/*
* Run with your own PDF or use the sample Amtrak-Financials file
* provided in this repo.
*/
async static Task Main(string[] args)
{
if (args.Length == 0)
{
Console.WriteLine("Please provide the filepath of a PDF to OCR");
Environment.Exit(0);
}
IndicoConfig config = new IndicoConfig(
host: "app.indico.io"
);
IndicoClient client = new IndicoClient(config);
JObject extractConfig = new JObject()
{
{ "preset_config", "standard" }
};
DocumentExtraction ocrQuery = client.DocumentExtraction(extractConfig);
Job job = await ocrQuery.Exec(args[0]);
JObject result = await job.Result();
string url = (string)result.GetValue("url");
Blob blob = await client.RetrieveBlob(url).Exec();
Console.WriteLine(blob.AsJSONObject());
}
/// <summary>
/// Create a new DocumentExtraction client to OCR files
/// </summary>
/// <param name="jsonConfig">DocumentExtraction passed in as a JSON Object. Defaults to null</param>
/// <returns>DocumentExtraction</returns>
public DocumentExtraction DocumentExtraction(JObject jsonConfig = null)
{
DocumentExtraction ocr = new DocumentExtraction(this);
if (jsonConfig != null)
{
ocr.JsonConfig = jsonConfig;
}
return(ocr);
}
public void DocumentExtraction_Initialize()
{
try
{
_extractionService = new DocumentExtraction();
}
catch (Exception ex)
{
Assert.Fail(ex.ToString());
}
}
/*
* To actually run this example, you'll need to train a sequence
* model on the Indico IPA Platform with the labeled-swaps-200.csv
* file contained with this repo. Be aware that training will likely
* take a couple hours. Once training is complete, you can run the
* example by passing in the directory of the two sample PDF files in
* the repo (Confirmation letter and Confirmation of Interest Rate Swap).
*
* Before running, replace the Model Group ID (mgId) with the
* ID for your trained model. You can find it on the model's Review page.
*/
async static Task Main(string[] args)
{
// Replace this with your Model Group ID
int mgId = 4352;
List <string> targetFiles = GetTargetFiles(args[0]);
if (targetFiles.Count == 0)
{
Console.WriteLine("No files to process");
Environment.Exit(0);
}
IndicoClient client = new IndicoClient();
JObject extractConfig = new JObject()
{
{ "preset_config", "legacy" }
};
List <string> texts = new List <string>();
DocumentExtraction ocrQuery = client.DocumentExtraction(extractConfig);
foreach (string path in targetFiles)
{
Console.WriteLine(path);
Job ocrJob = await ocrQuery.Exec(path);
JObject result = await ocrJob.Result();
string resUrl = (string)result.GetValue("url");
Blob blob = await client.RetrieveBlob(resUrl).Exec();
JObject obj = blob.AsJSONObject();
texts.Add((string)obj.GetValue("text"));
}
ModelGroup mg = await client.ModelGroupQuery(mgId).Exec();
string status = await client.ModelGroupLoad(mg).Exec();
Console.WriteLine($"Model status = {status}");
Job job = await client.ModelGroupPredict(mg).Data(texts).Exec();
JArray jobResult = await job.Results();
Console.WriteLine(jobResult);
}
public async Task <string> ExtractDocumentAsync(string filePath, DocumentExtractionPreset preset, CancellationToken cancellationToken)
{
var config = new JObject
{
{ "preset_config", preset.ToString("F").ToLower() }
};
var docExtraction =
new DocumentExtraction(_indicoClientLegacy)
{
JsonConfig = config,
};
var job = await docExtraction.Exec(filePath, cancellationToken);
return(job.Id);
}
public static string Invoke(this DocumentExtraction documentExtractionActivity) => documentExtractionActivity.Invoke <DocumentExtraction, string>((a, outArg) => a.Results = outArg);
