Azure 文档智能/客户端。AnalyzeDocumentAsync 不起作用

Question

client.AnalyzeDocumentFromUriAsync 工作正常。

但是当我提交与 FileFromUri 完全相同的本地文件的流时，它不会返回任何字段。这是为什么？

    Dim invoiceUri As Uri = New Uri("https://blablabla.ch/testfile_xy.pdf")
        Dim operation As AnalyzeDocumentOperation = Await client.AnalyzeDocumentFromUriAsync(WaitUntil.Completed, "prebuilt-invoice", invoiceUri)
        Dim result As AnalyzeResult = operation.Value
        
        Dim filePath = "c:\users\myname\desktop\testfile_xy.pdf" 
        Dim stream = New IO.FileStream(filePath, IO.FileMode.Open)
        Dim operation2 As AnalyzeDocumentOperation = Await client.AnalyzeDocumentAsync(WaitUntil.Completed, "prebuilt-layout", stream)
        Dim result2 As AnalyzeResult = operation2.Value
        stream.Close()

    Debug.Print(result.Documents.Count) '1  :-)
        Debug.Print(result2.Documents.Count) '0  :-(
        

为什么没有传输文档属性？

Answer 1

在从流分析文档时，

AnalyzeDocumentAsync

方法不会返回任何字段，但在从 URI 分析时却工作正常，因此遇到了错误。您可以修改代码以确保正确的流位置。

• 使用 URL 和本地路径是两种不同的方法。
• 我们不应该同时使用两者，除非我们想要同时来自本地和 URL 源的结果。该 URL 应有效。例如，我使用此示例 URL 和 Azure 存储 Blob SAS URL。

附网址：

Imports System
Imports System.IO
Imports System.Threading.Tasks
Imports Azure
Imports Azure.AI.FormRecognizer
Imports Azure.AI.FormRecognizer.DocumentAnalysis
Imports Azure.AI.FormRecognizer.Models

Module Program
    Sub Main(args As String())
        MainAsync().GetAwaiter().GetResult()
    End Sub

    Async Function MainAsync() As Task
        Dim filePath = "c:\users\myname\desktop\testfile_xy.pdf"
        Dim endpoint As String = "<endpoint>" ' Replace <endpoint> with your Form Recognizer endpoint
        Dim apiKey As String = "<apiKey>" ' Replace <apiKey> with your Form Recognizer API key

        Dim client As New DocumentAnalysisClient(New Uri(endpoint), New AzureKeyCredential(apiKey))

        ' Analyzing document from URI
        Dim invoiceUri As Uri = New Uri("https://blablabla.ch/testfile_xy.pdf")
        Dim operation As AnalyzeDocumentOperation = Await client.AnalyzeDocumentFromUriAsync(WaitUntil.Completed, "prebuilt-invoice", invoiceUri)
        Dim result As AnalyzeResult = operation.Value

        ' Analyzing document from local file stream
        Dim stream = New IO.FileStream(filePath, IO.FileMode.Open)
        stream.Seek(0, SeekOrigin.Begin) ' Reset the position of the stream to the beginning
        Dim operation2 As AnalyzeDocumentOperation = Await client.AnalyzeDocumentAsync(WaitUntil.Completed, "prebuilt-layout", stream)
        Dim result2 As AnalyzeResult = operation2.Value
        stream.Close()

        ' Displaying results
        Console.WriteLine("Results from analyzing document from URI:")
        Console.WriteLine($"Document count: {result.Documents.Count}")
        For Each document As AnalyzedDocument In result.Documents
            Console.WriteLine($"Document of type: {document.DocumentType}")
            For Each field In document.Fields
                Console.WriteLine($"Field '{field.Key}': ")
                Console.WriteLine($"  Content: '{field.Value.Content}'")
                Console.WriteLine($"  Confidence: '{field.Value.Confidence}'")
            Next
        Next

        Console.WriteLine()

        Console.WriteLine("Results from analyzing document from local file stream:")
        Console.WriteLine($"Document count: {result2.Documents.Count}")
        For Each document As AnalyzedDocument In result2.Documents
            Console.WriteLine($"Document of type: {document.DocumentType}")
            For Each field In document.Fields
                Console.WriteLine($"Field '{field.Key}': ")
                Console.WriteLine($"  Content: '{field.Value.Content}'")
                Console.WriteLine($"  Confidence: '{field.Value.Confidence}'")
            Next
        Next
    End Function
End Module

输出：

没有来自本地路径的 URL：

Imports System
Imports System.IO
Imports System.Threading.Tasks
Imports Azure
Imports Azure.AI.FormRecognizer
Imports Azure.AI.FormRecognizer.DocumentAnalysis
Imports Azure.AI.FormRecognizer.Models

Module Program
    Sub Main(args As String())
        Dim filePath As String = "C://Users//sample-1.pdf" ' Replace <filePath> with the actual path to your file

        If Not File.Exists(filePath) Then
            Console.WriteLine($"File not found: {filePath}")
            Return
        End If

        Dim endpoint As String = "<endpoint>";' Replace <endpoint> with your Form Recognizer endpoint
        Dim apiKey As String = "<apiKey>" ' Replace <apiKey> with your Form Recognizer API key

        Dim client As New DocumentAnalysisClient(New Uri(endpoint), New AzureKeyCredential(apiKey))

        Using stream As New FileStream(filePath, FileMode.Open)
            Dim operation As AnalyzeDocumentOperation = client.AnalyzeDocumentAsync(WaitUntil.Completed, "prebuilt-invoice", stream).Result
            Dim result As AnalyzeResult = operation.Value

            ' Add your processing logic here
            ' For example, accessing the extracted data:
            For Each document As AnalyzedDocument In result.Documents
                Console.WriteLine($"Document of type: {document.DocumentType}")

                For Each field In document.Fields
                    Console.WriteLine($"Field '{field.Key}': ")
                    Console.WriteLine($"  Content: '{field.Value.Content}'")
                    Console.WriteLine($"  Confidence: '{field.Value.Confidence}'")
                Next
            Next
        End Using
    End Sub
End Module