Get values from JSON result C#

Question

I am trying to get the values from a computer vision api result from reading text on an image using Azure. The output is JSON data but the syntax of the result looks strange.

Ultimately I am trying to get the value "text" stripped out of it and written to a text file without any escape characters etc.

Here is the code I am using the parse the result.

static async Task MakeOCRRequest(string imageFilePath)
{
    try
    {
        HttpClient client = new HttpClient();
        client.DefaultRequestHeaders.Add("Ocp-Apim-Subscription-Key", subscriptionKey);
        string requestParameters = "language=unk&detectOrientation=true";
        string uri = uriBase + "?" + requestParameters;
        HttpResponseMessage response;
        byte[] byteData = GetImageAsByteArray(imageFilePath);

        using (ByteArrayContent content = new ByteArrayContent(byteData))
        {
            content.Headers.ContentType = new MediaTypeHeaderValue("application/octet-stream");
            response = await client.PostAsync(uri, content);
        }

        string contentString = await response.Content.ReadAsStringAsync();

        ///////  It is at this point that I want to get the values from the "text" field
        JToken token = JToken.Parse(contentString).ToString();
        String[] result = contentString.Split(',');
        Console.writeline("\nResponse:\n\n{}\n", JToken.Parse(contentString).ToString());

    }
    catch (Exception e)
    {
        Console.WriteLine("\n" + e.Message);
    }
}

And here is the result that I get from the OCR process. I havent included the full result as it represents over 1700 lines.

  "language": "en",
  "textAngle": 0.0,
  "orientation": "Right",
  "regions": [
    {
      "boundingBox": "140,300,639,420",
      "lines": [
        {
          "boundingBox": "419,300,87,15",
          "words": [
            {
              "boundingBox": "419,300,87,15",
              "text": "0000175351"
            }
          ]
        },
        {
          "boundingBox": "140,342,337,47",
          "words": [
            {
              "boundingBox": "140,347,92,38",
              "text": "WE."
            },
            {
              "boundingBox": "241,347,13,36",
              "text": "1"
            },
            {
              "boundingBox": "266,342,211,47",
              "text": "0/1-1.9(2)"
            }
          ]
        },

With the current code I get the error message

JObject textResult = token["regions"]["text"].Value<JObject>();

Cannot access child value on NewtonSoft.Json.Linq.JValue.

I wonder if I am requesting the wrong key?

Answer 1

If you need to retrieve all text property value regardless of boundingBox then you can use below linq after parsing your json to JToken.

JToken jToken = JToken.Parse(json);

var allTexts = jToken["regions"].SelectMany(reg => reg["lines"].SelectMany(line => line["words"]).Select(word => word["text"].ToString()).ToList()).ToList();

Output: (From Debugger)

Answer 2

Suppose that you have a valid JSON string now you can use Newtonsoft.Json package and deserialize your json string to object and then use object to get values:

ResponseModel res = JsonConvert.DeserializeObject<ResponseModel>(contentString);

your response model could be a pocco class like this:

public class ResponseModel
{
    public string language { get; set; }
    public string textAngle { get; set; }
    public string orientation { get; set; }
    //you have to create pocco class for RegionModel
    public List<RegionModel> regions { get; set; }
    ....
}