registrystack / registrystack/registry-stack

Add classification review and safeguard explanation to authoring and catalog UX

オープン
#596 コメント 0 件 リアクション 0 件 担当者 0 名 GitHub で見る
area:docs area:manifest area:registryctl area:relay criticality:p3 enhancement post-1.0 triage:roadmap
主要言語
Rust
スター
2
フォーク
0
平均マージ
2時間 55分
マージ済み PR(30日)
130

説明

Tracked by #590. Depends on #592 and #593. Coordinate with editor epic #399 and structured diagnostics #396.

## Outcome

Authors, reviewers, operators, and catalog consumers can inspect field concepts, classifications, provenance, review state, and effective safeguard requirements without reading raw YAML or exposing source values.

## Requirements

- Provide a field-oriented review view showing:
- dataset, entity, and field name;
- physical type;
- primary and additional semantic concepts;
- classification facet and term;
- vocabulary/profile identity and version;
- declared, derived, or proposed provenance;
- review state;
- effective safeguard requirements and enforcement status.
- Support filtering by concept, classification, vocabulary/profile, review state, and enforcement status.
- Keep labels human-readable while retaining immutable IRIs and stable rule IDs underneath.
- Explain why a classification or control applies and which reviewed artifact supplied it.
- Use the exact compiled metadata and explanation contracts rather than reimplementing semantic or policy logic in a UI.
- Integrate authoring diagnostics with #396 and later editor clients under #399.
- Keep unreviewed suggestions visually and structurally distinct from accepted classifications.
- Do not inspect live registry values or trigger scripts, connectors, or network vocabulary lookups as a side effect of browsing metadata.
- Respect existing catalog metadata authorization and avoid widening public schema discovery accidentally.

## Acceptance criteria

- [ ] A reviewer can inspect a table-like field inventory with typed classification badges and authoritative concept sources.
- [ ] Every displayed control links to a value-free explanation and stable rule ID.
- [ ] Unreviewed suggestions cannot be mistaken for released metadata or active policy.
- [ ] Filtering and search operate only over bounded compiled metadata.
- [ ] The UI renders PublicSchema, EU SEMIC, and local vocabulary bindings without privileging one hardcoded namespace.
- [ ] Authorization tests prove restricted metadata does not become public through the new view.
- [ ] A clean authoring journey is documented and tested separately from runtime enforcement.

## Non-goals

- A general data marketplace
- Source-data profiling or sampling
- Policy activation from a browser click without the governed config workflow
- Duplicating Registry compiler or PDP logic in frontend code
- Editor-specific language-server implementation already tracked by #399

コントリビューションガイド

コントリビューションガイドを開く

調査の方向性

まず追跡対象および依存する issue #590、#592、#593、#396、#399 を読み、その後、既存のコンパイル済みメタデータと説明のコントラクトを特定します。フィールド指向のレビュー画面、範囲を限定したフィルタリング、値を含まない説明、認可のカバレッジ、そして別途ドキュメント化されたオーサリングの手順が、記載された受け入れ基準を満たした時点で作業は完了です。

索引モデルが issue の本文から書いたものです。

評価

領域
authorization, frontend, testing
issue の種類
機能追加
難易度
5/5
見積もり時間
1週間以上
活発さ
静か
明瞭さ
おおむね明確
初心者へのやさしさ
35/100

新しい issue をメールで受け取る

初心者向けの GitHub issue を短くまとめたダイジェスト。