/// <summary>
/// Converts the single json file into several PAGE xml files.
/// </summary>
/// <param name="inputFilePath">The path to the json file.</param>
/// <param name="outputFolderPath">The folder that will contain the PAGE xml files.</param>
public static void Convert(string inputFilePath, string outputFolderPath)
{
using (FileStream s = File.Open(inputFilePath, FileMode.Open))
using (StreamReader sr = new StreamReader(s))
using (JsonReader reader = new JsonTextReader(sr))
{
JsonSerializer serializer = new JsonSerializer();
var cocoFile = serializer.Deserialize <CocoFile>(reader);
Dictionary <int, string> categories = cocoFile.categories.ToDictionary(k => k.id, k => k.name);
int totalImageCount = cocoFile.images.Count;
for (int i = 0; i < totalImageCount; i++)
{
var image = cocoFile.images[i];
string outputFilePath = Path.ChangeExtension(Path.Combine(outputFolderPath, image.file_name), "xml");
if (File.Exists(outputFilePath))
{
continue;
}
var entry = new CocoEntry(image.id, image.file_name, image.height, image.width);
var annotations = cocoFile.annotations.Where(a => a.image_id == image.id);
foreach (var annotation in annotations)
{
entry.Annotations.Add(annotation);
}
var pageXml = Get(entry, categories);
File.WriteAllText(outputFilePath, pageXml);
Console.WriteLine("Done: \t" + Path.GetFileName(outputFilePath) + "\t" + i + @"/" + totalImageCount);
}
}
}
private static PageXmlDocument.PageXmlPage ToPageXmlPage(CocoEntry page, Dictionary <int, string> categories)
{
var pageXmlPage = new PageXmlDocument.PageXmlPage()
{
ImageFilename = page.FileName,
ImageHeight = (int)Math.Round(page.Height),
ImageWidth = (int)Math.Round(page.Width),
};
var regions = new List <PageXmlDocument.PageXmlRegion>();
foreach (var annotation in page.Annotations)
{
var category = categories[annotation.category_id];
var segmentations = annotation.GetSegmentationPoints()[0];
switch (category)
{
case "title":
regions.Add(ToPageXmlTextRegion(annotation.id, segmentations, PageXmlDocument.PageXmlTextSimpleType.Heading));
break;
case "text":
regions.Add(ToPageXmlTextRegion(annotation.id, segmentations, PageXmlDocument.PageXmlTextSimpleType.Paragraph));
break;
case "list":
regions.Add(ToPageXmlTextRegion(annotation.id, segmentations, PageXmlDocument.PageXmlTextSimpleType.LisLabel));
break;
case "figure":
regions.Add(ToPageXmlImageRegion(annotation.id, segmentations));
break;
case "table":
regions.Add(ToPageXmlTableRegion(annotation.id, segmentations));
break;
default:
regions.Add(ToPageXmlUnknownRegion(annotation.id, segmentations));
break;
}
}
pageXmlPage.Items = regions.ToArray();
return(pageXmlPage);
}
/// <summary>
/// Get the PAGE-XML (XML) string of the pages layout.
/// </summary>
/// <param name="page"></param>
/// <param name="includePaths">Draw <see cref="PdfPath"/>s present in the page.</param>
private static string Get(CocoEntry page, Dictionary <int, string> categories)
{
PageXmlDocument pageXmlDocument = new PageXmlDocument()
{
Metadata = new PageXmlDocument.PageXmlMetadata()
{
Created = DateTime.UtcNow,
LastChange = DateTime.UtcNow,
Creator = "PublayNetConverter",
Comments = "PubLayNet dataset"
},
PcGtsId = "pc" + page.Id.ToString()
};
pageXmlDocument.Page = ToPageXmlPage(page, categories);
return(Serialize(pageXmlDocument));
}