public static void Run()
{
string altoFilePath = @"D:\MachineLearning\Document Layout Analysis\hocr\Glyph_Sample01_General.xml"; // Glyph_Sample01_General.xml";
AltoDocument alto = AltoDocument.Deserialize(altoFilePath);
var xml = alto.Serialize();
File.WriteAllText(Path.ChangeExtension(altoFilePath, "new.xml"), xml);
}
private string Serialize(AltoDocument altoDocument)
{
var serializer = new XmlSerializer(typeof(AltoDocument));
var settings = new XmlWriterSettings
{
Encoding = System.Text.Encoding.UTF8,
Indent = true,
IndentChars = indentChar,
};
using (var memoryStream = new System.IO.MemoryStream())
using (var xmlWriter = XmlWriter.Create(memoryStream, settings))
{
serializer.Serialize(xmlWriter, altoDocument);
return(System.Text.Encoding.UTF8.GetString(memoryStream.ToArray()));
}
}
public static void Run(string path)
{
float zoom = 10;
var pinkPen = new Pen(Color.HotPink, zoom * 0.4f);
var greenPen = new Pen(Color.GreenYellow, zoom * 0.6f);
var bluePen = new Pen(Color.Blue, zoom * 1.0f);
AltoDocument alto = new AltoDocument()
{
Layout = new AltoLayout()
{
StyleRefs = null
},
Description = GetAltoDescription("filename"),
SchemaVersion = "SCHEMAVERSION1",
Styles = new AltoStyles()
{
},
Tags = new AltoTags()
{
}
};
List <AltoPage> altoPages = new List <AltoPage>();
using (PdfDocument document = PdfDocument.Open(path))
{
var testAlto = AltoDocument.FromPdfDocument(document);
for (var i = 0; i < document.NumberOfPages; i++)
{
var pagePdf = document.GetPage(i + 1);
Stopwatch stopwatch = new Stopwatch();
stopwatch.Start();
var words = pagePdf.GetWords(NearestNeighbourWordExtractor.Instance);
stopwatch.Stop();
Console.WriteLine("GetWords() - Time elapsed: {0}", stopwatch.Elapsed);
var pageWordsH = words.Where(x => x.TextDirection == TextDirection.Horizontal || x.TextDirection == TextDirection.Rotate180).ToArray();
stopwatch.Reset();
stopwatch.Start();
var blocks = RecursiveXYCut.Instance.GetBlocks(pageWordsH);
stopwatch.Stop();
Console.WriteLine("RecursiveXYCut() - Time elapsed: {0}", stopwatch.Elapsed);
altoPages.Add(new AltoPage()
{
Height = (float)pagePdf.Height,
Width = (float)pagePdf.Width,
Accuracy = float.NaN,
Quality = AltoQuality.OK,
QualityDetail = null,
BottomMargin = null,
LeftMargin = null,
RightMargin = null,
TopMargin = null,
Pc = float.NaN,
PhysicalImgNr = pagePdf.Number,
PrintedImgNr = null,
PageClass = null,
Position = AltoPosition.Cover,
Processing = null,
ProcessingRefs = null,
StyleRefs = null,
PrintSpace = new AltoPageSpace()
{
Height = (float)pagePdf.Height, // TBD
Width = (float)pagePdf.Width, // TBD
VPos = 0f, // TBD
HPos = 0f, // TBD
ComposedBlocks = null, // TBD
GraphicalElements = null, // TBD
Illustrations = null, // TBD
ProcessingRefs = null, // TBD
StyleRefs = null, // TBD
TextBlock = blocks.Select(b => ToAltoTextBlock(b, pagePdf.Height)).ToArray(),
Id = "NA"
},
Id = "NA"
});
}
}
alto.Layout.Pages = altoPages.ToArray();
XmlSerializer xsSubmit = new XmlSerializer(typeof(AltoDocument));
var xml = "";
using (var sww = new StringWriter())
{
using (XmlTextWriter writer = new XmlTextWriter(sww)
{
Formatting = Formatting.Indented
})
{
xsSubmit.Serialize(writer, alto);
xml = sww.ToString(); // Your XML
}
}
File.WriteAllText(Path.ChangeExtension(path, "xml"), xml);
}