bosprimigenious commented on issue #376:
URL: https://github.com/apache/hugegraph-ai/issues/376#issuecomment-6771389931

   Hi @simon824 — thanks for writing this up. The merge plan is unusually 
concrete for a v0.3 proposal, especially the three ground rules (single 
governance channel, anti-corruption layer, IDs re-assembled through the engine).
   
   Context so you can place me: I have been contributing to hugegraph-ai (#398 
integer-zero vertex IDs, #411 opt-in `/rag` trace) and to the toolchain (#894 
Go-client body logging). I would rather help land the merge than start 
something parallel.
   
   Two slices I could pick up, if either is useful:
   
   1. **Step 1 (M0) acceptance automation.** A CI check that (a) nothing under 
`server/` imports upstream `hugegraph` packages directly, and (b) the pinned 
`third_party/hugegraph-ai/` subtree plus NOTICE/licence provenance matches the 
recorded release tag. That turns two of the ground rules into something the 
build enforces instead of a manual review pass.
   2. **Step 2 (M1/M2), the `ask_graph` masked-and-cited path.** My #411 work 
is about getting graph-recall intermediates (`keywords`, `match_vids`, 
`gremlin`, `graph_result`) out of the RAG flow without a second channel, so I 
have context on where citation data can be produced and tested.
   
   Which one is more useful right now? I am happy to open a scoped child issue 
under the merge plan first, keep the PR small, and wait for a reviewer's OK 
before writing code.
   
   Drafted with AI assistance (OpenAI Codex); reviewed and owned by me.
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to