Add ppmatAgent with HEA and KnowMat integration

[JinZongxiao]

No description provided.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.

---

jinzongxiao seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[Copilot]

Copilot wasn't able to review this pull request because it exceeds the maximum number of lines (20,000). Try reducing the number of changed lines and requesting a review from Copilot again.

[JinZongxiao]

Fix ppmatAgent migration smoke paths

• 将 KnowMat 的 LLM key 校验从 import 阶段后移到实际运行 LLM extraction 前，避免基础导入、CI、文档构建在没有 LLM_API_KEY 时失败。
• 修复 PaddleOCR API / MinerU / OpenAI OCR 云端后端对 PyMuPDF 的非必要依赖；只有本地 OCR 或指定页码范围时才需要 PyMuPDF。
• 修复 CLI 在已有 .md/.txt 输入配合 --force-rerun 时没有任务可执行的问题。
• 修复 HEA 示例中 tdb_files 使用当前工作目录相对路径的问题，改为使用包内默认 TDB 路径。
• 新增 ppmatAgent 迁移轻量测试，覆盖基础 import、资源加载、dotenv 校验、包内 TDB registry 和 surrogate registry。

本地已验证：

• python -m compileall -q ppmatAgent test
• pytest -q test/test_ppmat_agent_migration.py
• python -m ppmatAgent.knowmat --help
• python -m ppmatAgent.hea_crewai_agent.run_crewai --help
• PaddleOCR API OCR-only smoke test
• DeepSeek LLM extraction smoke test