html table 表格 转 markdown 标准表格

/**
     * table网页 表格转换为 markdown形式
     * @return
     */
    public static String convert(String html){
        if(StringUtils.isBlank(html) || !html.contains("<table")){
            return html;
        }

        Document document = Jsoup.parse(html);
        document.outputSettings().prettyPrint(false);
        Elements tables = document.select(Label.table);
        if(CollectionUtils.isEmpty(tables)){
            return html;
        }

        for(Element table : tables){
            String subTable = convertTable(table);
            table.after(subTable);
            table.remove();
        }
        return document.body().html();
    }

/**
     * 单个table网页 表格转换为 markdown形式
     * @param table
     * @return
     */
    public static String convertTable(Element table) {
        MarkdownCell[][] matrix = convert2Matrix(table);
        return printOutMatrix2MarkDown(matrix);
    }


/**
     * 转换成二维数组
     * @param table
     * @return
     */
    public static MarkdownCell[][] convert2Matrix(Element table){
        int rowsLength = getRowLength(table);
        int colsLength = getColsLength(table);
        MarkdownCell[][] matrix = new MarkdownCell[rowsLength][colsLength];

        Elements trs = table.select(Label.tr);
        for(int i = 0; i < trs.size(); ++i){
            Element tr = trs.get(i);
            Elements children = tr.children();
            if(CollectionUtils.isEmpty(children)){
                continue;
            }

            for(int j = 0 , index = 0; index < children.size() ; ){
                //已经填充过说明是合并过的单元格
                if(matrix[i][j] != null){
                    ++ j;
                    continue;
                }

                Element node = children.get(index++);
                int rowSpan = getRowspan(node);
                int colSpan = getColspan(node);
                String content = node.text();
                fillCells(matrix,i,j,rowSpan,colSpan,content);
                j +=colSpan;
            }
        }

        return matrix;
    }



/**
     * 打印单元格
     * @param matrix
     * @return
     */
    private static String printOutMatrix2MarkDown(MarkdownCell[][] matrix) {
        int rows = matrix.length;
        int cols = matrix[0].length;

        StringBuilder sb = new StringBuilder();
        for(int i = 0; i < rows; ++ i){
            sb.append("|");
            for(int j = 0 ; j < cols; ++ j){
                String content = matrix[i][j] == null?"":matrix[i][j].getContent();
                sb.append(" ").append(content)
                  .append(" ").append("|");
            }
            sb.append("\n");

            //markdown首行
            if(i == 0 ){
                sb.append("|");
                for(int j = 0 ; j < cols; ++ j){
                    sb.append(" ").append("---")
                      .append(" ").append("|");
                }
                sb.append("\n");
            }
        }
        return sb.toString();
    }

/**
     * 填充单元格
     * @param matrix
     * @param rowSpan
     * @param colSpan
     * @param content
     */
    private static void fillCells(MarkdownCell[][] matrix, int startRows, int startClos,int rowSpan, int colSpan, String content) {
        for(int i = startRows; i < startRows + rowSpan; ++i){

            for(int j = startClos; j < startClos + colSpan; ++j){

                matrix[i][j] = MarkdownCell.builder().content(content).build();

            }
        }
    }

/**
     * 获取table有多少列
     * @param table
     * @return
     */
    public static int getColsLength(Element table){
        int total = 0;
        Element tr = table.select(Label.tr).first();
        Elements headers = tr.children();

        for(Element element: headers){
            total += getColspan(element);
        }
        return total;
    }
    
   /**
     * 获取table有多少行
     * @param table
     * @return
     */
    public static int getRowLength(Element table){
        Elements trs = table.select(Label.tr);
        if(CollectionUtils.isEmpty(trs)){
            throw new RuntimeException("resolving table error : tr not found");
        }
        return trs.size();
    }


/**
     * 获取合并的列数
     * @param cell
     * @return
     */
    private static int getColspan(Element cell) {
        int colspan = 1;
        if(cell.hasAttr(attr.colspan)) {
            try {
                colspan = Integer.parseInt(cell.attr(attr.colspan));
            } catch(NumberFormatException ex) {
                // ignore invalid numbers
            }
        }
        return colspan;
    }

    /**
     * 获取合并的行数
     * @param cell
     * @return
     */
    private static int getRowspan(Element cell) {
        int rowspan = 1;
        if(cell.hasAttr(attr.rowspan)) {
            try {
                rowspan = Integer.parseInt(cell.attr(attr.rowspan));
            } catch(NumberFormatException ex) {
                // ignore invalid numbers
            }
        }
        return rowspan;
    }