Method merge_with_text_xml

modules/xml_merger.py:527–635

合并图形XML和文字XML 这是最常用的合并场景：SAM3提取的图形 + OCR提取的文字 Args: shape_xml_path: 图形XML路径 text_xml_path: 文字XML路径 output_path: 输出路径 image_path: 原始图片路径（用于获取尺寸） Returns: 输出文件路径

(self,
                            shape_xml_path: str,
                            text_xml_path: str,
                            output_path: str,
                            image_path: str = None)

        return output_path
    
    def merge_with_text_xml(self,
                            shape_xml_path: str,
                            text_xml_path: str,
                            output_path: str,
                            image_path: str = None) -> str:
        """
        合并图形XML和文字XML
        
        这是最常用的合并场景：SAM3提取的图形 + OCR提取的文字
        
        Args:
            shape_xml_path: 图形XML路径
            text_xml_path: 文字XML路径
            output_path: 输出路径
            image_path: 原始图片路径（用于获取尺寸）
            
        Returns:
            输出文件路径
        """
        self._log("合并图形和文字XML")
        
        fragments = []
        
        # 解析图形XML
        shape_tree = ET.parse(shape_xml_path)
        shape_root = shape_tree.getroot()
        
        # 获取画布尺寸
        model = shape_root.find(".//mxGraphModel")
        canvas_width = int(model.get("pageWidth", 800))
        canvas_height = int(model.get("pageHeight", 600))
        
        # 提取图形cells并判断层级
        root_elem = shape_root.find(".//root")
        for cell in list(root_elem):
            cell_id = cell.get("id")
            if cell_id in ["0", "1"]:
                continue
            
            cell_xml = ET.tostring(cell, encoding='unicode')
            style = cell.get("style", "")
            
            # 根据style判断层级
            if "image=data:image" in style:
                layer = LayerLevel.IMAGE.value
            else:
                layer = LayerLevel.BASIC_SHAPE.value
            
            # 提取geometry信息用于排序
            geom = cell.find("mxGeometry")
            bbox = None
            if geom is not None:
                try:
                    bbox = BoundingBox(
                        x1=int(float(geom.get("x", 0))),
                        y1=int(float(geom.get("y", 0))),
                        x2=int(float(geom.get("x", 0))) + int(float(geom.get("width", 0))),
                        y2=int(float(geom.get("y", 0))) + int(float(geom.get("height", 0)))