Release v2.0.2

🚀 Features

(rapidocr) Support img url as input and optimize vis (#356)

rapidocr -img "https://github.com/RapidAI/RapidOCR/blob/main/python/tests/test_files/ch_en_num.jpg?raw=true" --vis_res

📚 Documentation

• Update README

Bump! Release rapidocr v2.0.1

⚠️注意：该版本不与rapidocr_onnxruntime<=1.4.4兼容。遵循语义化版本号规则，请谨慎使用，以免出现各种问题。

更新要点

• 发布rapidocr库，整合onnxruntime、openvino、torch和paddlepaddle四个推理引擎，大家可按需安装对应的推理引擎来使用。
• 支持torch版本作为推理引擎
• 与PaddleOCR参数部分对齐
• 整理模型到modelscope，纳入版本管理，相应模型自动下载对齐
• 采用dataclass类来封装函数返回值，简化使用
• 优化输入配置文件，保证可以通过配置文件使用，便于分享参数设置
• 文档添加版本管理 → RapidOCR Doc，尚未来得及完善，doing

更新主要想法

• 考虑到终端命令行使用有限，且维护有些繁琐，新版本会弱化终端部分代码，仅保留快速验证代码。

进度：

• 2025-03-09: 完善参数校验和自动化发版流程，发布rapidocr==2.0.1
• 2025-03-05: 添加文档版本管理，为后续编写v2.0.0文档做准备
• 2025-03-04: 优化调用默认server模型代码，通过添加lang_det和lang_rec来实现。至此基本功能均已经实现。
• 2025-02-18: 整理pytorch版本到rapidocr中；添加自动下载字体文件功能；添加对命令行执行的单元测试
• 2025-02-14: 完善各个引擎下模型自动下载并调用功能
• 2025-02-13: 优化logger，解决重复输出问题
• 2025-02-12: 添加已有模型的自动下载功能
• 2025-02-11: 优化初始化RapidOCR传参问题
• 2025-02-10: 采用抽象基类，统一四个推理引擎代码接口
• 2025-02-08：
  • 将PP-OCRv4所有模型均转换并托管在modelscope上
  • 将推理引擎部分单独出来，声明基类，统一不同模态间的调用方式

相关issue #329

Full Changelog: v2.0.0...v2.0.1

Release v1.4.4

常规更新，去掉无效代码

🚀 Features

• Support det mean and std config (#312)
• (rapidocr_api) Add uvicorn timestamp
• (rapidocr_api) 为uvicorn日志添加时间戳 (#309)

🐛 Bug Fixes

• (rapidocr_api) 补充依赖，指定版本号 (#305)
• (rapidocr_openvino) Merge PR #293 #294
• (rapidocr_paddle) Merge PR #293 #294

⚙️ Miscellaneous Tasks

• (rapidocr_paddle) Update paddpaddle version
• Update python version
• Optimize code
• Update modelscope online demo link
• (rapidocr_api) Update python version
• Remove the stale issue bot

New Contributors

• @imweijh made their first contribution in #305

Full Changelog: v1.4.3...v1.4.4

Release v1.4.3

What's Changed

• 优化单字符坐标及返回单字符置信度 by @jesse01 in #293
• fix: ensure 4 point for char box rec by @Joker1212 in #294

New Contributors

• @jesse01 made their first contribution in #293

Full Changelog: v1.4.2...v1.4.3

Release v1.4.2

主要修复issue #288 问题

What's Changed

• add extras_require by @hbh112233abc in #281
• try import RapidOCR by @hbh112233abc in #282

New Contributors

• @hbh112233abc made their first contribution in #281

Full Changelog: v1.4.1...v1.4.2

Release v1.4.1

中文版（by @Joker1212 ）：

1. 支持英文单字识别
2. 解决只切割中文字符导致英文字符识别框缩小的问题
3. 修正单字切割后排序偶发的点重叠问题

What's Changed

• Update README_zh.md by @Decem-Y in #268
• Sup en char rec by @Joker1212 in #272

New Contributors

• @Decem-Y made their first contribution in #268

Release v1.4.0

🚀 Features

Char rec(单字识别) (#254)。支持返回单字坐标，汉字返回单字的坐标，英文返回单词级别坐标。
大概原理是基于文本行长度和文字多少来分割出单字坐标的，泛化性比较有限。

主要借鉴PP-Structure和fanqie03/char-detection

具体用法可参加：docs
可视化结果参见：docs

示例如下：

🐛 Bug Fixes

• (python) Fix incorrect usage of the exists function
• (rapidocr_onnxruntime) Fix normalize value of det (#246)

📚 Documentation

• (rapidocr_api) Fix typo errors
• Add modelscope demo link
• (rapidocr_paddle) Fixed bad links

🧪 Testing

• Update unit tests for different platform

⚙️ Miscellaneous Tasks

• Optimze char rec (#260)
• (rapidocr_onnxruntime) Change read_yaml order (#251)
• Upgrade actions/checkout version to 4

New Contributors

• @xmxoxo made their first contribution in #242
• @whmlol made their first contribution in #253
• @Pumpkinor made their first contribution in #251
• @Joker1212 made their first contribution in #254

Release v1.3.25

🐛 Bug Fixes

• Fixed issue #232：
• 目前输入一张图像的处理逻辑是：
  • 当输入图像最大边大于max_side_len时，会将该图像最大边等比缩放到max_side_len（默认2000px），另一边按照宽高比等比缩放
  • 当输入图像最小边小于min_side_len时，会将该图像最小边放大到min_side_len(默认30px)，另一边按照宽高比缩放

📚 Documentation

• Update QQ group
• Update demo line and colab demo
• Update docs
• Update docs link of api and ocrweb
• Update README
• (README) Add OpenSSF badge
• Update README
• Add more visualization
• Update README
• Fix typo
• Update README

⚙️ Miscellaneous Tasks

• Compatible with python 3.7
• Remove scorecard workflow
• (rapidocr_openvino/paddle) Optim code
• Optimize code structure
• Fix the Sponsor link
• Update LICENSE authors
• (rapidocr_api) Optim auto package logic
• (python) Remove the project urls of whl packages

New Contributors

• @benben17 made their first contribution in #200
• @dependabot made their first contribution in #208

Release v1.3.24

🐛 Bug Fixes

• (python) Fixed issue #202 and add Shapely version limited

⚙️ Miscellaneous Tasks

• (submodule) Remove submodules of others program languages
• (rapidocr_web) Fix setup error when auto publish whl

Release v1.3.23

🐛 Bug Fixes

• (python) Restrict dependent package version, especially numpy versions

🚜 Refactor

• (rapidocr_paddle) Refactor the entire code
• (rapidocr_openvino) Refactor the entire code
• (rapidocr_onnxruntime) Refactor the pre-processing code for the det part
• (rapidocr_onnxruntime) Add typing annotations of RapidOCR call
• (rapidocr_onnxruntime) Add typing annotations and remove invalid code

📚 Documentation

• (README) Remove invalid QQ number
• ()* Update docs path
• (README) Add AI Studio online demo