后端架构概览

.
├── __init__.py
├── about
│   ├── __init__.py
│   └── update_about.py
├── courses
│   ├── __init__.py
│   ├── build_course_pages.py
│   ├── fetch_opened_prs_and_issues.py
│   ├── gen_links.py
│   ├── gen_repo_update_time.py
│   └── wrap_badges.py
├── filetrees
│   ├── __init__.py
│   ├── filetree_manager.py
│   └── filetrees.json
├── infos
│   ├── netdisk_notice.txt
│   ├── notice.txt
│   └── sponsor.txt
├── news
│   ├── __init__.py
│   ├── gen_image.py
│   └── gen_news.py
├── requirements.txt
└── workflows
    ├── build_course_pages.sh
    ├── build_directory_pages.sh
    ├── build_semester_pages.sh
    ├── build_single_course_page.sh
    ├── check_before_build_single_course_page.sh
    └── fetch_repos.sh

jobs:
  build-documentation:
    runs-on: ubuntu-22.04
    permissions:
      contents: write
      id-token: write
    env:
      PYTHONPATH: /home/runner/work/hoa-moe/hoa-moe   
    steps:
      - name: Setup Python
        uses: actions/setup-python@v5
        with:
          python-version: "3.11"

      - name: Install Python dependencies
        run: |
          python -m pip install --upgrade pip
          pip install -r scripts/requirements.txt          

      # 获取 OpenAuto 组织下仓库名等步骤已省去，具体内容可自行查看 .github/workflow/course.yaml
      
      - name: Build course pages
        run: python scripts/courses/build_course_pages.py "HITSZ-OpenAuto" ${{ secrets.PERSONAL_ACCESS_TOKEN }}

if __name__ == "__main__":
    # 1. 设置命令行参数
    #    命令行内使用 python scripts/courses/build_course_pages.py HITSZ-OpenAuto token 调用
    parser = ArgumentParser(description="Generate course pages from GitHub repositories.")
    parser.add_argument("owner", help="GitHub repository owner", default="HITSZ-OpenAuto")
    parser.add_argument("token", help="GitHub token")

    # 2. 从环境变量获取仓库列表
    #    查看 course.yaml 可以发现，在之前步骤中已经把可用仓库名设置成了 repos_array 的环境变量
    repos_json = os.environ.get("repos_array")
    repos = json.loads(repos_json) if repos_json else []

    # 3. 记录执行时间并运行主要处理函数
    #    进入下一个函数 -> process_multiple_repos
    start_time = time.perf_counter()
    asyncio.run(process_multiple_repos(args.owner, repos, args.token))
    end_time = time.perf_counter()
    execution_time = end_time - start_time

async def process_multiple_repos(owner: str, repos: list, token: str) -> None:
    # 1. 过滤掉特定仓库（.github, hoa-moe, HITSZ-OpenAuto）
    repos = [repo for repo in repos if repo not in [".github", "hoa-moe", "HITSZ-OpenAuto"]]
    sorted_repos = sorted(repos)  # 排序，用于在并行的情况下保证构建网页时的顺序

    # 2. 为每个仓库创建一个 GitHubAPIClient 实例
    #    只需要知道这个类主要是用来和 GitHub 网络通信拿取有关信息就行
    clients = [GitHubAPIClient(owner, repo, token, index) for index, repo in enumerate(sorted_repos)]
    tasks: list[asyncio.Task] = []

    # 3. 使用异步方式并行处理所有仓库
    #    为每个仓库都单独创建一个执行 process_repo 函数的 task，每个 task 可以同时执行（只消耗一份时间）
    for client in clients:
        await client.init_session()
        task = asyncio.create_task(process_repo(client))
        tasks.append(task)

    await asyncio.gather(*tasks) # 等待所有 task 执行完毕

    # 4. 最后更新学期类别文件并关闭所有会话
    for client in clients:
        client.update_semester_category_file()  # 最后执行更新学期类别文件，以固定构建网页时的顺序
        await client.close_session()

async def process_repo(client: GitHubAPIClient) -> None:
    # 1. 读取仓库中的 tag.txt 文件获取课程信息（类别、学期、课程名等）
    tag_content: str = await client.fetch_file_content("tag.txt")

    # 2. 根据学期和类别信息创建或更新相应的分类文件
    category_match = re.search(r"category:\s*(.*)", tag_content)
    if category_match:
        category_raw = category_match.group(1)
        category, extra_info = category_mapping.get(category_raw.strip())
        log += f"Matched category: {category}\n"

    semesters_match = re.search(r"semester:\s*(.*)", tag_content)
    # ...
    name_match = re.search(r"name:\s*(.*)", tag_content)
    # ...

    # 一个课程可能会有多个学期
    for semester in semesters:

        # 3. 获取 README.md 的内容
        readme_content: str = await client.fetch_file_content("README.md")

        # 4. 生成包含课程信息、更新时间、README 内容等的完整课程页面
        s: str = ""
        s += "---\n"
        s += f"title: {course_name}\n"
        s += (
            "toc: true\n"
            + f'editURL: "https://github.com/{client.owner}/{client.repo}/edit/main/README.md"\n'
            + "math: true\n"
            + "---\n\n"
        )

        s += readme_content + "\n"

        # 5. 添加资料下载、网盘提示和赞助信息
        s += "## 资料下载\n\n"
        with open("scripts/infos/netdisk_notice.txt", "r", encoding="utf-8") as notice_file:
            s += notice_file.read() + "\n"
        with open(f"{client.repo}_cards.txt", "r", encoding="utf-8") as cards_file:
            s += cards_file.read() + "\n"
        with open("scripts/infos/sponsor.txt", "r", encoding="utf-8") as sponsor_file:
            s += sponsor_file.read() + "\n"

        # 6. 将生成的界面内容写入对应路径下的 Markdown 文档
        repo_md_filename = f"./content/docs/{semester_en}/{client.repo}.md"
        with open(repo_md_filename, "w", encoding="utf-8") as f:
            f.write(s)