feat(llm): use poetry to manage the dependencies

[returnToInnocence]

Background

Originally, we used pip to install modules by downloading and installing all the dependencies. pip does not manage this, but simply removes the module, leaving behind a bunch of dependencies that were installed when the module was downloaded and installed, which can cause conflicts when installing other libraries with this legacy and no longer useful library

Which file change, why?

• In file hugegraph-llm/README.md

  1. Python version we use 3.10, to prepare for the gradio5
  2. Add poetry usage

• In file hugegraph-llm/pyproject.toml

  1. Write the tool.poetry configuration according to hugegraph-python-client
     chore(hg-llm): release details synchronously
  2. use pyproject.toml instead of requierments.txt
     feature(llm): use poetry
     chore(hg-llm): Add the library apscheduler
  3. use gradio5 to avoid several critical vulnerabilities in the previous version
     chore(hg-llm): change graio5 & authors
  4. Use higher versions of nltk to avoid critical vulnerabilities in lower versions
     chore(hg-llm): update nltk & readme

• In file .licenserc.yaml

  1. chore(hg-llm): ignore lock config

References

1. https://python-poetry.org/docs/
2. https://github.com/microsoft/graphrag
3. https://python-poetry.org/docs/pyproject/
4. https://python-poetry.org/docs/dependency-specification/
5. https://pypi.org/project/gradio/

[ChenZiHong-Gavin]

we don't need requirements.txt now?

[returnToInnocence]

we don't need requirements.txt now?

Yes, instead of the requirements.txt, we'll record versions through 'tool.poetry.dependencies' and poetry.lock

[MrJs133]

Perhaps we can add the command poetry shell after setting up the environment.
Beginners may not know that they need to execute poetry shell.

[returnToInnocence]

Perhaps we can add the command poetry shell after setting up the environment. Beginners may not know that they need to execute poetry shell.

@MrJs133 Thanks for your review and tips! 🫡

💡By creating a new virtual environment and using poetry to install the dependencies, instead of using poetry shell, we run it with python3 -m hugegraph_llm.demo.rag_demo.app. poetry shell should not be necessary

[imbajin]

TODO: we need to improve the doc for poetry usage (for devs)