[Task] code-doc sync before 1.8.0

首先更新 HugeGraph、Toolchain、AI 和 Computer / Vermeer 中处于开放状态且尚未合并的 PR 清单，然后阅读链接的双语指南以及引用的入口点，例如 docker-entrypoint.sh、KafkaReader.java、GraphsAPI.java、VertexAPI.java 和 master.ini。更新指定的文档页面，验证链接和可运行示例，并在范围内的每个 PR 都有文档处理结论且所有列出的验收检查都通过后完成。

Release 1.8.0 gate

Update (2026-09-18): PR #492 was merged as 58b99a1. It completes the HugeGraph, AI, and Computer / Vermeer documentation items below. The Toolchain item and this release gate remain open.

Update (2026-09-26): PR #501 carries the main Chinese documentation synchronization against current Server, Toolchain, AI, and Computer / Vermeer mainline sources. It remains open and unmerged; final CI verification is tracked in the PR. English synchronization, benchmark documentation, and the audit of unmerged 1.8.0 release-candidate PRs are separate follow-ups. For the Kafka skipped_line item below, current Toolchain master accepts the field but does not filter Kafka messages; PR #501 clarifies this in Chinese, while the English page remains pending. This update does not complete the release gate or close this issue.

Update (2026-10-01): PR #501 was merged on 2026-09-30 as a8c6857. PR #506 now tracks the focused bilingual Toolchain follow-up below. Its checked implementation items are implemented and locally verified; the PR is non-draft but remains unmerged. This does not complete the release-candidate audit or close the release gate.

• [P1] Audit all unmerged PRs intended for the 1.8.0 release and synchronize their user-facing documentation before the release branch is cut.
  • TODO steps:
    • Enumerate open and unmerged PRs labeled or otherwise identified as 1.8.0 release candidates.
    • Check each PR for user-visible behavior, configuration, API, CLI, deployment, or compatibility changes.
    • Map every documentation-impacting PR to the affected Chinese and English pages.
    • Update/add the docs and cross-link the PR or issue when the change is merged.
    • Record the mapping and verification status in this issue.
  • Evidence: PR #492 is now merged and completes the tracked documentation-sync items below; this issue still does not track the still-unmerged PRs intended for 1.8.0.
  • Impact: Release notes and user guides can remain incomplete for changes that are already merged or are about to be released.
  • Recommendation: Keep this gate open until the audit and mapping are complete.
  • Acceptance: Every relevant unmerged 1.8.0 PR has a documented impact assessment, an owner, and a verified CN/EN documentation update or an explicit no-docs-needed decision.

HugeGraph

• [P2] Document the current Server startup timeout setting in the Docker guide.

  • TODO:
    • Add HG_SERVER_STARTUP_TIMEOUT_S, default 120, valid range 1–86400, and failure behavior to the Chinese and English guides.
    • Add Docker/Compose examples and explain the distinction from the health check.
  • Completed by PR #492, which updates both language versions.
  • Acceptance: A user can find the variable, understand its default/range/failure behavior, and run the example without guessing.

• [P2] Add existing backend config pages to the Server configuration index.

  • TODO:
    • Add HBase, HStore, and RocksDB configuration links to the Chinese and English indexes.
    • Keep the links aligned with the existing configuration-page naming and navigation structure.
  • Completed by PR #492, which updates both language versions.
  • Acceptance: The Server configuration indexes expose the existing backend configuration pages in both languages.

Toolchain

Implementation below follows hugegraph-toolchain/master (currently 1.8.0). PR #506 contains: Client/Loader examples, Hubble guide restructuring, reuse of upstream Docker Compose, and expanded workspaces/permissions with real distributed screenshots. Latest commit: 89c677c48 (review comments addressed). A checked child item means implemented/local evidence is available, not that the release gate or PR merge is complete.

Hubble

• [P1] Merge the refreshed bilingual Hubble guides and their navigation (#506).
  • Make standalone RocksDB the shared guide and HStore/PD a difference-only supplement.
  • Replace obsolete step-by-step images with nine real localized screenshots per language across the two guides.
  • Exercise anonymous standalone mode, sample/schema/query, CSV import, and asynchronous queries; record actual latest image versions.
  • Reuse apache/hugegraph/docker/docker-compose.yml, conf/hubble/standalone.properties, and the Docker README instead of maintaining another Compose.
  • Preserve the existing Hubble URLs and language switching; add a latest-only supplement route and historical fallback.
  • Expand each core workspace section with useful features, controls and verified shortcuts.
  • Verify account and GraphSpace permission boundaries against master source and authenticated UI.
  • Add real distributed overview/node screenshots and localized supporting images; keep introductory wording natural and security guidance concise.
  • Acceptance: Readers can complete the shared core flows; images match their text; distributed conditions are explicit; required checks and review pass before merge.

Client

• [P2] Merge the master-aligned Java Client examples and API guidance (#506).
  • Explain source installation for the current master version and older-client API boundaries.
  • Repair the vertex index example and schema prerequisites.
  • Compile four complete examples and two API snippets against a source-derived master Client.
  • Document the configTimeout(0) unit-conversion defect and a safe explicit timeout.
  • Track/fix the Client timeout-zero implementation separately; product-code changes are outside #506.

Loader

• [P1] Merge the corrected bilingual Loader core example (#506).

  • Align nullable properties, custom numeric IDs, endpoint columns, dates and mapping with the built-in source example.
  • Check endpoint existence, data types and the expected 8-vertex/6-edge dataset offline.
  • Correct header semantics, disabled direct mode, runtime Oracle driver placement and master/release boundaries without expanding the tutorial.
  • Make bilingual local Docker examples anonymous and bind Server access to loopback.
  • Acceptance: Schema/data/mapping agree with master; offline checks are explicit and are not presented as successful server integration.

• [P2] Remove or implement Kafka skipped_line option so docs and runtime behavior agree.

  • TODO:
    • Decide whether skipped_line is supported, deprecated, or invalid.
    • If supported, document the semantics and add/verify coverage.
    • If unsupported, remove it from the guide and add validation or migration guidance as needed.
    • Clarify the actual accepted-but-ignored behavior and preprocessing guidance in both language versions.
    • Cross-link a runtime implementation/removal PR if that behavior is changed.
  • Location: English Loader source reference and its Chinese counterpart.
  • Evidence: Current Toolchain master declares and accepts skipped_line, but KafkaReader does not apply it as a filter. The current Chinese and English guides both explain this limitation and upstream preprocessing. PR #501 is merged. Runtime removal/implementation and dedicated coverage remain separate work; do not infer those are complete from the documentation update.
  • Impact: Users may configure an option that is ignored or misunderstand the loader's behavior.
  • Acceptance: The docs describe the actual behavior, with a verified example or explicit removal/migration guidance in both languages.

Validation and delivery

• Finish the PR #506 delivery gate.
  • Link checks, Python 3.13 source tests (193 run; one aggregate-artifact-only skip), strict Hugo 0.165 Extended build, fresh latest validation and 24 search-ranking checks.
  • Independent documentation review and affected-fix re-review.
  • Push each validated milestone to the same non-draft PR.
  • Finish expanded bilingual browser/image acceptance and independent review.
  • Immediately remove containers, networks, named/anonymous volumes, images, caches and temporary image archives after each completed test round; verify residual Docker storage, especially after distributed testing.
  • Verify GitHub CI on 89c677c48: all required checks and visual evidence passed; publication is intentionally skipped for this PR.
  • Obtain maintainer approval and merge; merging/publishing is outside this task.
  • Evidence: authenticated distributed latest ran Hubble JAR 1.8.0 with Server/PD/Store 1.7.0. Verified 8 vertices/6 edges, algorithm result 4 vertices/3 edges, async result [8], and read-only-user menu restrictions. Client/Loader checks remain compilation/offline rather than server integration; distributed failover was not tested. All task Docker resources, temporary image archives and credentials were removed; final Docker storage is 0B.

AI

• [P3] Add Vermeer Python Client page to AI quick-start follow-up links.

  • TODO:
    • Add the English Vermeer Python Client link.
    • Add the Chinese Vermeer Python Client link.
  • Completed by PR #492, which updates both AI quick-start pages.
  • Acceptance: Users following either quick-start page can reach the corresponding Vermeer Python Client documentation.

• [P1] Make the Vermeer Python Client example actually wait for task completion.

  • TODO:
    • Replace the single status read with bounded polling.
    • Handle terminal success and failure states and show a clear timeout/failure message.
    • Keep the Chinese and English examples behaviorally identical.
  • Completed by PR #492, which adds bounded polling and failure handling to both examples.
  • Acceptance: A copied example waits for completion, exits on success/failure, and cannot poll forever.

Computer / Vermeer

• [P1] Update Computer quick start to current graph-space REST paths.

  • TODO:
    • Replace the read mode URL in both languages with /graphspaces/DEFAULT/graphs/hugegraph/graph_read_mode.
    • Replace the vertex query URL in both languages with /graphspaces/DEFAULT/graphs/hugegraph/graph/vertices.
    • Re-run the examples against the documented endpoint shape.
  • Completed by PR #492, which updates the Computer examples in both languages.
  • Acceptance: The documented read-mode and vertex-query examples use the current graph-space paths and are runnable.

• [P1] Align documented Vermeer master HTTP port with the runnable deployment.

  • TODO:
    • Use HTTP port 6688 as the default, or explicitly map 8688 if that remains the documented port.
    • Update the Chinese and English REST examples.
    • Update the Python client example/configuration.
    • Add a runnable Docker/curl verification example.
  • Completed by PR #492, which aligns the HTTP port at 6688, documents gRPC on 6689, and adds Docker/curl verification in both languages.
  • Acceptance: A fresh deployment and the documented REST/Python examples use the same reachable port and protocol.