Imports System
Imports System.Collections.Generic
Imports System.IO
Imports System.Net.Http
Imports System.Text.Json
Imports System.Threading
Imports System.Threading.Tasks

' OCR.space: what you build before the first line of your actual application
Public Class OcrSpaceApiClient
    Implements IDisposable

    Private ReadOnly _httpClient As HttpClient
    Private ReadOnly _apiKey As String
    Private ReadOnly _rateLimiter As SemaphoreSlim
    Private Const ApiEndpoint As String = "https://api.ocr.space/parse/image"
    Private Const MaxFileSizeFree As Integer = 5 * 1024 * 1024 ' 5MB — hard limit on free tier
    Private Const RateLimitPerMinute As Integer = 60

    Public Sub New(apiKey As String)
        _apiKey = If(apiKey, Throw New ArgumentNullException(NameOf(apiKey)))
        _httpClient = New HttpClient()
        _httpClient.Timeout = TimeSpan.FromSeconds(120)
        _rateLimiter = New SemaphoreSlim(RateLimitPerMinute, RateLimitPerMinute)
    End Sub

    Public Async Function ExtractTextFromFileAsync(filePath As String, Optional language As String = "eng", Optional cancellationToken As CancellationToken = Nothing) As Task(Of OcrResult)
        If Not File.Exists(filePath) Then
            Throw New FileNotFoundException("Image file not found", filePath)
        End If

        Dim fileInfo = New FileInfo(filePath)
        If fileInfo.Length > MaxFileSizeFree Then
            Throw New InvalidOperationException($"File size {fileInfo.Length \ 1024 \ 1024}MB exceeds free tier limit of 5MB.")
        End If

        Await _rateLimiter.WaitAsync(cancellationToken)

        Try
            ' Document leaves your infrastructure here
            Dim imageBytes As Byte() = Await File.ReadAllBytesAsync(filePath, cancellationToken)
            Dim base64Image As String = Convert.ToBase64String(imageBytes)
            Dim mimeType As String = GetMimeType(filePath)

            Dim formContent = New FormUrlEncodedContent(New KeyValuePair(Of String, String)() {
                New KeyValuePair(Of String, String)("apikey", _apiKey),
                New KeyValuePair(Of String, String)("base64Image", $"data:{mimeType};base64,{base64Image}"),
                New KeyValuePair(Of String, String)("language", language),
                New KeyValuePair(Of String, String)("isOverlayRequired", "false"),
                New KeyValuePair(Of String, String)("filetype", Path.GetExtension(filePath).TrimStart("."c)),
                New KeyValuePair(Of String, String)("detectOrientation", "true"),
                New KeyValuePair(Of String, String)("scale", "true"),
                New KeyValuePair(Of String, String)("OCREngine", "2")
            })

            Dim response = Await _httpClient.PostAsync(ApiEndpoint, formContent, cancellationToken)

            If Not response.IsSuccessStatusCode Then
                Dim errorBody = Await response.Content.ReadAsStringAsync(cancellationToken)
                Throw New OcrSpaceException($"OCR.space API returned {response.StatusCode}: {errorBody}", CInt(response.StatusCode))
            End If

            Dim jsonResponse = Await response.Content.ReadAsStringAsync(cancellationToken)
            Return ParseOcrResponse(jsonResponse)
        Finally
            _ = Task.Run(Async Function()
                             Await Task.Delay(60000 \ RateLimitPerMinute)
                             _rateLimiter.Release()
                         End Function)
        End Try
    End Function

    Private Function ParseOcrResponse(jsonResponse As String) As OcrResult
        Using doc = JsonDocument.Parse(jsonResponse)
            Dim root = doc.RootElement

            If root.TryGetProperty("IsErroredOnProcessing", ByRef isErrored) AndAlso isErrored.GetBoolean() Then
                Dim errorMessage As String = "OCR processing failed"
                If root.TryGetProperty("ErrorMessage", ByRef errorMessages) Then
                    ' Parse the array of error strings manually
                    Dim messages = New List(Of String)()
                    If errorMessages.ValueKind = JsonValueKind.Array Then
                        For Each msg In errorMessages.EnumerateArray()
                            messages.Add(msg.GetString() OrElse "Unknown error")
                        Next
                    End If
                    errorMessage = String.Join("; ", messages)
                End If
                Throw New OcrSpaceException(errorMessage, 500)
            End If

            Dim result = New OcrResult()

            If root.TryGetProperty("ParsedResults", ByRef parsedResults) Then
                For Each parsedResult In parsedResults.EnumerateArray()
                    If parsedResult.TryGetProperty("ParsedText", ByRef parsedText) Then
                        result.ParsedText &= parsedText.GetString()
                    End If

                    If parsedResult.TryGetProperty("FileParseExitCode", ByRef exitCode) Then
                        result.ExitCodes.Add(exitCode.GetInt32())
                    End If
                Next
            End If

            Return result
        End Using
    End Function

    Private Function GetMimeType(filePath As String) As String
        Return Path.GetExtension(filePath).ToLowerInvariant() Select Case
            Case ".png" : Return "image/png"
            Case ".jpg", ".jpeg" : Return "image/jpeg"
            Case ".bmp" : Return "image/bmp"
            Case ".tiff", ".tif" : Return "image/tiff"
            Case ".pdf" : Return "application/pdf"
            Case Else : Return "application/octet-stream"
        End Select
    End Function

    Public Sub Dispose() Implements IDisposable.Dispose
        _httpClient?.Dispose()
        _rateLimiter?.Dispose()
    End Sub
End Class

' Custom models — no SDK means you define these yourself
Public Class OcrResult
    Public Property ParsedText As String = String.Empty
    Public ReadOnly Property Warnings As New List(Of String)()
    Public ReadOnly Property ExitCodes As New List(Of Integer)()
End Class

Public Class OcrSpaceException
    Inherits Exception

    Public ReadOnly Property StatusCode As Integer

    Public Sub New(message As String, statusCode As Integer)
        MyBase.New(message)
        Me.StatusCode = statusCode
    End Sub
End Class

Imports System
Imports System.Collections.Generic
Imports System.IO
Imports System.Net.Http
Imports System.Text.Json
Imports System.Threading
Imports System.Threading.Tasks

' OCR.space: what you build before the first line of your actual application
Public Class OcrSpaceApiClient
    Implements IDisposable

    Private ReadOnly _httpClient As HttpClient
    Private ReadOnly _apiKey As String
    Private ReadOnly _rateLimiter As SemaphoreSlim
    Private Const ApiEndpoint As String = "https://api.ocr.space/parse/image"
    Private Const MaxFileSizeFree As Integer = 5 * 1024 * 1024 ' 5MB — hard limit on free tier
    Private Const RateLimitPerMinute As Integer = 60

    Public Sub New(apiKey As String)
        _apiKey = If(apiKey, Throw New ArgumentNullException(NameOf(apiKey)))
        _httpClient = New HttpClient()
        _httpClient.Timeout = TimeSpan.FromSeconds(120)
        _rateLimiter = New SemaphoreSlim(RateLimitPerMinute, RateLimitPerMinute)
    End Sub

    Public Async Function ExtractTextFromFileAsync(filePath As String, Optional language As String = "eng", Optional cancellationToken As CancellationToken = Nothing) As Task(Of OcrResult)
        If Not File.Exists(filePath) Then
            Throw New FileNotFoundException("Image file not found", filePath)
        End If

        Dim fileInfo = New FileInfo(filePath)
        If fileInfo.Length > MaxFileSizeFree Then
            Throw New InvalidOperationException($"File size {fileInfo.Length \ 1024 \ 1024}MB exceeds free tier limit of 5MB.")
        End If

        Await _rateLimiter.WaitAsync(cancellationToken)

        Try
            ' Document leaves your infrastructure here
            Dim imageBytes As Byte() = Await File.ReadAllBytesAsync(filePath, cancellationToken)
            Dim base64Image As String = Convert.ToBase64String(imageBytes)
            Dim mimeType As String = GetMimeType(filePath)

            Dim formContent = New FormUrlEncodedContent(New KeyValuePair(Of String, String)() {
                New KeyValuePair(Of String, String)("apikey", _apiKey),
                New KeyValuePair(Of String, String)("base64Image", $"data:{mimeType};base64,{base64Image}"),
                New KeyValuePair(Of String, String)("language", language),
                New KeyValuePair(Of String, String)("isOverlayRequired", "false"),
                New KeyValuePair(Of String, String)("filetype", Path.GetExtension(filePath).TrimStart("."c)),
                New KeyValuePair(Of String, String)("detectOrientation", "true"),
                New KeyValuePair(Of String, String)("scale", "true"),
                New KeyValuePair(Of String, String)("OCREngine", "2")
            })

            Dim response = Await _httpClient.PostAsync(ApiEndpoint, formContent, cancellationToken)

            If Not response.IsSuccessStatusCode Then
                Dim errorBody = Await response.Content.ReadAsStringAsync(cancellationToken)
                Throw New OcrSpaceException($"OCR.space API returned {response.StatusCode}: {errorBody}", CInt(response.StatusCode))
            End If

            Dim jsonResponse = Await response.Content.ReadAsStringAsync(cancellationToken)
            Return ParseOcrResponse(jsonResponse)
        Finally
            _ = Task.Run(Async Function()
                             Await Task.Delay(60000 \ RateLimitPerMinute)
                             _rateLimiter.Release()
                         End Function)
        End Try
    End Function

    Private Function ParseOcrResponse(jsonResponse As String) As OcrResult
        Using doc = JsonDocument.Parse(jsonResponse)
            Dim root = doc.RootElement

            If root.TryGetProperty("IsErroredOnProcessing", ByRef isErrored) AndAlso isErrored.GetBoolean() Then
                Dim errorMessage As String = "OCR processing failed"
                If root.TryGetProperty("ErrorMessage", ByRef errorMessages) Then
                    ' Parse the array of error strings manually
                    Dim messages = New List(Of String)()
                    If errorMessages.ValueKind = JsonValueKind.Array Then
                        For Each msg In errorMessages.EnumerateArray()
                            messages.Add(msg.GetString() OrElse "Unknown error")
                        Next
                    End If
                    errorMessage = String.Join("; ", messages)
                End If
                Throw New OcrSpaceException(errorMessage, 500)
            End If

            Dim result = New OcrResult()

            If root.TryGetProperty("ParsedResults", ByRef parsedResults) Then
                For Each parsedResult In parsedResults.EnumerateArray()
                    If parsedResult.TryGetProperty("ParsedText", ByRef parsedText) Then
                        result.ParsedText &= parsedText.GetString()
                    End If

                    If parsedResult.TryGetProperty("FileParseExitCode", ByRef exitCode) Then
                        result.ExitCodes.Add(exitCode.GetInt32())
                    End If
                Next
            End If

            Return result
        End Using
    End Function

    Private Function GetMimeType(filePath As String) As String
        Return Path.GetExtension(filePath).ToLowerInvariant() Select Case
            Case ".png" : Return "image/png"
            Case ".jpg", ".jpeg" : Return "image/jpeg"
            Case ".bmp" : Return "image/bmp"
            Case ".tiff", ".tif" : Return "image/tiff"
            Case ".pdf" : Return "application/pdf"
            Case Else : Return "application/octet-stream"
        End Select
    End Function

    Public Sub Dispose() Implements IDisposable.Dispose
        _httpClient?.Dispose()
        _rateLimiter?.Dispose()
    End Sub
End Class

' Custom models — no SDK means you define these yourself
Public Class OcrResult
    Public Property ParsedText As String = String.Empty
    Public ReadOnly Property Warnings As New List(Of String)()
    Public ReadOnly Property ExitCodes As New List(Of Integer)()
End Class

Public Class OcrSpaceException
    Inherits Exception

    Public ReadOnly Property StatusCode As Integer

    Public Sub New(message As String, statusCode As Integer)
        MyBase.New(message)
        Me.StatusCode = statusCode
    End Sub
End Class

net

net

Imports System.IO
Imports System.Net.Http
Imports System.Text.Json
Imports System.Threading.Tasks

Public Class OCRSpaceClient
    Private ReadOnly _apiKey As String
    Private ReadOnly _httpClient As HttpClient

    Public Sub New(apiKey As String, httpClient As HttpClient)
        _apiKey = apiKey
        _httpClient = httpClient
    End Sub

    Public Async Function ExtractTextFromPdf(pdfPath As String) As Task(Of List(Of String))
        Dim pageTexts As New List(Of String)()

        ' Reject before wasting a request quota entry
        Dim fileInfo As New FileInfo(pdfPath)
        If fileInfo.Length > 5 * 1024 * 1024 Then
            Throw New InvalidOperationException("File exceeds 5MB free tier limit. Upgrade or split PDF.")
        End If

        Dim pdfBytes As Byte() = File.ReadAllBytes(pdfPath)
        Dim base64 As String = Convert.ToBase64String(pdfBytes)

        Dim content As New FormUrlEncodedContent(New KeyValuePair(Of String, String)() {
            New KeyValuePair(Of String, String)("apikey", _apiKey),
            New KeyValuePair(Of String, String)("base64Image", $"data:application/pdf;base64,{base64}"),
            New KeyValuePair(Of String, String)("language", "eng"),
            New KeyValuePair(Of String, String)("filetype", "PDF"),
            New KeyValuePair(Of String, String)("isCreateSearchablePdf", "false") ' Watermarked on free tier
        })

        Dim response As HttpResponseMessage = Await _httpClient.PostAsync("https://api.ocr.space/parse/image", content)

        If Not response.IsSuccessStatusCode Then
            Throw New Exception($"API error: {response.StatusCode}")
        End If

        Dim json As String = Await response.Content.ReadAsStringAsync()
        Using doc As JsonDocument = JsonDocument.Parse(json)
            If doc.RootElement.TryGetProperty("ParsedResults", results) Then
                For Each result In results.EnumerateArray()
                    If result.TryGetProperty("ParsedText", text) Then
                        pageTexts.Add(If(text.GetString(), String.Empty))
                    End If
                Next
            End If
        End Using

        Return pageTexts
    End Function
End Class

Imports System.IO
Imports System.Net.Http
Imports System.Text.Json
Imports System.Threading.Tasks

Public Class OCRSpaceClient
    Private ReadOnly _apiKey As String
    Private ReadOnly _httpClient As HttpClient

    Public Sub New(apiKey As String, httpClient As HttpClient)
        _apiKey = apiKey
        _httpClient = httpClient
    End Sub

    Public Async Function ExtractTextFromPdf(pdfPath As String) As Task(Of List(Of String))
        Dim pageTexts As New List(Of String)()

        ' Reject before wasting a request quota entry
        Dim fileInfo As New FileInfo(pdfPath)
        If fileInfo.Length > 5 * 1024 * 1024 Then
            Throw New InvalidOperationException("File exceeds 5MB free tier limit. Upgrade or split PDF.")
        End If

        Dim pdfBytes As Byte() = File.ReadAllBytes(pdfPath)
        Dim base64 As String = Convert.ToBase64String(pdfBytes)

        Dim content As New FormUrlEncodedContent(New KeyValuePair(Of String, String)() {
            New KeyValuePair(Of String, String)("apikey", _apiKey),
            New KeyValuePair(Of String, String)("base64Image", $"data:application/pdf;base64,{base64}"),
            New KeyValuePair(Of String, String)("language", "eng"),
            New KeyValuePair(Of String, String)("filetype", "PDF"),
            New KeyValuePair(Of String, String)("isCreateSearchablePdf", "false") ' Watermarked on free tier
        })

        Dim response As HttpResponseMessage = Await _httpClient.PostAsync("https://api.ocr.space/parse/image", content)

        If Not response.IsSuccessStatusCode Then
            Throw New Exception($"API error: {response.StatusCode}")
        End If

        Dim json As String = Await response.Content.ReadAsStringAsync()
        Using doc As JsonDocument = JsonDocument.Parse(json)
            If doc.RootElement.TryGetProperty("ParsedResults", results) Then
                For Each result In results.EnumerateArray()
                    If result.TryGetProperty("ParsedText", text) Then
                        pageTexts.Add(If(text.GetString(), String.Empty))
                    End If
                Next
            End If
        End Using

        Return pageTexts
    End Function
End Class

Imports System
Imports System.Collections.Generic
Imports System.IO
Imports System.Net.Http
Imports System.Text.Json
Imports System.Threading.Tasks

Public Class OCRSpace
    Public Async Function SafeExtract(client As HttpClient, apiKey As String, imagePath As String) As Task(Of String)
        Try
            Dim imageBytes As Byte() = File.ReadAllBytes(imagePath)
            Dim base64 As String = Convert.ToBase64String(imageBytes)

            Dim content As New FormUrlEncodedContent(New KeyValuePair(Of String, String)() {
                New KeyValuePair(Of String, String)("apikey", apiKey),
                New KeyValuePair(Of String, String)("base64Image", $"data:image/png;base64,{base64}")
            })

            Dim response As HttpResponseMessage = Await client.PostAsync("https://api.ocr.space/parse/image", content)

            If Not response.IsSuccessStatusCode Then
                Select Case response.StatusCode
                    Case System.Net.HttpStatusCode.Unauthorized
                        Throw New Exception("Invalid API key")
                    Case System.Net.HttpStatusCode.TooManyRequests
                        Throw New Exception("Rate limit exceeded — wait and retry")
                    Case System.Net.HttpStatusCode.PaymentRequired
                        Throw New Exception("Quota exceeded — upgrade plan")
                    Case Else
                        Throw New Exception($"API error: {response.StatusCode}")
                End Select
            End If

            ' Second error layer: JSON-level failures
            Dim json As String = Await response.Content.ReadAsStringAsync()
            Using doc As JsonDocument = JsonDocument.Parse(json)
                If doc.RootElement.TryGetProperty("IsErroredOnProcessing", ByRef isError) AndAlso isError.GetBoolean() Then
                    Dim messages As New List(Of String)()
                    If doc.RootElement.TryGetProperty("ErrorMessage", ByRef errors) Then
                        For Each e In errors.EnumerateArray()
                            messages.Add(e.GetString() ?? "Unknown error")
                        Next
                    End If
                    Throw New Exception(String.Join("; ", messages))
                End If

                ' Third layer: per-page exit codes
                If doc.RootElement.TryGetProperty("ParsedResults", ByRef results) Then
                    For Each result In results.EnumerateArray()
                        If result.TryGetProperty("FileParseExitCode", ByRef exitCode) AndAlso exitCode.GetInt32() <> 1 Then
                            Throw New Exception($"Page parse failed: exit code {exitCode.GetInt32()}")
                        End If
                    Next
                End If

                Return doc.RootElement _
                    .GetProperty("ParsedResults")(0) _
                    .GetProperty("ParsedText") _
                    .GetString() ?? String.Empty
            End Using
        Catch ex As JsonException
            Throw New Exception($"Invalid API response: {ex.Message}")
        Catch ex As HttpRequestException
            Throw New Exception($"Network error: {ex.Message}")
        End Try
    End Function
End Class

Imports System
Imports System.Collections.Generic
Imports System.IO
Imports System.Net.Http
Imports System.Text.Json
Imports System.Threading.Tasks

Public Class OCRSpace
    Public Async Function SafeExtract(client As HttpClient, apiKey As String, imagePath As String) As Task(Of String)
        Try
            Dim imageBytes As Byte() = File.ReadAllBytes(imagePath)
            Dim base64 As String = Convert.ToBase64String(imageBytes)

            Dim content As New FormUrlEncodedContent(New KeyValuePair(Of String, String)() {
                New KeyValuePair(Of String, String)("apikey", apiKey),
                New KeyValuePair(Of String, String)("base64Image", $"data:image/png;base64,{base64}")
            })

            Dim response As HttpResponseMessage = Await client.PostAsync("https://api.ocr.space/parse/image", content)

            If Not response.IsSuccessStatusCode Then
                Select Case response.StatusCode
                    Case System.Net.HttpStatusCode.Unauthorized
                        Throw New Exception("Invalid API key")
                    Case System.Net.HttpStatusCode.TooManyRequests
                        Throw New Exception("Rate limit exceeded — wait and retry")
                    Case System.Net.HttpStatusCode.PaymentRequired
                        Throw New Exception("Quota exceeded — upgrade plan")
                    Case Else
                        Throw New Exception($"API error: {response.StatusCode}")
                End Select
            End If

            ' Second error layer: JSON-level failures
            Dim json As String = Await response.Content.ReadAsStringAsync()
            Using doc As JsonDocument = JsonDocument.Parse(json)
                If doc.RootElement.TryGetProperty("IsErroredOnProcessing", ByRef isError) AndAlso isError.GetBoolean() Then
                    Dim messages As New List(Of String)()
                    If doc.RootElement.TryGetProperty("ErrorMessage", ByRef errors) Then
                        For Each e In errors.EnumerateArray()
                            messages.Add(e.GetString() ?? "Unknown error")
                        Next
                    End If
                    Throw New Exception(String.Join("; ", messages))
                End If

                ' Third layer: per-page exit codes
                If doc.RootElement.TryGetProperty("ParsedResults", ByRef results) Then
                    For Each result In results.EnumerateArray()
                        If result.TryGetProperty("FileParseExitCode", ByRef exitCode) AndAlso exitCode.GetInt32() <> 1 Then
                            Throw New Exception($"Page parse failed: exit code {exitCode.GetInt32()}")
                        End If
                    Next
                End If

                Return doc.RootElement _
                    .GetProperty("ParsedResults")(0) _
                    .GetProperty("ParsedText") _
                    .GetString() ?? String.Empty
            End Using
        Catch ex As JsonException
            Throw New Exception($"Invalid API response: {ex.Message}")
        Catch ex As HttpRequestException
            Throw New Exception($"Network error: {ex.Message}")
        End Try
    End Function
End Class

' At application startup — once, not per request
IronOcr.License.LicenseKey = Environment.GetEnvironmentVariable("IRONOCR_LICENSE")

' At application startup — once, not per request
IronOcr.License.LicenseKey = Environment.GetEnvironmentVariable("IRONOCR_LICENSE")

Imports IronOcr

Using input As New OcrInput()
    input.LoadImage("low-quality-scan.jpg")
    input.Deskew()
    input.DeNoise()
    input.Contrast()
    input.Binarize()
    input.EnhanceResolution(300)
    Dim result = New IronTesseract().Read(input)
End Using

Imports IronOcr

Using input As New OcrInput()
    input.LoadImage("low-quality-scan.jpg")
    input.Deskew()
    input.DeNoise()
    input.Contrast()
    input.Binarize()
    input.EnhanceResolution(300)
    Dim result = New IronTesseract().Read(input)
End Using