如何使用生成器（yield）递归查找几何模型层级结构？

问题背景 #

我们有一个包含链式关联对象的几何模型文件mymodel.txt，目标是通过层级关联名称或继承类，找出父对象（如surfname1）的所有依赖子对象。注意所有对象均以分号;结尾。

模型文件示例 #

[mymodel.txt]...
 FramePoint pointname0 * / 0.0 0.000 0.000 ;
 FramePoint pointname1 * / 1.0 0.000 0.000 ;
 FramePoint pointname2 * / 10.0 0.000 10.000 ;
 FramePoint pointname3 * / 10.0 10.000 10.000 ;
 FramePoint pointname4 * / 10.0 10.000 10.000 ;
#  ... 
 FramePoint pointnameX * / 0.0 0.000 0.000 ;
# ...
 CCurve curvename1 / pointname1 pointname2 pointname3
                     pointname4 ;  
 IntBead bead1 / curvename1 *Z=0 pointname0 ;
 BSubCurve curvename2 / pointname0 ;
# BSubCurve ...
 CLoftSurf surfname1 / curvename1 curvename2 ;
# ...

现有Python代码 #

import os
import logging
import pathlib
import re
def findstr_in_file(str2find:str, inpfile, pattern:str='', G_line=None):

    G_line += G_line if G_line else ''

    file_path = pathlib.Path(inpfile)
    try:
        with file_path.open(mode="r") as file:
            for l_no, line in enumerate(file):
                for match in re.finditer(pattern + str2find, line):
                    # 排除包含三个连续数值的行（如1.0002 2.0005 4.0009）
                    if not re.findall('[0-9]+\.[0-9]+ +[0-9]+\.[0-9]+ +[0-9]+\.[0-9]+ +', line[line.rfind('/') + 1:]):
                        print(f'Line Number:{l_no}; Line:{line}')
                        # 拼接后续行直到找到结束标记;
                        while not ';' in line:
                            line += next(file)
                        yield line
                        # 提取/后的关联对象名称
                        obj_names = re.findall(r'\b[^\*|=]\w+\b', line[line.rfind('/') + 1:])
                        for objname in obj_names:
                            if objname:
                                findstr_in_file(objname, inpfile, pattern, G_line)
                        break
                    else:
                        yield line
                        break
    except OSError as error:
        logging.error(f'Reading file {file_path} failed due to: {error}')

# 对象定义规则：ObjectType ObjectName + 规格说明
# 因此搜索模式为\w+ 空格 \w+ 
G_line = ''
findstr_in_file('surfname1', 'mymodel.txt', '(\w+ +)', G_line)

需求说明 #

需实现使用**生成器（yield）**递归查找模型层级结构的功能。

内容的提问来源于stack exchange，提问作者MathArt