist: torch==2.5.0
+Requires-Dist: torch>=2.5.0
 Requires-Dist: torchvision>=0.20.0
 Description-Content-Type: text/markdown
 
@@ -27,7 +27,7 @@ Description-Content-Type: text/markdown
 ![Pytorch](https://img.shields.io/badge/Pytorch-2.5-EE4C2C.svg?logo=Pytorch&style=fla)
 ![CUDA](https://img.shields.io/badge/CUDA->=11.8-76B900.svg?logo=NVIDIA&style=fla)
 ![OS](https://img.shields.io/badge/OS-Linux|Mac|Win-1793D1.svg?&style=fla)
-[![Document](https://img.shields.io/badge/docs-live-brightgreen)](https://kotaro-kinoshita.github.io/yomitoku-dev/)
+[![Document](https://img.shields.io/badge/docs-live-brightgreen)](https://kotaro-kinoshita.github.io/yomitoku/)
 
 ## 🌟 概要
 
@@ -61,7 +61,7 @@ Markdown でエクスポートした結果は関してはリポジトリ内の[s
 
 ## 📣 リリース情報
 
-- 2024 年 11 月 26 YomiToku v0.5.0 ベータ版を公開
+- 2024 年 11 月 26 日 YomiToku v0.5.1 (beta) を公開
 
 ## 💡 インストールの方法
 
@@ -101,7 +101,7 @@ yomitoku --help
 
 ## 📝 ドキュメント
 
-パッケージの詳細は[ドキュメント](https://kotaro-kinoshita.github.io/yomitoku-dev/)を確認してください。
+パッケージの詳細は[ドキュメント](https://kotaro-kinoshita.github.io/yomitoku/)を確認してください。
 
 ## LICENSE
 

{yomitoku-0.5.0 → yomitoku-0.5.2}/README.md

@@ -6,7 +6,7 @@
 ![Pytorch](https://img.shields.io/badge/Pytorch-2.5-EE4C2C.svg?logo=Pytorch&style=fla)
 ![CUDA](https://img.shields.io/badge/CUDA->=11.8-76B900.svg?logo=NVIDIA&style=fla)
 ![OS](https://img.shields.io/badge/OS-Linux|Mac|Win-1793D1.svg?&style=fla)
-[![Document](https://img.shields.io/badge/docs-live-brightgreen)](https://kotaro-kinoshita.github.io/yomitoku-dev/)
+[![Document](https://img.shields.io/badge/docs-live-brightgreen)](https://kotaro-kinoshita.github.io/yomitoku/)
 
 ## 🌟 概要
 
@@ -40,7 +40,7 @@ Markdown でエクスポートした結果は関してはリポジトリ内の[s
 
 ## 📣 リリース情報
 
-- 2024 年 11 月 26 YomiToku v0.5.0 ベータ版を公開
+- 2024 年 11 月 26 日 YomiToku v0.5.1 (beta) を公開
 
 ## 💡 インストールの方法
 
@@ -80,7 +80,7 @@ yomitoku --help
 
 ## 📝 ドキュメント
 
-パッケージの詳細は[ドキュメント](https://kotaro-kinoshita.github.io/yomitoku-dev/)を確認してください。
+パッケージの詳細は[ドキュメント](https://kotaro-kinoshita.github.io/yomitoku/)を確認してください。
 
 ## LICENSE
 

{yomitoku-0.5.0 → yomitoku-0.5.2}/README_EN.md

@@ -41,7 +41,7 @@ Source of the image: Created by processing content from “Reiwa 6 Edition Infor
 
 ## 📣 Release
 
-- Released YomiToku vX.X.X on December XX, 2024.
+- Released YomiToku v0.5.1 (beta) on November 26, 2024.
 
 ## 💡 Installation
 

{yomitoku-0.5.0 → yomitoku-0.5.2}/mkdocs.yml

@@ -74,11 +74,10 @@ repo_url: https://github.com/kotaro-kinoshita/yomitoku-dev
 plugins:
   - search:
   - i18n:
-      default_language: ja
       languages:
         - locale: ja
           name: 日本語
           default: true
         - locale: en
           name: English
-          link: /en/
+          link: /yomitoku/en/

{yomitoku-0.5.0 → yomitoku-0.5.2}/pyproject.toml

@@ -12,7 +12,7 @@ authors = [{name = "Kotaro Kinoshita", email = "kotaro.kinoshita@mlism.com"}]
 description = "Yomitoku is an AI-powered document image analysis package designed specifically for the Japanese language."
 readme = "README.md"
 license = {text = "CC BY-NC-SA 4.0"}
-requires-python = ">=3.9"
+requires-python = ">=3.9,<3.13"
 keywords = ["Japanese", "OCR", "Deep Learning"]
 dependencies = [
     "huggingface-hub>=0.26.1",
@@ -24,7 +24,7 @@ dependencies = [
     "shapely>=2.0.6",
     "timm>=1.0.11",
     "torchvision>=0.20.0",
-    "torch==2.5.0",
+    "torch>=2.5.0",
     "pypdfium2>=4.30.0",
 ]
 
@@ -79,12 +79,18 @@ commands =
     pytest tests
 
 [testenv:lint]
+basepython = python3.12
 deps = ruff
 commands = 
     ruff check
 
 [testenv:docs]
-deps = mkdocs-material
+basepython = python3.12
+deps = 
+    mkdocs-material
+    pygments 
+    plantuml_markdown 
+    mkdocs-static-i18n
 commands = 
     mkdocs gh-deploy --force
 """

{yomitoku-0.5.0 → yomitoku-0.5.2}/src/yomitoku/text_recognizer.py

@@ -2,6 +2,7 @@ from typing import List
 
 import numpy as np
 import torch
+import unicodedata
 from pydantic import conlist
 
 from .base import BaseModelCatalog, BaseModule, BaseSchema
@@ -72,6 +73,8 @@ class TextRecognizer(BaseModule):
 
     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)