What's Changed

• 修复_load_imgs中isinstance(img_content,InputType)错误 by @hbh112233abc in #140

Full Changelog: v3.0.0...v3.0.1

rapid_tablev3.0.0 主要更新是支持batch推理，模型并没有升级哈！

因为版本号是根据语义化版本号来走的，这次更改了接口的返回值。因此从v2.0.3升级到了v3.0.0。
返回值具体变化如下：

# v2.0.3
class RapidTableOutput:
    img: Optional[np.ndarray] = None
    pred_html: Optional[str] = None
    cell_bboxes: Optional[np.ndarray] = None
    logic_points: Optional[np.ndarray] = None
    elapse: Optional[float] = None
    
# v3.0.0
class RapidTableOutput:
    imgs: List[np.ndarray] = field(default_factory=list)
    pred_htmls: List[str] = field(default_factory=list)
    cell_bboxes: List[np.ndarray] = field(default_factory=list)
    logic_points: List[np.ndarray] = field(default_factory=list)
    elapse: float = 0.0

重构代码。自己逐渐发现修改代码的水平会随着想法逐渐改变的。这也像改PPT，永远没有尽头。哈哈！

我深知自己写的代码远没有达到优雅，但是我在努力中，希望大家在这个旅途中，能够收获快乐，玩得开心！

🚀 Features

• support batch inference both model and refacter code by @SWHL in a4fd085
• finished unitable refacter by @SWHL in 4fbba21
• add batch processing for images by @deeperrrr in 6a57308

🐛 Bug Fixes

• fixed issue #126 by @SWHL in 7d18543
• fix path error when inputing a url by @SWHL in 02e0a0e

📚 Documentation

• update README by @SWHL in 03cbe1e
• update README by @SWHL in a0d56d5

🧪 Testing

• adapted batch infer by @SWHL in 9f02fcd

⚙️ Miscellaneous Tasks

• update files by @SWHL in 6bb3e10
• update files by @SWHL in 732c29e

🎉 Contributors

• @SWHL
• @rming
• @deeperrrr

Full Changelog: 3.0.0

🐛 Bug Fixes

• fixed issue #114 by @SWHL in 7a871ef
• fixed issue #112 by @SWHL in 51bb0ec
• fix the image path error in the url input from the cli by @SWHL in d1ffbf4

📚 Documentation

• update README by @SWHL in 1d26ad5

Full Changelog: 2.0.3

What's Changed

• remove table_structure.init.py imports by @hbh112233abc in #103

Full Changelog: v2.0.1...v2.0.2

修复打包过程中丢失yaml问题

What's Changed

• fix: python>=3.12正则产生的warnings by @koevas1226 in #98
• config传参默认可以不传,解决默认onnx,报须torch库的问题 by @hbh112233abc in #97
• feat: adapt rapidocr v3 and refactor code by @SWHL in #99

New Contributors

• @koevas1226 made their first contribution in #98
• @hbh112233abc made their first contribution in #97

Full Changelog: v1.0.5...v2.0.0

修复限制python3.13问题

主要是放开了numpy<2.0的限制，适配rapidocr 2.0版本

🚀 Features

• Add colorlog
• Add unit testing of cli

🐛 Bug Fixes

• Fix cli error
• Fixed find_packages module
• Logic points vis adapt

⚙️ Miscellaneous Tasks

• Add rapidocr package
• Optimize code

What's Changed

• fix: logic points vis adapt by @Joker1212 in #70
• chore: optimize code by @SWHL in #73

Full Changelog: v1.0.3...v1.0.4

⚠️注意：本次更新版本为v1.x，不兼容v0.x版本，请谨慎更新，避免导致接口调用有误。

主要更新

1. RapidTable的输入输出做了更新，采用dataclasses来封装，简化参数传递，便于后续使用，更新和维护。示例如下：

   # 输入
   @dataclass
   class RapidTableInput:
      model_type: Optional[str] = ModelType.SLANETPLUS.value
      model_path: Union[str, Path, None, Dict[str, str]] = None
      use_cuda: bool = False
      device: str = "cpu"
   
   # 输出
   @dataclass
   class RapidTableOutput:
      pred_html: Optional[str] = None
      cell_bboxes: Optional[np.ndarray] = None
      logic_points: Optional[np.ndarray] = None
      elapse: Optional[float] = None
   
   # 使用示例
   input_args = RapidTableInput(model_type="unitable")
   table_engine = RapidTable(input_args)
   
   img_path = 'test_images/table.jpg'
   table_results = table_engine(img_path)
   
   print(table_results.pred_html)

2. 集成了Unitable项目最新表格识别算法，推理引擎为torch，提升了RapidTable的上限。
3. 优化了模型下载和托管问题，模型托管在modelscope上。在使用时，只需要指定对应的model_type，即可自动下载对应模型。当然，也可以通过model_path来具体指定模型路径。

What's Changed in v1.0.3

• fix: fix cli error
• Merge pull request #43 from Joker1212/fix_ci
• ci: fix setup

Full Changelog: https://github.com///compare/v1.0.2...v1.0.3

⚠️注意：本次更新版本为v1.x，不兼容v0.x版本，请谨慎更新，避免导致接口调用有误。

主要更新

1. RapidTable的输入输出做了更新，采用dataclasses来封装，简化参数传递，便于后续使用，更新和维护。示例如下：

   # 输入
   @dataclass
   class RapidTableInput:
      model_type: Optional[str] = ModelType.SLANETPLUS.value
      model_path: Union[str, Path, None, Dict[str, str]] = None
      use_cuda: bool = False
      device: str = "cpu"
   
   # 输出
   @dataclass
   class RapidTableOutput:
      pred_html: Optional[str] = None
      cell_bboxes: Optional[np.ndarray] = None
      logic_points: Optional[np.ndarray] = None
      elapse: Optional[float] = None
   
   # 使用示例
   input_args = RapidTableInput(model_type="unitable")
   table_engine = RapidTable(input_args)
   
   img_path = 'test_images/table.jpg'
   table_results = table_engine(img_path)
   
   print(table_results.pred_html)

2. 集成了Unitable项目最新表格识别算法，推理引擎为torch，提升了RapidTable的上限。
3. 优化了模型下载和托管问题，模型托管在modelscope上。在使用时，只需要指定对应的model_type，即可自动下载对应模型。当然，也可以通过model_path来具体指定模型路径。

What's Changed

• feat: add unitable torch inference by @Joker1212 in #35
• Dev unitable by @Joker1212 in #37
• feat: support unitable and optimize code by @SWHL in #40
• fix: adapt row&col span decode for unitable by @Joker1212 in #42

New Contributors

• @SWHL made their first contribution in #40

Full Changelog: v0.1.0...v1.0.2