Document Understanding 活动

发行说明

v2.0.0 - 预览​

提取结果处理程序类​

• 基本字段，
• 字段组，
• 表格,
• 表格行，
• 通用（通配符）数据点。

字段组和货币数量支持​

v1.31.0​

新增功能​

v1.29.4​

新增功能​

v1.27.1​

v1.26.0​

新功能和改进功能​

v1.23.1​

新特性和改进功能​

v1.21.1​

错误修复​

v1.21.0​

新特性和改进功能​

• The ResultsValue Class received a new property, TextType that can provide the text origin of the extracted value.
• We've added new overload for the .Serialize method that accepts a SerializationSettings object. You can now configure whether serialization will be done using a Pascal- or camelCase convention. The default is PascalCase. This applies to the following classes: Document, DocumentTaxonomy, DocumentSplittingResult, ClassificationResult, and ExtractionResult. You can still use all the existing methods in PascalCase and everything is backwards compatible. You can use deserialization with either Pascal- or camelCase serialized objects.
• New classes have been added to the UiPath.DocumentProcessing.Contracts package that contains information about the Actions created in Action Center.
• A new helper method, IsTextTypeInDocument, is available for detecting the presence of a type of text (handwritten or checkboxes) in a document by using a TextType parameter.

错误修复​

弃用时间表​

v1.18.0​

新特性和改进功能​

• The TextType enumeration has been enriched with the following new classification fields: Barcode, QRcode, Stamp, Logo, and Unknown.
• Task Capture 用户界面现已提供繁体中文版本。

v1.17.1​

新特性和改进功能​

• 我们重构了“数据提取结果”类，以支持在工作流中使用它。 我们已实施新的表格结构和表格帮助程序方法，目前新旧结构都可用。 验证站点及其他组件（训练器、提取程序等）支持新格式，以后建议使用这种方法访问和操作结果对象中的表格数据。
• Every word from Document Object Model can now be set as type Text, Handwriting, or Checkbox.
• The ResultsValue.Components property, the ResultsValue constructors which have ResultsDataPoint[] components as parameters, the ResultsValue.CreateTableValue methods, and the ExtractionResult.FlattenFields method are now marked as obsolete.
• There is a new ProcessingSource value, named PdfAndOcr, that reflects a PDF page which was processed with both native PDF processing and OCR processing.

v1.14.0​

新特性和改进功能​

• The UiPath.DocumentProcessing.Contracts package has been upgraded to .NET5 portable, allowing you to run them on Linux robots.
• Some of the classes included in the UiPath.DocumentProcessing.Contracts package have been updated. Among the updated ones are the Simplified Value, Results Value, and Extraction Result classes.

v1.11.0​

新特性和改进功能​

• The ClassificationResult class has received new methods, Serialize and Deserialize, meant to help you serialize and deserialize the classification output.
• New methods, FlattenFields and GetFields, have also been added to the ExtractionResult class meant to help you filter the fields based on the given condition.
• The UiPath.DocumentProcessing.Contracts package has been upgraded to .NET5. While both .NET versions continue to be supported, the .NET5 projects can only work on 64-bit architectures.

错误修复​

v1.10.1​

改进​

错误修复​

• 修复了缺少 API 文档的问题。
• 修复了与版本低于 4.11.0 的UiPath.IntelligentOCR.Activities 的向后兼容性。

v1.9.1​

错误修复​

v1.9.0​

新特性和改进功能​

• 在以下情况下，允许值提取为 null（空）：
  • 没有引用的数据提取字段；
  • 没有引用的文档分类；
  • Values created by the used in Validation Station with no reference.
• ExtractionResults 中增加了对自定义提取程序有效负载的支持。
• 在分类法中增加了对字段级元数据的支持。

v1.6.1​

新特性和改进功能​

• 名为 VisualLineNumber 的新属性已添加到公共类词中，指明该词放置在哪个可视行上。这可在 UiPath.DocumentProcessing.Contracts.Dom 命名空间下找到。
• 名为 GetVisualTextProjection 的新扩展方法已添加到文档类中，使您可以访问词的视觉排列。这可在 UiPath.DocumentProcessing.Contracts.Dom 命名空间下找到。

v1.4.0​

新特性和改进功能​

v1.3.0​

新特性和改进功能​

• This release brings a new feature for the UiPath.DocumentProcessing.Contracts pack.
• 公共类词在词水平框中添加了一个附加内容，现在支持词级别的多边形信息。

v1.2.0​

新特性和改进功能​

v1.1.0​

新特性和改进功能​

v1.0.0​

新特性和改进功能​

• 请引用此包中提供的合同。
• 在工作流中实现自己的分类器和提取程序。