java分离pdf的段落内容和表格内容

提示：文章写完后，目录可以自动生成，如何生成可参考右边的帮助文档

---

前言

PDF中提取全部文本信息比较容易，但是表格内容和文字一般是混在一起的，需要拆分开。

---

一、依赖？

        <dependency>
            <groupId>org.apache.pdfbox</groupId>
            <artifactId>pdfbox</artifactId>
            <version>3.0.0-alpha2</version>
        </dependency>
        <dependency>
            <groupId>technology.tabula</groupId>
            <artifactId>tabula</artifactId>
            <version>1.0.3</version>
        </dependency>

二、代码示例

代码如下（示例）：

private static void readTextAndTable(String path) throws IOException {
        FileOutputStream fileOutputStream;
        FileOutputStream fileOutputStream1;
        fileOutputStream = new FileOutputStream("text.txt");
        fileOutputStream1 = new FileOutputStream("table.txt");

        PDDocument pd = Loader.loadPDF(new File(path));
        ObjectExtractor oe = new ObjectExtractor(pd);
        SpreadsheetExtractionAlgorithm sea = new SpreadsheetExtractionAlgorithm(); // Tabula algo.
        int pageSize = pd.getNumberOfPages();
        // 一页一页读取
        for (int i = 0; i < pageSize; i++) {
            System.out.println("第" + (i + 1) + "页");
            Page page = oe.extract(i + 1);
            //抽取出所有的表格内容
            List<Table> table = sea.extract(page);
            List<double[]> loc = new ArrayList<>();
            for (Table table1 : table) {
                loc.add(new double[]{table1.y, table1.y + table1.height});
                for (List<RectangularTextContainer> row : table1.getRows()) {
                    fileOutputStream1.write((row.stream().map(RectangularTextContainer::getText).map(o -> o.replace("\r", "")).collect(Collectors.joining(" | ")) + "\n").getBytes("utf-8"));
                }
                fileOutputStream1.write("\n".getBytes("utf-8"));
            }
            //抽取出所有的文本内容
            MyPDFTextStripper stripper = new MyPDFTextStripper();
            //设置按顺序输出
            stripper.setSortByPosition(true);
            stripper.setStartPage(i + 1);
            stripper.setEndPage(i + 1);
            stripper.setParagraphStart("\t");
            stripper.setParagraphEnd(" ");
            stripper.getText(pd);
            double lastY = 0;
            List<List<TextPosition>> charactersByArticle = stripper.myGetCharactersByArticle();
            StringBuilder pagraph = new StringBuilder();
            outer:
            for (int m = 0; m < charactersByArticle.get(0).size(); m++) {
                TextPosition word = charactersByArticle.get(0).get(m);
                float endY = word.getY();
                // 如果文本的y在表格的上下边框的内部，则不输出继续遍历文本
                for (double[] l : loc) {
                    if (l[0] < endY && l[1] > endY) {
                        continue outer;
                    }
                }
                if (!word.getUnicode().equals(" ")) {
                    pagraph.append(word.getUnicode());
                } else {
                    if (lastY == endY) {
                        pagraph.append(word.getUnicode());
                    } else {
                        if (!pagraph.toString().equals(" ")) {
                            pagraph.append("\n");
                            fileOutputStream.write(pagraph.toString().getBytes("utf-8"));
                        }
                        pagraph = new StringBuilder();
                        pagraph.append(word.getUnicode());
                        lastY = endY;
                    }
                }
            }
        }
        fileOutputStream.flush();
        fileOutputStream.close();
        fileOutputStream1.flush();
        fileOutputStream1.close();
    }

---

总结

以上就是今天要讲的内容，本文仅仅简单介绍了如何拆分文字内容和表格内容，输出的文本内容还需要优化调整。