/// <summary>
/// hOCR v1.2 (HTML)
/// <para>See http://kba.cloud/hocr-spec/1.2/ </para>
/// </summary>
public HOcrTextExporter(IWordExtractor wordExtractor, IPageSegmenter pageSegmenter, double scale = 1.0, string indent = "\t")
{
this.wordExtractor = wordExtractor;
this.pageSegmenter = pageSegmenter;
this.scale = scale;
indentChar = indent;
}
/// <summary>
/// Alto 4.1 (XML).
/// <para>See https://github.com/altoxml/schema </para>
/// </summary>
/// <param name="wordExtractor">Extractor used to identify words in the document.</param>
/// <param name="pageSegmenter">Segmenter used to split page into blocks.</param>
/// <param name="scale">Scale multiplier to apply to output document, defaults to 1.</param>
/// <param name="indent">Character to use for indentation, defaults to tab.</param>
public AltoXmlTextExporter(IWordExtractor wordExtractor, IPageSegmenter pageSegmenter, decimal scale = 1.0m, string indent = "\t")
{
this.wordExtractor = wordExtractor ?? throw new ArgumentNullException(nameof(wordExtractor));
this.pageSegmenter = pageSegmenter ?? throw new ArgumentNullException(nameof(pageSegmenter));
this.scale = scale;
indentChar = indent ?? string.Empty;
}
public hOCR(IWordExtractor wordExtractor, IPageSegmenter pageSegmenter, double scale = 1.0, string indent = "\t")
{
_wordExtractor = wordExtractor;
_pageSegmenter = pageSegmenter;
_scale = (decimal)scale;
_indent = indent;
}
/// <summary>
/// PAGE-XML 2019-07-15 (XML) text exporter.
/// <para>See https://github.com/PRImA-Research-Lab/PAGE-XML </para>
/// </summary>
/// <param name="wordExtractor"></param>
/// <param name="pageSegmenter"></param>
/// <param name="readingOrderDetector"></param>
/// <param name="scale"></param>
/// <param name="indent">Indent character.</param>
public PageXmlTextExporter(IWordExtractor wordExtractor, IPageSegmenter pageSegmenter, IReadingOrderDetector readingOrderDetector = null, double scale = 1.0, string indent = "\t")
{
this.wordExtractor = wordExtractor;
this.pageSegmenter = pageSegmenter;
this.readingOrderDetector = readingOrderDetector;
this.scale = scale;
this.indentChar = indent;
}
public void SetPageSegmenter(Type pageSegmenter)
{
if (pageSegmenter == null)
{
return;
}
this.pageSegmenter = (IPageSegmenter)Activator.CreateInstance(pageSegmenter);
}
internal PdfPageModel(Page page)
{
this.page = page;
wordExtractor = NearestNeighbourWordExtractor.Instance;
pageSegmenter = DocstrumBoundingBoxes.Instance;
}
/// <summary>
/// Get blocks that are labelled as decoration for each page in the document, using a content and a geometric similarity measure.
/// <para>Decoration blocks are blocks that contains information such as author names, publication titles, page numbers, etc.
/// They are printed repeatedly at the border of each page, usually placed inside headers or footers, but sometimes also at the
/// left or right edge of the page.</para>
/// </summary>
/// <param name="pages">The <see cref="Page"/>s in the document. All of them are needed for the algorithm to work.</param>
/// <param name="wordExtractor"></param>
/// <param name="pageSegmenter"></param>
/// <param name="minimumEditDistanceNormalised">Minimum edit distance normalised. A value of 0 means both strings are exactly equal.</param>
/// <param name="similarityThreshold">Minimum similarity score to decide wether a block is labelled as decoration or not.</param>
/// <param name="n">Number of blocks in a page to be considered when looking for decoration blocks.</param>
/// <param name="maxDegreeOfParallelism">Sets the maximum number of concurrent tasks enabled.
/// <para>A positive property value limits the number of concurrent operations to the set value.
/// If it is -1, there is no limit on the number of concurrently running operations.</para></param>
public static IReadOnlyList <IReadOnlyList <TextBlock> > Get(IReadOnlyList <Page> pages,
IWordExtractor wordExtractor, IPageSegmenter pageSegmenter, Func <string, string, double> minimumEditDistanceNormalised,
double similarityThreshold = 0.25, int n = 5, int maxDegreeOfParallelism = -1)
{
if (pages.Count < 2)
{
throw new ArgumentException("The algorithm cannot be used with a document of less than 2 pages.", nameof(pages));
}
ConcurrentDictionary <int, IReadOnlyList <TextBlock> > pagesBlocks = new ConcurrentDictionary <int, IReadOnlyList <TextBlock> >();
ParallelOptions parallelOptions = new ParallelOptions()
{
MaxDegreeOfParallelism = maxDegreeOfParallelism
};
Parallel.For(0, pages.Count(), parallelOptions, p =>
{
var words = pages[p].GetWords(wordExtractor);
var blocks = pageSegmenter.GetBlocks(words);
if (!pagesBlocks.TryAdd(p, blocks))
{
throw new ArgumentException("Cannot add element with index " + p + " in ConcurrentDictionary.");
}
});
return(Get(pagesBlocks.OrderBy(x => x.Key).Select(x => x.Value).ToList(),
minimumEditDistanceNormalised,
similarityThreshold,
n,
maxDegreeOfParallelism));
}
/// <summary>
/// Get blocks that are labelled as decoration for each page in the document, using a content and a geometric similarity measure.
/// <para>Decoration blocks are blocks that contains information such as author names, publication titles, page numbers, etc.
/// They are printed repeatedly at the border of each page, usually placed inside headers or footers, but sometimes also at the
/// left or right edge of the page.</para>
/// </summary>
/// <param name="pages">The <see cref="Page"/>s in the document. All of them are needed for the algorithm to work.</param>
/// <param name="wordExtractor"></param>
/// <param name="pageSegmenter"></param>
/// <param name="similarityThreshold">Minimum similarity score to decide wether a block is labelled as decoration or not.</param>
/// <param name="n">Number of blocks in a page to be considered when looking for decoration blocks.</param>
/// <param name="maxDegreeOfParallelism">Sets the maximum number of concurrent tasks enabled.
/// <para>A positive property value limits the number of concurrent operations to the set value.
/// If it is -1, there is no limit on the number of concurrently running operations.</para></param>
public static IReadOnlyList <IReadOnlyList <TextBlock> > Get(IReadOnlyList <Page> pages,
IWordExtractor wordExtractor, IPageSegmenter pageSegmenter,
double similarityThreshold = 0.25, int n = 5, int maxDegreeOfParallelism = -1)
{
return(Get(pages, wordExtractor, pageSegmenter, Distances.MinimumEditDistanceNormalised, similarityThreshold, n, maxDegreeOfParallelism));
}
public IEnumerable <(string Prediction, float Score, TextBlock Block)> Classify(Page page, IWordExtractor wordExtractor,
IPageSegmenter pageSegmenter, Bookmarks bookmarks = null)
{
List <DocumentBookmarkNode> bookmarksNodes = bookmarks?.GetNodes()
.Where(b => b is DocumentBookmarkNode)
.Select(b => b as DocumentBookmarkNode)
.Cast <DocumentBookmarkNode>()
.Where(b => b.PageNumber == page.Number).ToList();
var avgPageFontHeight = page.Letters.Select(l => l.GlyphRectangle.Height).Average();
var words = wordExtractor.GetWords(page.Letters);
var blocks = pageSegmenter.GetBlocks(words);
foreach (var block in blocks)
{
var letters = block.TextLines.SelectMany(li => li.Words).SelectMany(w => w.Letters);
var paths = FeatureHelper.GetPathsInside(block.BoundingBox, page.ExperimentalAccess.Paths);
var images = FeatureHelper.GetImagesInside(block.BoundingBox, page.GetImages());
var features = FeatureHelper.GetFeatures(
block, paths,
images, avgPageFontHeight,
block.BoundingBox.Area,
bookmarksNodes);
BlockFeatures blockFeatures = new BlockFeatures()
{
BlockAspectRatio = features[0],
CharsCount = features[1],
WordsCount = features[2],
LinesCount = features[3],
PctNumericChars = features[4],
PctAlphabeticalChars = features[5],
PctSymbolicChars = features[6],
PctBulletChars = features[7],
DeltaToHeight = features[8],
PathsCount = features[9],
PctBezierPaths = features[10],
PctHorPaths = features[11],
PctVertPaths = features[12],
PctOblPaths = features[13],
ImagesCount = features[14],
ImageAvgProportion = features[15],
BestNormEditDistance = features[16],
};
var result = predEngine.Predict(blockFeatures);
yield return(FeatureHelper.Categories[(int)result.Prediction], result.Score.Max(), block);
}
}
public IEnumerable <(string Prediction, float Score, TextBlock Block)> Classify(Page page, IWordExtractor wordExtractor, IPageSegmenter pageSegmenter)
{
var words = wordExtractor.GetWords(page.Letters);
var blocks = pageSegmenter.GetBlocks(words);
foreach (var block in blocks)
{
var letters = block.TextLines.SelectMany(li => li.Words).SelectMany(w => w.Letters);
var paths = FeatureHelper.GetPathsInside(block.BoundingBox, page.ExperimentalAccess.Paths);
var images = FeatureHelper.GetImagesInside(block.BoundingBox, page.GetImages());
var features = FeatureHelper.GetFeatures(page, block.BoundingBox, letters, paths, images);
BlockFeatures blockFeatures = new BlockFeatures()
{
CharsCount = features[0],
PctNumericChars = features[1],
PctAlphabeticalChars = features[2],
PctSymbolicChars = features[3],
PctBulletChars = features[4],
DeltaToHeight = features[5],
PathsCount = features[6],
PctBezierPaths = features[7],
PctHorPaths = features[8],
PctVertPaths = features[9],
PctOblPaths = features[10],
ImagesCount = features[11],
ImageAvgProportion = features[12]
};
var result = predEngine.Predict(blockFeatures);
yield return(FeatureHelper.Categories[(int)result.Prediction], result.Score.Max(), block);
}
}