将每个单独单词的坐标提取到pdf文件中的TextChunk中

Question

继this actual solution之后，我试图获得TextChunk及其每个坐标（actual page，top，bottom，left，right）内的所有单词。

由于TextChunk可能是一个短语，一个单词或其他什么，我试图手动执行此操作，指望最后一个单词的矩形并每次切割它。我注意到这种手动方法可能很麻烦（我需要手动指望特殊字符等），所以我问自己ITextSharp是否提供了更简单的方法来执行此操作。

我的Chunk和LocationTextExtractionStragy继承的课程如下：

public class Chunk
{
    public Guid Id { get; set; }
    public Rectangle Rect { get; set; }
    public TextRenderInfo Render { get; set; }
    public BaseFont BF { get; set; }
    public string Text { get; set; }
    public int FontSize { get; set; }


    public Chunk(Rectangle rect, TextRenderInfo renderInfo)
    {
        this.Rect = rect;
        this.Render = renderInfo;
        this.Text = Render.GetText();
        Initialize();
    }


    public Chunk(Rectangle rect, TextRenderInfo renderInfo, string text)
    {
        this.Rect = rect;
        this.Render = renderInfo;
        this.Text = text;
        Initialize();
    }


    private void Initialize()
    {
        this.Id = Guid.NewGuid();
        this.BF = Render.GetFont();
        this.FontSize = ObtainFontSize();
    }

    private int ObtainFontSize()
    {
        return Convert.ToInt32(this.Render.GetSingleSpaceWidth() * 12 / this.BF.GetWidthPoint(" ", 12));
    }
}

public class LocationTextExtractionPersonalizada : LocationTextExtractionStrategy
{
    //Save each coordinate
    public List<Chunk> ChunksInPage = new List<Chunk>();

    //Automatically called on each chunk on PDF
    public override void RenderText(TextRenderInfo renderInfo)
    {
        base.RenderText(renderInfo);
        if (string.IsNullOrWhiteSpace(renderInfo.GetText())
                || renderInfo == null)
                return;

        //Get chunk Vectors
        var bottomLeft = renderInfo.GetDescentLine().GetStartPoint();
        var topRight = renderInfo.GetAscentLine().GetEndPoint();

        //Create Rectangle based on previous Vectors
        var rect = new Rectangle(
                           bottomLeft[Vector.I1],
                           bottomLeft[Vector.I2],
                           topRight[Vector.I1],
                           topRight[Vector.I2]);

        if (rect == null)
                return;

        //Add each chunk with its coordinates
        ChunksInPage.Add(new Chunk(rect, renderInfo));
    }
}

所以一旦我得到文件等等，我就这样继续：

private void ProcessContent()
{
    for (int page= 1; page <= pdfReader.NumberOfPages; page++)
    {
        var strategy = new LocationTextExtractionPersonalizada();

        var currentPageText = PdfTextExtractor.GetTextFromPage(
                                          pdfReader,
                                          page,
                                          strategy);

        //Here is where I want to get each word with its coordinates
        var chunksWords= ChunkRawToWord(strategy.ChunksInPage);
    }
}

private List<Chunk> ChunkRawToWord(IList<Chunk> chunks)
{
    if (chunks == null || chunks[0] == null)
            return null;

    var words = new List<Chunk>();
    //Poor RegEx pattern to get the word and its wathever
    string pattern = @"[@&\w+]*(-*\/*\s*\:*\;*\,*\.*\(*\)*\%*\>*\<*)?";

    var something = chunks[0].Render.GetCharacterRenderInfos();

    for (int i = 0; i < chunks.Count; i++)
    {
        var wordsInChunk = Regex.Matches(
                                          chunks[i].Text,
                                          pattern,
                                          RegexOptions.IgnoreCase);


        var rectangleChunk = new Rectangle(chunks[i].Rect);
        for (int j = 0; j < wordsInChunk.Count; j++)
        {
            if (string.IsNullOrWhiteSpace(wordsInChunk[j].Value))
                continue;

        var word = new Chunk(
                                   rectangleChunk, 
                                   chunks[i].Render, 
                                   wordsInChunk[j].ToString());

            if (j == 0)
            {
                word.Rect.Right = word.BF.GetWidthPoint(word.Text, word.FontSize);
                    words.Add(word);
                    continue;
            }

            if (words.Count <= 0)
                continue;

            word.Rect.Left = words[j - 1].Rect.Right;
            word.Rect.Right = words[j - 1].Rect.Right + word.BF.GetWidthPoint(word.Text, word.FontSize);
            words.Add(word);
        }
    }

    return words;
}

之后，我写了一篇关于Mkl解决方案的评论，回复了“使用getCharacterRenderInfos（）”，我使用它，我将每个字符都放入TextRenderInfo的列表中。

对不起，我开始混合概念，找出如何应用该解决方案的方法，并让我大吃一惊。

我真的很感激这里的一只手。

Answer 1

您可以使用TextRenderInfo.GetCharacterRenderInfos()方法为块中的每个字符集获取TextRenderInfo的集合。然后，您可以将单个字符重组为单词，并使用该单词中第一个和最后一个TextRenderInfo的坐标计算包含该单词的矩形。

在您的自定义文本提取策略中：

 var _separators = new[] { "-", "(", ")", "/", " ", ":", ";", ",", "."};
 protected virtual void ParseRenderInfo(TextRenderInfo currentInfo)
    {
        var resultInfo = new List<TextRenderInfo>();
        var chars = currentInfo.GetCharacterRenderInfos();

        foreach (var charRenderInfo in chars)
        {
            resultInfo.Add(charRenderInfo);
            var currentChar = charRenderInfo.GetText();
            if (_separators.Contains(currentChar))
            {
                ProcessWord(currentInfo, resultInfo);
                resultInfo.Clear();
            }
        }
        ProcessWord(currentInfo, resultInfo);
    }
 private void ProcessWord(TextRenderInfo charChunk, List<TextRenderInfo> wordChunks)
    {
        var firstRender = wordChunks.FirstOrDefault();
        var lastRender = wordChunks.LastOrDefault();
        if (firstRender == null || lastRender == null)
        {
            return;
        }
        var startCoords = firstRender.GetDescentLine().GetStartPoint();
        var endCoords = lastRender.GetAscentLine().GetEndPoint();
        var wordText = string.Join("", wordChunks.Select(x => x.GetText()));
        var wordLocation = new LocationTextExtractionStrategy.TextChunkLocationDefaultImp(startCoords, endCoords, charChunk.GetSingleSpaceWidth());
        _chunks.Add(new CustomTextChunk(wordText, wordLocation));
    }