feat(web): in-browser chunk.json capture tool (ADR-0026)

[PhysShell]

What

A minimal, phone-friendly capture tool on the web playground: load a tab/MIDI → pick a track → assign rights + tags + quality + reviewer → preview phrase boundaries → download <id>.chunk.json. Collect the files and fold them into the corpus with griff manifest — no in-browser persistence or assembly.

Carves a thin, download-only path out of ADR-0024's "corpus curation / persistence on web — later" (see ADR-0026). The full S8 web curation dock (IndexedDB, split/merge/rename, boundary editing, ensemble) stays deferred.

How

• Reuses griff-core, no schema fork. New #[wasm_bindgen] exports build_chunk_json / detect_boundaries_json call the same boundary + structure/gesture/complexity code as griff curate and serialize a real corpus::ChunkMeta via serde_json — byte-compatible with what griff manifest reads.
• tag_palette_json feeds the tag picker from SwancoreTag::all_variants(), so the UI's labels and indices can't drift from the enum.
• created_at/updated_at are supplied by the page, keeping the wasm a pure function of its inputs (SPEC §6 determinism).
• Capture panel (hidden until a tab loads) across index.html / app.js / style.css.

Tests / gates

• Strict TDD: red → green commits per AGENTS.md (failing tests committed before each pub fn).
• build_chunk_json output round-trips back into corpus::ChunkMeta; the rights assertions pin the non-derivable provenance datum S5 requires at curation time.
• 16/16 web tests, clippy -D warnings clean, host and wasm32-unknown-unknown release builds green.

Deploy

The capture UI reaches the live Pages site on merge to main (web.yml deploys from main).

🤖 Generated with Claude Code

https://claude.ai/code/session_01TTUbGjzD8ysnVnCJnZJE95

---

Generated by Claude Code

Summary by CodeRabbit

• New Features
  • Added a “Capture · chunk.json” panel to the web playground to generate and download per-track chunk.json metadata files.
  • Added phrase-boundary detection for the selected track, with a preview/count shown in the capture UI.
  • Added a tag palette and capture inputs for tags, quality flags, reviewer decision, rights/acquisition info, tuning, and timestamps.
• Documentation
  • Added an Architecture Decision Record (ADR-0026) describing the minimal in-browser chunk capture workflow.

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, you can upgrade your account or add credits to your account and enable them for code reviews in your settings.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: d298e972-9273-4fbc-9974-52a5912d1102

📥 Commits

Reviewing files that changed from the base of the PR and between 7860feb and 808ffcd.

📒 Files selected for processing (2)

• web/static/app.js
• web/static/index.html

🚧 Files skipped from review as they are similar to previous changes (2)

• web/static/index.html
• web/static/app.js

---

📝 Walkthrough

Walkthrough

Adds ADR-0026 documenting a minimal download-only in-browser chunk.json capture tool. The implementation adds serde_json to the web crate, extends web/src/lib.rs with three wasm_bindgen exports (tag_palette_json, detect_boundaries_json, build_chunk_json), and adds a hidden capture panel to the web playground via HTML, CSS, and JS updates.

Changes

In-browser chunk.json capture tool (ADR-0026)

Layer / File(s)	Summary
Decision record and ADR index docs/adr/0026-web-chunk-capture-tool.md, docs/adr/README.md	ADR-0026 documents the download-only scope carve-out, the decision to reuse griff-core via wasm_bindgen with serde_json schema alignment, no IndexedDB persistence, page-supplied timestamps, deferred S8 features, and is registered in the ADR index.
WASM ChunkMeta assembly, serialization, and entrypoints web/Cargo.toml, web/src/lib.rs	Adds serde_json dependency; adds imports for corpus/boundary types; implements source-format detection, PPQN-scaled boundary detection, index/enum parsing, and ChunkMeta assembly; serializes to pretty JSON with {"error":...} envelopes; exposes tag_palette_json, detect_boundaries_json, and build_chunk_json as wasm_bindgen exports; adds unit tests for index parsing, field mapping, serde round-trips, error envelopes, and tag palette completeness.
Browser capture panel: HTML, CSS, and JS wiring web/static/index.html, web/static/style.css, web/static/app.js	Inserts hidden #capture section with all chunk metadata and rights form fields; adds capture panel CSS; updates wasm imports; adds DOM bindings and fileName/current state; reveals panel on file load; wires populateTagPalette, boundary detection, and chunk download handlers.

Sequence Diagram

sequenceDiagram
  participant User
  participant appjs as app.js
  participant WASM as lib.rs (wasm_bindgen)
  participant Core as griff_core

  User->>appjs: upload score file
  appjs->>WASM: tag_palette_json()
  WASM-->>appjs: palette JSON → populate multi-selects
  appjs->>appjs: unhide `#capture` panel, set default title

  User->>appjs: click "Detect Boundaries"
  appjs->>WASM: detect_boundaries_json(track)
  WASM->>Core: scale BoundaryConfig to PPQN, run boundary detection
  Core-->>WASM: phrase boundary list
  WASM-->>appjs: {"boundaries": [...]} or {"error": ...}
  appjs->>User: display bounds in `#capBounds`

  User->>appjs: click "Download chunk.json"
  appjs->>WASM: build_chunk_json(track, id, title, ..., created_at, updated_at)
  WASM->>Core: assemble ChunkMeta with boundaries, tags, rights
  Core-->>WASM: ChunkMeta
  WASM-->>appjs: {"ok": {<ChunkMeta JSON>}} or {"error": ...}
  appjs->>User: trigger <id>.chunk.json download

Loading

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~25 minutes

Possibly related PRs

• PhysShell/griff#66: Adds RightsInfo and boundary filling to ChunkMeta and a manifest builder that reads the *.chunk.json files this PR produces via build_chunk_json.

Poem

🐇 Hop, hop — a little form appears,
Pick your tags, your rights, your tuning clear.
Press detect, the boundaries align,
Then download JSON — oh, how divine!
No IndexedDB, just a tidy file,
griff manifest will sort the rest meanwhile. ✨

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The PR title 'feat(web): in-browser chunk.json capture tool (ADR-0026)' accurately and concisely describes the main change—adding an in-browser chunk.json capture tool to the web playground, with ADR reference for context.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch claude/modest-maxwell-0zecwn

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 3

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@web/static/app.js`:
- Around line 89-91: The capture title field (els.capTitle) is only auto-filled
when it is empty, so loading a second file leaves the previous file's title in
place. Update the file loading logic to always reset and populate the capTitle
value with the new file's name by removing the conditional check
(!els.capTitle.value) and directly assigning file.name.replace(/\.[^.]+$/, '')
to els.capTitle.value whenever a new source file is loaded.
- Around line 132-142: In the detectBoundaries() function, after successfully
parsing the WASM result and confirming there is no error in the res object,
clear the previous error message by calling capMsg() with an appropriate message
(likely an empty string or success message) before updating
els.capBounds.textContent with the new boundary detection results. This ensures
that any stale error styling or messages from previous failed detections are
cleared before displaying the successful detection feedback.

In `@web/static/index.html`:
- Around line 152-153: The elements with ids capBounds and capStatus are
dynamically updated but lack ARIA live region attributes, preventing screen
readers from announcing changes to users. Add the aria-live attribute set to
"polite" to both the capBounds output element and the capStatus paragraph
element to enable screen reader announcements of dynamic updates.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 66ce0bb8-e4bd-43bd-a7bb-b7db3462aac0

📥 Commits

Reviewing files that changed from the base of the PR and between 4a9f487 and 7860feb.

📒 Files selected for processing (7)

• docs/adr/0026-web-chunk-capture-tool.md
• docs/adr/README.md
• web/Cargo.toml
• web/src/lib.rs
• web/static/app.js
• web/static/index.html
• web/static/style.css