feat: Enabling Natural Language Graph Queries in GraphScope with chatGPT #2648

[Suchun-sv]

What do these changes do?

[TL;DR] We add LLM support to enhance the basic graph queries, allowing users to interact with graphs more casually.

Specifically, the detailed steps can be shown as:

1. Using the Langchain to send natural language sentences to LLM (chatGPT, chatGLM, and so on) with manually designed prompts.
2. Adding the schema of graphs to form more expected results from LLM.
3. Extract Cypher statements from the results.
4. Running Cypher on the Graphs through the Cypher interactive backend.
5. Return results from the backend.

Our contribution can be summarized as：

1. Well-designed Langchain pipeline as well as prompt templates (langchain_cypher.py).
2. Exposed the query interface (query.py::query_to_cypher).

Related issue number

We are related to the issue #2648

Fixes

[welcome]

Thanks for submitting your first pull request! You are awesome! 🤗
Please make sure you have signed the CLA, as this is a mandatory check before a PR being merged.
Welcome to the GraphScope community! 🎉 You can meet the community on DingTalk or Slack.

[CLAassistant]

All committers have signed the CLA.

[codecov-commenter]

Codecov Report

Merging #3271 (0d75359) into main (4d3a83e) will decrease coverage by 0.35%.
Report is 2 commits behind head on main.
The diff coverage is 0.00%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3271      +/-   ##
==========================================
- Coverage   41.39%   41.04%   -0.35%     
==========================================
  Files         107      109       +2     
  Lines       11203    11298      +95     
==========================================
  Hits         4637     4637              
- Misses       6566     6661      +95     

Files	Coverage Δ
python/graphscope/langchain_prompt/query.py	0.00% <0.00%> (ø)
...on/graphscope/langchain_prompt/langchain_cypher.py	0.00% <0.00%> (ø)

---

Continue to review full report in Codecov by Sentry.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 1942a67...0d75359. Read the comment docs.

[longbinlai]

LGTM