private static OcrWord ReadWord(BinaryReader reader)
{
OcrWord word = new OcrWord();
word.Rect = ReadRect(reader);
word.Confidence = reader.ReadInt32();
word.Text = reader.ReadString();
return(word);
}
public OcrLine AddLine(string line)
{
if (line == null)
{
line = string.Empty;
}
string[] words = line.Split();
OcrLine ocrLine = new OcrLine();
foreach (string word in words)
{
OcrWord ocrWord = new OcrWord();
ocrWord.Text = word;
ocrLine.Words.Add(ocrWord);
}
Lines.Add(ocrLine);
return(ocrLine);
}
public static OcrResult Parse(string xml)
{
Dictionary <string, string> meta = new Dictionary <string, string>();
List <Page> pages = new List <Page>();
XmlReaderSettings settings = new XmlReaderSettings();
settings.DtdProcessing = DtdProcessing.Parse;
settings.XmlResolver = new XmlPreloadedResolver(XmlKnownDtds.Xhtml10);
using (StringReader stringReader = new StringReader(xml))
using (XmlReader reader = XmlReader.Create(stringReader, settings))
{
while (reader.Read())
{
switch (reader.NodeType)
{
case XmlNodeType.Element:
switch (reader.Name)
{
case "meta":
{
string metaName = reader.GetAttribute("name");
string metaContent = reader.GetAttribute("content");
string metaValue = reader.GetAttribute("value");
if (!string.IsNullOrEmpty(metaName))
{
meta[metaName] = metaContent == null ? metaValue : metaContent;
}
}
break;
case "div":
case "span":
case "p":
{
switch (reader.GetAttribute("class"))
{
case "ocr_page":
pages.Add(Parse <Page>(reader, null));
break;
}
}
break;
}
break;
case XmlNodeType.EndElement:
break;
}
}
}
OcrResult result = new OcrResult();
foreach (Page page in pages)
{
foreach (Area area in page.Areas)
{
OcrArea ocrArea = new OcrArea(area.Rect.ToRect());
List <Line> lines = new List <Line>();
lines.AddRange(area.Lines);
foreach (Paragraph paragraph in area.Paragraphs)
{
lines.AddRange(paragraph.Lines);
}
foreach (Line line in lines)
{
OcrLine ocrLine = new OcrLine();
ocrLine.Rect = line.Rect.ToRect();
foreach (Word word in line.Words)
{
OcrWord ocrWord = new OcrWord();
ocrWord.Rect = word.Rect.ToRect();
ocrWord.Confidence = word.Confidence;
ocrWord.Text = word.Text;
ocrLine.Words.Add(ocrWord);
}
ocrArea.Lines.Add(ocrLine);
}
if (!string.IsNullOrEmpty(ocrArea.Text.Trim()))
{
result.AddArea(ocrArea);
}
}
}
return(result);
}
private static void WriteWord(BinaryWriter writer, OcrWord word)
{
WriteRect(writer, word.Rect);
writer.Write(word.Confidence);
writer.Write(word.Text == null ? string.Empty : word.Text);
}