public AnnotatedPage(OcrResult hw, ImageReference image, int pageNumber)
{
// page
metadata.WriteLine($" <div class='ocr_page' id='page_{pageNumber}' title='image \"{image.Url}\"; bbox 0 0 {image.Width} {image.Height}; ppageno {pageNumber}'>");
metadata.WriteLine($" <div class='ocr_carea' id='block_{pageNumber}_1'>");
var allwords = new List <WordResult>();
int li = 0;
int wi = 0;
foreach (var line in hw.lines)
{
metadata.WriteLine($" <span class='ocr_line' id='line_{pageNumber}_{li}' title='baseline -0.002 -5; x_size 30; x_descenders 6; x_ascenders 6'>");
var words = line.words.FirstOrDefault()?.boundingBox == null ? line.words : line.words.OrderBy(l => l.boundingBox[0]).ToArray();
foreach (var word in words)
{
var bbox = word.boundingBox != null && word.boundingBox.Length == 8 ? $"bbox {word.boundingBox[0]} {word.boundingBox[1]} {word.boundingBox[4]} {word.boundingBox[5]}" : "";
metadata.WriteLine($" <span class='ocrx_word' id='word_{pageNumber}_{li}_{wi}' title='{bbox}'>{word.text}</span>");
text.WriteLine(word.text);
wi++;
allwords.Add(word);
}
li++;
metadata.WriteLine(" </span>"); // line
}
metadata.WriteLine(" </div>"); // reading area
metadata.WriteLine(" </div>"); // page
}
public async Task <OcrResult> GetText(Stream stream, string url = null, string name = null)
{
var visionResult = await(stream != null ? visionClient.RecognizeTextAsync(stream, "en", true) : visionClient.RecognizeTextAsync(url, "en", true));
var lines = visionResult.Regions.SelectMany(r => r.Lines).Select(l =>
new lineResult()
{
boundingBox = ConvertBoundingBox(l.BoundingBox),
words = l.Words.Select(w => new WordResult()
{
boundingBox = ConvertBoundingBox(w.BoundingBox),
text = w.Text
}).ToArray()
}
);
var result = new OcrResult()
{
lines = lines.ToArray(),
Orientation = visionResult.Orientation
};
return(result);
}
public async Task <OcrResult> GetVision(Stream stream, string url = null)
{
var features = new[] { VisualFeature.Tags, VisualFeature.ImageType, VisualFeature.Description, VisualFeature.Adult };
var visionResult = stream != null ?
await visionClient.AnalyzeImageAsync(stream, features)
: await visionClient.AnalyzeImageAsync(url, features);
List <lineResult> lines = new List <lineResult>();
lines.AddRange(visionResult.Description.Captions.Select(c => new lineResult()
{
words = c.Text.Split(' ').Select(w => new WordResult()
{
text = w
}).ToArray()
}
));
lines.Add(new lineResult()
{
words = new[] { "(" }
.Concat(visionResult.Tags.Select(t => t.Name))
.Concat(new[] { ")" })
.Select(t => new WordResult()
{
text = t
}).ToArray()
});
var result = new OcrResult()
{
lines = lines.ToArray()
};
return(result);
}
public AnnotatedPage(OcrResult hw, ImageReference image) : this(hw, image, 0)
{
}
private async Task <OcrResult> GetHandwritingTextImpl(Stream stream, string url)
{
var client = new HttpClient();
client.DefaultRequestHeaders.Add("Ocp-Apim-Subscription-Key", apiKey);
var uri = "https://westus.api.cognitive.microsoft.com/vision/v1.0/recognizeText?handwriting=true";
HttpResponseMessage response;
// Request body
if (stream != null)
{
using (var content = new StreamContent(stream))
{
content.Headers.ContentType = new MediaTypeHeaderValue("application/octet-stream");
response = await client.PostAsync(uri, content);
}
}
else
{
var json = JsonConvert.SerializeObject(new { url = url });
using (var content = new StringContent(json))
{
content.Headers.ContentType = new MediaTypeHeaderValue("application/json");
response = await client.PostAsync(uri, content);
}
}
OcrResult result = null;
IEnumerable <string> opLocation;
if (!response.IsSuccessStatusCode)
{
var err = await response.Content.ReadAsStringAsync();
response.EnsureSuccessStatusCode();
}
if (response.Headers.TryGetValues("Operation-Location", out opLocation))
{
while (true)
{
response = await client.GetAsync(opLocation.First());
var txt = await response.Content.ReadAsStringAsync();
var status = JsonConvert.DeserializeObject <AsyncStatusResult>(txt);
if (status.status == "Running" || status.status == "NotStarted")
{
await Task.Delay(TimeSpan.FromMilliseconds(100));
}
else
{
result = status.recognitionResult;
break;
}
}
}
return(result);
}
