Support for hORC, AtloXml and PageXml output formats

Tested with:
- 'hocrjs' for hORC (see https://unpkg.com/hocrjs)
- 'PAGE Viewer' for hORC, AtloXml and PageXml (see http://www.primaresearch.org/tools/PAGEViewer)

src/UglyToad.PdfPig/Content/Page.cs

@@ -177,6 +177,49 @@
             {
                 return letter.PointSize;
             }
+
+            /// <summary>
+            /// Get the hOCR (html) string of the page layout.
+            /// <para>This is considered experimental because it needs more testing.</para>
+            /// </summary>
+            /// <param name="wordExtractor">The word extractor to use to generate words.</param>
+            /// <param name="pageSegmenter">The page segmenter to use.</param>
+            /// <param name="indent">Indent character to use.</param>
+            /// <param name="drawPaths">Draw <see cref="PdfPath"/>s present in the page.</param>
+            /// <param name="useHocrjs">Will add a reference to the 'hocrjs' script just before the closing 'body' tag, adding the interface to a plain hOCR file.<para>See https://github.com/kba/hocrjs for more information.</para></param>
+            public string GetHOCR(IWordExtractor wordExtractor, DocumentLayoutAnalysis.IPageSegmenter pageSegmenter, string indent = "\t", bool drawPaths = false, bool useHocrjs = false)
+            {
+                var hocr = new Export.HOcrTextExporter(wordExtractor, pageSegmenter, 2, indent);
+                return hocr.Get(page, drawPaths, useHocrjs: useHocrjs);
+            }
+
+            /// <summary>
+            /// Get the Alto (xml) string of the page layout.
+            /// <para>This is considered experimental because it needs more testing.</para>
+            /// </summary>
+            /// <param name="wordExtractor">The word extractor to use to generate words.</param>
+            /// <param name="pageSegmenter">The page segmenter to use.</param>
+            /// <param name="indent">Indent character to use.</param>
+            /// <param name="drawPaths">Draw <see cref="PdfPath"/>s present in the page.</param>
+            public string GetAltoXml(IWordExtractor wordExtractor, DocumentLayoutAnalysis.IPageSegmenter pageSegmenter, string indent = "\t", bool drawPaths = false)
+            {
+                var alto = new Export.AltoXmlTextExporter(wordExtractor, pageSegmenter, 2, indent);
+                return alto.Get(page, drawPaths);
+            }
+
+            /// <summary>
+            /// Get the PageXml (xml) string of the page layout.
+            /// <para>This is considered experimental because it needs more testing.</para>
+            /// </summary>
+            /// <param name="wordExtractor">The word extractor to use to generate words.</param>
+            /// <param name="pageSegmenter">The page segmenter to use.</param>
+            /// <param name="indent">Indent character to use.</param>
+            /// <param name="drawPaths">Draw <see cref="PdfPath"/>s present in the page.</param>
+            public string GetPageXml(IWordExtractor wordExtractor, DocumentLayoutAnalysis.IPageSegmenter pageSegmenter, string indent = "\t", bool drawPaths = false)
+            {
+                var pageXml = new Export.PageXmlTextExporter(wordExtractor, pageSegmenter, 2, indent);
+                return pageXml.Get(page, drawPaths);
+            }
         }
     }
 }

src/UglyToad.PdfPig/DocumentLayoutAnalysis/DefaultPageSegmenter.cs

@@ -0,0 +1,29 @@
+using System.Collections.Generic;
+using System.Linq;
+using UglyToad.PdfPig.Content;
+using UglyToad.PdfPig.Util;
+
+namespace UglyToad.PdfPig.DocumentLayoutAnalysis
+{
+    /// <summary>
+    /// Default Page Segmenter. All words are included in one block.
+    /// </summary>
+    public class DefaultPageSegmenter : IPageSegmenter
+    {
+        /// <summary>
+        /// Create an instance of default page segmenter, <see cref="DefaultPageSegmenter"/>.
+        /// </summary>
+        public static DefaultPageSegmenter Instance { get; } = new DefaultPageSegmenter();
+
+        /// <summary>
+        /// Get the blocks.
+        /// </summary>
+        /// <param name="pageWords">The words in the page.</param>
+        public IReadOnlyList<TextBlock> GetBlocks(IEnumerable<Word> pageWords)
+        {
+            if (pageWords.Count() == 0) return EmptyArray<TextBlock>.Instance;
+
+            return new List<TextBlock>() { new TextBlock(new XYLeaf(pageWords).GetLines()) };
+        }
+    }
+}

src/UglyToad.PdfPig/Export/HOcrTextExporter.cs

@@ -0,0 +1,345 @@
+using System;
+using System.Linq;
+using UglyToad.PdfPig.Content;
+using UglyToad.PdfPig.DocumentLayoutAnalysis;
+using UglyToad.PdfPig.Geometry;
+using UglyToad.PdfPig.Util;
+
+namespace UglyToad.PdfPig.Export
+{
+    /// <summary>
+    /// hOCR v1.2 (HTML) text exporter.
+    /// <para>See http://kba.cloud/hocr-spec/1.2/ </para>
+    /// </summary>
+    internal class HOcrTextExporter : ITextExporter
+    {
+        private const string xmlHeader = "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n<!DOCTYPE html PUBLIC \"-//W3C//DTD XHTML 1.0 Transitional//EN\"\n\t\"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd\">\n";
+        private const string hocrjs = "<script src='https://unpkg.com/hocrjs'></script>\n";
+
+        private IPageSegmenter pageSegmenter;
+        private IWordExtractor wordExtractor;
+
+        private decimal scale;
+        private string indentChar;
+
+        private int pageCount = 0;
+        private int areaCount = 0;
+        private int lineCount = 0;
+        private int wordCount = 0;
+        private int pathCount = 0;
+        private int paraCount = 0;
+
+        /// <summary>
+        /// hOCR v1.2 (HTML)
+        /// <para>See http://kba.cloud/hocr-spec/1.2/ </para>
+        /// </summary>
+        /// <param name="wordExtractor"></param>
+        /// <param name="pageSegmenter"></param>
+        /// <param name="scale"></param>
+        /// <param name="indent">Indent character.</param>
+        public HOcrTextExporter(IWordExtractor wordExtractor, IPageSegmenter pageSegmenter, double scale = 1.0, string indent = "\t")
+        {
+            this.wordExtractor = wordExtractor;
+            this.pageSegmenter = pageSegmenter;
+            this.scale = (decimal)scale;
+            this.indentChar = indent;
+        }
+
+        /// <summary>
+        /// Get the hORC (HTML) string of the page layout.
+        /// </summary>
+        /// <param name="document">The document.</param>
+        /// <param name="includePaths">Draw <see cref="PdfPath"/>s present in the page.</param>
+        /// <param name="useHocrjs">Will add a reference to the 'hocrjs' script just before the closing 'body' tag, adding the 
+        /// interface to a plain hOCR file.<para>See https://github.com/kba/hocrjs for more information.</para></param>
+        public string Get(PdfDocument document, bool includePaths = false, bool useHocrjs = false)
+        {
+            string hocr = GetHead() + indentChar + "<body>\n";
+
+            for (var i = 0; i < document.NumberOfPages; i++)
+            {
+                var page = document.GetPage(i + 1);
+                hocr += GetCode(page, includePaths) + "\n";
+            }
+
+            if (useHocrjs) hocr += indentChar + indentChar + hocrjs;
+            hocr += indentChar + "</body>";
+            hocr = xmlHeader + AddHtmlHeader(hocr);
+            return hocr;
+        }
+
+        /// <summary>
+        /// Get the hORC (HTML) string of the page layout. Excludes <see cref="PdfPath"/>s.
+        /// </summary>
+        /// <param name="page">The page.</param>
+        /// <returns></returns>
+        public string Get(Page page)
+        {
+            return Get(page, false);
+        }
+
+        /// <summary>
+        /// Get the hORC (HTML) string of the page layout.
+        /// </summary>
+        /// <param name="page">The page.</param>
+        /// <param name="imageName">The image name, if any.</param>
+        /// <param name="includePaths">Draw <see cref="PdfPath"/>s present in the page.</param>
+        /// <param name="useHocrjs">Will add a reference to the 'hocrjs' script just before the closing 'body' tag, adding the interface to a plain hOCR file.<para>See https://github.com/kba/hocrjs for more information.</para></param>
+        public string Get(Page page, bool includePaths = false, string imageName = "unknown", bool useHocrjs = false)
+        {
+            string hocr = GetHead() + indentChar + "<body>\n";
+
+            hocr += GetCode(page, includePaths, imageName) + "\n";
+
+            if (useHocrjs) hocr += indentChar + indentChar + hocrjs;
+            hocr += indentChar + "</body>";
+            hocr = xmlHeader + AddHtmlHeader(hocr);
+            return hocr;
+        }
+
+        private string GetHead()
+        {
+            return indentChar + "<head>" +
+                "\n" + indentChar + indentChar + "<title></title>" +
+                "\n" + indentChar + indentChar + "<meta http-equiv='Content-Type' content='text/html;charset=utf-8' />" +
+                "\n" + indentChar + indentChar + "<meta name='ocr-system' content='" + pageSegmenter.GetType().Name + "|" + wordExtractor.GetType().Name + "' />" +
+                "\n" + indentChar + indentChar + "<meta name='ocr-capabilities' content='ocr_page ocr_carea ocr_par ocr_line ocrx_word ocr_linedrawing' />" +
+                "\n" + indentChar + "</head>\n";
+        }
+
+        private string AddHtmlHeader(string content)
+        {
+            return "<html xmlns=\"http://www.w3.org/1999/xhtml\" xml:lang=\"en\" lang=\"en\">\n" + content + "\n</html>";
+        }
+
+        /// <summary>
+        /// Get indent string from level.
+        /// </summary>
+        /// <param name="level">The indent level.</param>
+        /// <returns></returns>
+        private string GetIndent(int level)
+        {
+            string indent = "";
+            for (int i = 0; i < level; i++)
+            {
+                indent += indentChar;
+            }
+            return indent;
+        }
+
+        /// <summary>
+        /// Get the hORC string for the page.
+        /// <para>http://kba.cloud/hocr-spec/1.2/#elementdef-ocr_page</para>
+        /// </summary>
+        /// <param name="page"></param>
+        /// <param name="imageName"></param>
+        /// <param name="includePaths">Draw <see cref="PdfPath"/>s present in the page.</param>
+        private string GetCode(Page page, bool includePaths, string imageName = "unknown")
+        {
+            pageCount++;
+            int level = 2;
+
+            string hocr = GetIndent(level) + @"<div class='ocr_page' id='page_" + page.Number.ToString() +
+                "' title='image \"" + imageName + "\"; bbox 0 0 " +
+                (int)Math.Round(page.Width * scale) + " " + (int)Math.Round(page.Height * scale) +
+                "; ppageno " + (page.Number - 1) + "\'>";
+
+            if (includePaths)
+            {
+                foreach (var path in page.ExperimentalAccess.Paths)
+                {
+                    hocr += "\n" + GetCode(path, page.Height, true, level + 1);
+                }
+            }
+
+            var words = page.GetWords(wordExtractor);
+
+            if (words.Count() > 0)
+            {
+                var blocks = pageSegmenter.GetBlocks(words);
+                foreach (var block in blocks)
+                {
+                    hocr += "\n" + GetCodeArea(block, page.Height, level + 1);
+                }
+            }
+
+            hocr += "\n" + GetIndent(level) + @"</div>";
+            return hocr;
+        }
+
+        /// <summary>
+        /// Get the hORC string for the path.
+        /// <para>http://kba.cloud/hocr-spec/1.2/#elementdef-ocr_linedrawing</para>
+        /// </summary>
+        /// <param name="path"></param>
+        /// <param name="pageHeight"></param>
+        /// <param name="subPaths"></param>
+        /// <param name="level">The indent level.</param>
+        /// <returns></returns>
+        private string GetCode(PdfPath path, decimal pageHeight, bool subPaths, int level)
+        {
+            if (path == null) return string.Empty;
+
+            string hocr = string.Empty;
+
+            if (subPaths)
+            {
+                var bbox = path.GetBoundingRectangle();
+                if (bbox.HasValue)
+                {
+                    areaCount++;
+                    hocr += GetIndent(level) + @"<div class='ocr_carea' id='block_" + pageCount + "_"
+                        + areaCount + "' title='" + GetCode(bbox.Value, pageHeight) + "'>\n";
+                    foreach (var subPath in path.Commands)
+                    {
+                        var subBbox = subPath.GetBoundingRectangle();
+                        if (subBbox.HasValue)
+                        {
+                            pathCount++;
+                            hocr += GetIndent(level + 1) + @"<span class='ocr_linedrawing' id='drawing_" + pageCount + "_"
+                                + pathCount + "' title='" + GetCode(subBbox.Value, pageHeight) + "' />\n";
+                        }
+                    }
+                    hocr += GetIndent(level) + @"</div>";
+                }
+            }
+            else
+            {
+                var bbox = path.GetBoundingRectangle();
+                if (bbox.HasValue)
+                {
+                    pathCount++;
+                    hocr += GetIndent(level) + @"<span class='ocr_linedrawing' id='drawing_" + pageCount + "_"
+                            + pathCount + "' title='" + GetCode(bbox.Value, pageHeight) + "' />";
+                }
+
+            }
+
+            return hocr;
+        }
+
+        /// <summary>
+        /// Get the hORC string for the area.
+        /// <para>http://kba.cloud/hocr-spec/1.2/#elementdef-ocr_carea</para>
+        /// </summary>
+        /// <param name="block">The text area.</param>
+        /// <param name="pageHeight"></param>
+        /// <param name="level">The indent level.</param>
+        private string GetCodeArea(TextBlock block, decimal pageHeight, int level)
+        {
+            areaCount++;
+
+            string bbox = GetCode(block.BoundingBox, pageHeight);
+            string hocr = GetIndent(level) + @"<div class='ocr_carea' id='block_" + pageCount + "_"
+                + areaCount + "' title='" + GetCode(block.BoundingBox, pageHeight) + "'>";
+
+            hocr += GetCodeParagraph(block, pageHeight, level + 1); // we concider 1 area = 1 block. should change in the future
+            hocr += "\n" + GetIndent(level) + @"</div>";
+            return hocr;
+        }
+
+        /// <summary>
+        /// Get the hORC string for the paragraph.
+        /// <para>See http://kba.cloud/hocr-spec/1.2/#elementdef-ocr_par</para>
+        /// </summary>
+        /// <param name="block">The paragraph.</param>
+        /// <param name="pageHeight"></param>
+        /// <param name="level">The indent level.</param>
+        /// <returns></returns>
+        private string GetCodeParagraph(TextBlock block, decimal pageHeight, int level)
+        {
+            paraCount++;
+            string hocr = "\n" + GetIndent(level) + @"<p class='ocr_par' id='par_" + pageCount + "_"
+                + paraCount + "' title='" + GetCode(block.BoundingBox, pageHeight) + "'>"; // lang='eng'
+
+            foreach (var line in block.TextLines)
+            {
+                hocr += "\n" + GetCode(line, pageHeight, level + 1);
+            }
+            hocr += "\n" + GetIndent(level) + @"</p>";
+
+            return hocr;
+        }
+
+        /// <summary>
+        /// Get the hORC string for the text line.
+        /// <para>See http://kba.cloud/hocr-spec/1.2/#elementdef-ocr_line</para>
+        /// </summary>
+        /// <param name="line"></param>
+        /// <param name="pageHeight"></param>
+        /// <param name="level">The indent level.</param>
+        private string GetCode(TextLine line, decimal pageHeight, int level)
+        {
+            lineCount++;
+            double angle = 0;
+
+            // http://kba.cloud/hocr-spec/1.2/#propdef-baseline
+            // below will be 0 as long as the word's bounding box bottom is the BaseLine and not 'Bottom'
+            double baseLine = (double)line.Words[0].Letters[0].StartBaseLine.Y;
+            baseLine = (double)line.BoundingBox.Bottom - baseLine;
+
+            string hocr = GetIndent(level) + @"<span class='ocr_line' id='line_" + pageCount + "_" + lineCount + "' title='" +
+                GetCode(line.BoundingBox, pageHeight) + "; baseline " + angle + " 0'>"; //"; x_size 42; x_descenders 5; x_ascenders 12' >";
+
+            foreach (var word in line.Words)
+            {
+                hocr += "\n" + GetCode(word, pageHeight, level + 1);
+            }
+            hocr += "\n" + GetIndent(level) + @"</span>";
+            return hocr;
+        }
+
+        /// <summary>
+        /// Get the hORC string for the word.
+        /// <para>See http://kba.cloud/hocr-spec/1.2/#elementdef-ocrx_word</para>
+        /// </summary>
+        /// <param name="word"></param>
+        /// <param name="pageHeight"></param>
+        /// <param name="level">The indent level.</param>
+        private string GetCode(Word word, decimal pageHeight, int level)
+        {
+            wordCount++;
+            string hocr = GetIndent(level) +
+                @"<span class='ocrx_word' id='word_" + pageCount + "_" + wordCount +
+                "' title='" + GetCode(word.BoundingBox, pageHeight) + "; x_wconf " + GetConfidence(word);
+
+            hocr += "; x_font " + word.FontName;
+
+            if (word.Letters.Count > 0 && word.Letters[0].FontSize != 1)
+            {
+                hocr += "; x_fsize " + word.Letters[0].FontSize;
+            }
+            hocr += "'";
+
+            hocr += ">" + word.Text + "</span> ";
+            return hocr;
+        }
+
+        private int GetConfidence(Word word)
+        {
+            return 100;
+        }
+
+        /// <summary>
+        /// Get the hORC string for the bounding box.
+        /// <para>See http://kba.cloud/hocr-spec/1.2/#propdef-bbox</para>
+        /// </summary>
+        /// <param name="rectangle"></param>
+        /// <param name="pageHeight"></param>
+        private string GetCode(PdfRectangle rectangle, decimal pageHeight)
+        {
+            // the values are with reference to the the top-left 
+            // corner of the document image and measured in pixels
+
+            var left = (int)Math.Round(rectangle.Left * scale);
+            var top = (int)Math.Round((pageHeight - rectangle.Top) * scale);
+            var right = (int)Math.Round(rectangle.Right * scale);
+            var bottom = (int)Math.Round((pageHeight - rectangle.Bottom) * scale);
+
+            return @"bbox " + (left > 0 ? left : 0) + " "
+                            + (top > 0 ? top : 0) + " "
+                            + (right > 0 ? right : 0) + " "
+                            + (bottom > 0 ? bottom : 0);
+        }
+    }
+}

src/UglyToad.PdfPig/Export/ITextExporter.cs

@@ -0,0 +1,17 @@
+using UglyToad.PdfPig.Content;
+
+namespace UglyToad.PdfPig.Export
+{
+    /// <summary>
+    /// Exports the page's text into the desired format.
+    /// </summary>
+    public interface ITextExporter
+    {
+        /// <summary>
+        /// Get the text representation.
+        /// </summary>
+        /// <param name="page"></param>
+        /// <returns></returns>
+        string Get(Page page);
+    }
+}

src/UglyToad.PdfPig/Geometry/PdfRectangle.cs

@@ -33,7 +33,7 @@
         /// <summary>
         /// Centroid point of the rectangle.
         /// </summary>
-        public PdfPoint Centroid => new PdfPoint(TopLeft.X + (TopLeft.X - TopLeft.X) / 2, BottomLeft.Y + (TopLeft.Y - BottomLeft.Y) / 2);
+        public PdfPoint Centroid => new PdfPoint(Left + (Right - Left) / 2, Bottom + (Top - Bottom) / 2);
 
         /// <summary>
         /// Width of the rectangle.