Not sure where to post this or if I should make 2 posts. Currently we are trying to convert PDFs to CSV. Some PDFs convert without any problems, however, we have some that just return an empty string (Problem 1). As a work around we have attempted to convert the PDF into a jpeg then run OCR. The OCR is returning a random string (Problem 2). Any help would be appreciated. Thanks in advance.
Code:
public HttpResponseMessage GetText(string path)
{
var fileInfo = new FileInfo(path);
if (!fileInfo.Exists)
{
Log.Warning(string.Format("Method: GetImage(string). File not found : {0}", path));
return null;//this.NotFound();
}
//open document
var pdfDocument = new Document(path);
var excelsave = new ExcelSaveOptions { MinimizeTheNumberOfWorksheets = true };
var exceldoc = new Aspose.Cells.Workbook();
using (var stream = new MemoryStream())
{
pdfDocument.Flatten();
pdfDocument.Save(stream, excelsave);
exceldoc = new Aspose.Cells.Workbook(stream);
}
if (!IsPdf(exceldoc))
exceldoc = OcrPdf(pdfDocument);// throw new Exception("FIle cannot be parsed.");
using (var returnStream = new MemoryStream())
{
exceldoc.Save(returnStream, Aspose.Cells.SaveFormat.CSV);
returnStream.Seek(0, 0);
var result = new HttpResponseMessage(HttpStatusCode.OK)
{
Content = new ByteArrayContent(returnStream.GetBuffer())
};
result.Content.Headers.ContentDisposition = new System.Net.Http.Headers.ContentDispositionHeaderValue("attachment")
{
FileName = "my.csv"
};
result.Content.Headers.ContentType = new MediaTypeHeaderValue("application/octet-stream");
return result;
}
}
private Workbook OcrPdf(Document pdfDocument)
{
var ocr = new OcrEngine();
var sb = new StringBuilder();
for (int pageCount = 1; pageCount <= pdfDocument.Pages.Count; pageCount++)
{
using (var imageStream = new MemoryStream())
{
var resolution = new Resolution(300);
var jpegDevice = new JpegDevice(Convert.ToInt32(pdfDocument.Pages[pageCount].PageInfo.Width), Convert.ToInt32(pdfDocument.Pages[pageCount].PageInfo.Height), resolution, 100);
jpegDevice.Process(pdfDocument.Pages[pageCount], imageStream);
ocr.Image = ImageStream.FromMemoryStream(imageStream, ImageStreamFormat.Jpg);
if (ocr.Process())
{
sb.Append(ocr.Text);
sb.Append(Environment.NewLine);
}
imageStream.Close();
}
}
return null;
}
private bool IsPdf(Workbook exceldoc)
{
return exceldoc.Worksheets[0].Cells[0, 0].Value != null;
}
}
OCR result of first page:
",,Ny,NMANNNNT!%3%:TTEW,phipp:VNNWNNNNT}:n-npW#%#?%N\r\n"