Skip to content

docs(ai): align configuration, APIs and installation workflows - #572

Open
imbajin wants to merge 8 commits into
apache:masterfrom
hugegraph:docs/ai-workflows-refresh-20261115
Open

imbajin wants to merge 8 commits into
apache:masterfrom
hugegraph:docs/ai-workflows-refresh-20261115

Conversation

@imbajin

@imbajin imbajin commented Nov 16, 2026

Copy link
Copy Markdown
Member

Purpose of the PR

Update the bilingual AI guides so deployment, saved configuration, API calls and Python installation agree with current source and available packages.

  • Use the unified Helm entry and build the source-workspace Compose image before starting it.
  • Explain settings persistence and priority, per-role save/test/inheritance, credential retention and the actual REST/write-gate behavior.
  • Remove invalid workspace extras and distinguish source Python requirements from the published client packages.

Validation: source/image contract review, JSON/curl checks, dependency-resolution dry runs and strict artifact checks passed. Desktop/mobile checks covered the task entry, configuration tables and native reader controls. All eight configuration-table screenshot states passed with the current OpenAI role table. No live model, GPU or ML training acceptance is claimed.

Based on Helm #68, which provides the unified RAG deployment section. Existing links and bilingual heading compatibility are preserved.

Summary by CodeRabbit

  • 文档
    • 更新 HugeGraph-AI 1.8 的配置、部署、模型连接与 REST API 说明,涵盖配置优先级、持久化、密钥管理及服务部署方式。
    • 更新 ML 与 Python 客户端的环境要求和安装指南;区分源码与已发布客户端的 Python 版本要求,并固定稳定版安装示例。
    • 调整快速入门导航与配置面板说明,补充相关章节链接。
  • 测试
    • 改进配置目录表格的描述列识别,并更新配置筛选项。

@bitflicker64 bitflicker64 left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Blocking: no. Summary: The saved-connection precedence, path selection, key retention, role inheritance, test and use-chat routes, client_config handling and Thin API write gate match hugegraph-llm on apache/hugegraph-ai main, as do the extras, Compose image and ML/Vermeer packaging. One status-code statement is incomplete, in both languages.

@bitflicker64 bitflicker64 left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Blocking: no. Summary: The AI guides' saved-connection precedence, file paths, role inheritance, key retention, configuration and Thin API status codes, extras, Compose image and ML/Vermeer packaging match hugegraph-ai main, and the 422/400 split raised on the earlier head is fixed in both languages. One comment covers Helm text stacked from #574 that lacks its follow-up fixes.

@imbajin imbajin left a comment

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Score: 7.5/10 · Confidence: high

Blocking: yes

Summary

The existing Helm review comment remains applicable: the documented default shard replication condition can mislead operators when initial Store count and bootstrap Store replicas differ. No additional issue was verified.

Evidence

Compared the selected English and Chinese guide with current Server Helm _helpers.tpl and the existing inline discussion; inspected AI configuration and API changes against source.

Validation notes

Source-based review; no live Helm deployment, model connection or end-to-end application test was run.

@imbajin
imbajin force-pushed the docs/ai-workflows-refresh-20261115 branch from f16bd21 to 49acb6b Compare November 19, 2026 11:57

@bitflicker64 bitflicker64 left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Blocking: no. Summary: The AI guides are unchanged since the last reviewed head and still match hugegraph-ai main, and the rebase onto #574 resolves the earlier Helm comment, including the bootstrap.storeReplicas shard rule. One link in the stacked Chinese Helm text points to a missing anchor.


在 `values-cluster.yaml`(`store.updateStrategy.type=OnDelete`)预设下,`helm upgrade` 不会自动重建 Store Pod。
即使镜像名称匹配,旧 Pod 也并未以 `Always` 策略重新挂载 PVC。
请按[受控滚动升级流程](./hugegraph-helm-operations.md#6-安全逐个滚动更新-store-镜像)逐个删除并重建 Store Pod。

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Minor: #6-安全逐个滚动更新-store-镜像 matches no id on the rendered operations page, whose heading sets {#6-rolling-store-images-safely}, so the link opens the page top instead of the rollout procedure. Use #6-rolling-store-images-safely, as the English page does, in #574 and rebase.

@imbajin imbajin left a comment

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Score: 8.0/10 · Confidence: high

Blocking: no

Summary

The existing Chinese Store rollout link issue remains: the link lands at the operations page top instead of the controlled rollout procedure. No additional issue was verified.

Evidence

Compared the selected guide's fragment with the explicit ID in the target heading and reviewed the existing inline discussion. The earlier bootstrap replica-count issue is corrected on this head.

Validation notes

Source-based review; no rendered-site browser check, live Helm deployment, model connection or end-to-end application test was run.

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants