public static List <IGenericOcrRunner> GetStandardOcrs(IOcrCache ocrCache, StandardOcrSettings standardOcrSettings)
{
List <IGenericOcrRunner> genericOcrRunners = new List <IGenericOcrRunner>();
if (standardOcrSettings.AwsOcrSettings != null)
{
var awsSettings = standardOcrSettings.AwsOcrSettings;
genericOcrRunners.Add(new AwsOcrService(ocrCache, awsSettings.AccessKey, awsSettings.SecretKey));
}
if (standardOcrSettings.AzureOcrSettings != null)
{
var azureSettings = standardOcrSettings.AzureOcrSettings;
genericOcrRunners.Add(new AzureOcrExecutor(ocrCache, azureSettings.SubscriptionKey, azureSettings.Endpoint));
}
if (standardOcrSettings.GoogleOcrSettings != null)
{
var googleSettings = standardOcrSettings.GoogleOcrSettings;
genericOcrRunners.Add(new GoogleOcrService(ocrCache, googleSettings.ApiToken));
}
if (standardOcrSettings.TesseractOcrSettings != null)
{
var tesseractSettings = standardOcrSettings.TesseractOcrSettings;
genericOcrRunners.Add(new TesseractService(ocrCache, tesseractSettings.TesseractDir, tesseractSettings.DataDir));
}
if (standardOcrSettings.WindowsOcrSettings != null)
{
genericOcrRunners.Add(new WindowsOcrExecutor(ocrCache));
}
return(genericOcrRunners);
}
/// <summary>
/// Initializes a new instance of the <see cref="TesseractService"/> class.
/// </summary>
/// <param name="tesseractDir">The path for the Tesseract4 installation folder (C:\Program Files\Tesseract-OCR).</param>
/// <param name="dataDir">The data with the trained models (tessdata). Download the models from https://github.com/tesseract-ocr/tessdata_fast</param>
public TesseractService(IOcrCache ocrCache, string tesseractDir = @"C:\Program Files\Tesseract-OCR", string dataDir = null) : base(ocrCache)
{
// Tesseract configs.
_tesseractExePath = Path.Combine(tesseractDir, "tesseract.exe");
if (String.IsNullOrEmpty(dataDir))
{
dataDir = Path.Combine(tesseractDir, "tessdata");
}
Environment.SetEnvironmentVariable("TESSDATA_PREFIX", dataDir);
}
public WindowsOcrExecutor(IOcrCache ocrCache) : base(ocrCache)
{
var taskScheduler = new LimitedConcurrencyLevelTaskScheduler(1);
factory = new TaskFactory(taskScheduler);
powershell = PowerShell.Create();
string psScript = GetResourceText("Get-Text-Win-OCR.ps1");
powershell.AddScript(psScript, false);
powershell.Invoke();
this.ocrCache = ocrCache;
}
public static async Task <List <IGenericOcrRunner> > GetStandardOcrs(IOcrCache ocrCache,
StandardOcrSettings standardOcrSettings, IImageCompressor imageCompressor)
{
List <IGenericOcrRunner> genericOcrRunners = new List <IGenericOcrRunner>();
var ocrParams = new HighLevelOcrServiceParams {
OcrCache = ocrCache, ImageCompressor = imageCompressor
};
if (standardOcrSettings.AwsOcrSettings != null)
{
var awsSettings = standardOcrSettings.AwsOcrSettings;
genericOcrRunners.Add(new AwsOcrService(awsSettings.AccessKey, awsSettings.SecretKey, awsSettings.Region, ocrParams));
}
if (standardOcrSettings.AzureOcrSettings != null)
{
var azureSettings = standardOcrSettings.AzureOcrSettings;
genericOcrRunners.Add(new AzureOcrService(azureSettings.SubscriptionKey, azureSettings.Endpoint, ocrParams));
}
if (standardOcrSettings.GoogleOcrSettings != null)
{
var googleSettings = standardOcrSettings.GoogleOcrSettings;
genericOcrRunners.Add(new GoogleOcrService(googleSettings.ApiToken, ocrParams));
}
if (standardOcrSettings.TesseractOcrSettings != null)
{
var tesseractSettings = standardOcrSettings.TesseractOcrSettings;
if (tesseractSettings.ShouldTryToAutomaticallInstall && RuntimeInformation.IsOSPlatform(OSPlatform.Windows))
{
var tesseractInstaller = new WindowsTesseractInstaller(tesseractSettings.TesseractExecutable);
await tesseractInstaller.Install();
tesseractSettings.Installer = tesseractInstaller;
tesseractSettings.TesseractExecutable = tesseractInstaller.TesseractExePath;
}
if (RuntimeInformation.IsOSPlatform(OSPlatform.Linux))
{
tesseractSettings.TesseractExecutable = "tesseract";
}
genericOcrRunners.Add(new TesseractOcrService(tesseractSettings.TesseractExecutable, tesseractSettings.DataDir, ocrParams));
}
if (standardOcrSettings.WindowsOcrSettings != null && RuntimeInformation.IsOSPlatform(OSPlatform.Windows))
{
genericOcrRunners.Add(new WindowsOcrService(ocrParams));
}
return(genericOcrRunners);
}
public GenericOcrRunner(IOcrCache ocrCache)
{
this.ocrCache = ocrCache;
}
public AwsOcrService(IOcrCache ocrCache, string accessKey, string secretKey) : base(ocrCache)
{
rekognitionClient = new AmazonRekognitionClient(accessKey, secretKey);
}
public AzureOcrExecutor(IOcrCache ocrCache, string subscriptionKey, string endpoint) : base(ocrCache)
{
this.subscriptionKey = subscriptionKey;
this.endpoint = endpoint;
uriBase = endpoint + "vision/v2.1/ocr";
}
public GoogleOcrService(IOcrCache ocrCache, string apiToken) : base(ocrCache)
{
this.apiToken = apiToken;
}
public StandardMultiOcrRunnerFactory(StandardOcrSettings settings, IOcrCache ocrCache)
{
this.Settings = settings;
this.ocrCache = ocrCache;
this.imageCompressor = settings.ImageCompressor;
}
public StandardMultiOcrRunnerFactory(StandardOcrSettings settings, IOcrCache ocrCache)
{
this.settings = settings;
this.ocrCache = ocrCache;
}
