feat: manage existing-run experiment membership

[Cedric Vidal (cedricvidal)]

Summary

Adds complete-set membership revisions for creating Experiments from existing runs and atomically adding, moving, removing, or deleting members.

• Plans exact current run vectors before mutation, publishes one immutable revision/execution with CAS ownership, and supports deterministic replay plus interruption repair.
• Keeps Experiment history authoritative while maintaining request.experimentId as denormalized lookup state.
• Adds REST/OpenAPI and CLI parity, coordinated deletion, Runs-list bulk Create/Move, and current-attempt member removal within the compact Experiment Detail Runs tab.
• Covers inline/chunked storage transitions, stale/conflicting ownership, replay, retry/replan, deletion coordination, Portal interactions, and Storybook states.

Demos

Before

N/A — existing standalone runs could not be managed as Experiment membership.

After

Integrated Scope Portal walkthrough: bulk Create/Move conflict handling, Atom Experiment navigation, compact attempt history, completed/failed report versions, one-revision member removal, interrupted reconciliation, and retry recovery.

experiment-membership.mp4

Testing

• pnpm vitest run packages/shared/src/experiments/experiment-store.test.ts apps/api/src/experiment-membership.test.ts apps/api/src/endpoints.test.ts apps/api/src/routes/experiments.test.ts apps/cli/src/commands/experiment.test.ts — passed, 5 files / 210 tests.
• pnpm --filter portal test -- src/lib/experiment-membership.test.ts src/lib/api-experiment-membership.test.ts src/pages/ExperimentDetail.test.tsx src/pages/ExperimentDetail.play.test.tsx src/pages/ExperimentsList.test.tsx src/pages/ExperimentsList.play.test.tsx src/components/Layout.test.tsx src/components/ExperimentReportCard.test.tsx — passed, 8 files / 39 tests.
• pnpm --filter portal build — passed.
• pnpm --filter portal build-storybook — passed.
• pnpm test -- --testTimeout=15000 — passed before the later Portal-only PR B reconciliation, 286 files / 3,473 tests; affected Portal tests and builds were rerun above after the final merge.
• Website OpenAPI tests and production build — passed after restoring the website's frozen-lockfile dependencies; generated OpenAPI artifact matches the API snapshot.
• Retained demo take — H.264/yuv420p, 1440x900, 25 fps, 47.28 seconds / 1,182 frames; full FFmpeg decode and encoded first/last/contact-sheet inspection passed; moov precedes mdat.
• Compose smoke — environment-blocked, not a product failure. The normal attempt failed with shared Docker VM ENOSPC; the one permitted targeted retry failed at /var/tmp/libpod_builder with the same condition. Safe exact cleanup reclaimed 0 bytes because Docker metadata reported unable to open database file / retained references. Teardown verified zero project containers, networks, and volumes for scope-mt-app-9-cedricvidal-refactored-meme. No further Docker cleanup was attempted.
• Real Cosmos — not run because credentials were unavailable.

Documentation and compatibility

Updated docs/architecture/app-design.md, docs/architecture/experiments.md, apps/cli/README.md, and the generated website OpenAPI artifact.

The REST API adds existing-run creation plus bulk membership POST/DELETE routes. CLI adds experiment create --request-id, experiment add/move, and experiment remove. Existing APIs remain compatible. Request documents gain optional denormalized membership fields; no data migration or dependency change is required.

Checklist

• If Portal features changed, keep CLI capabilities in sync.
• If Portal components changed, update their Storybook stories.
• If database changes require a migration, include up() / down() and keep it CosmosDB-compatible. N/A — only optional schema fields were added.
• If dependencies changed, update the lockfile and regenerate NOTICE / NOTICE-REVIEW.txt with pnpm notice as needed. N/A — no dependency changes.
• Video showing the behavior before the suggested change. N/A — this adds a new membership workflow.
• Video showing the behavior after the suggested change.

[github-actions]

Test Results (Node.js 22)

test: Run #280

Tests 📝	Passed ✅	Failed ❌	Skipped ⏭️	Pending ⏳	Other ❓	Flaky 🍂	Duration ⏱️
3499	3498	1	0	0	0	0	1m56s

❌ Some tests failed!

Name	Failure Message
❌ ExperimentStore CAS contracts > plans chunked-to-chunked growth and chunked-to-inline shrink independently	Test timed out in 5000ms. If this is a long-running test, pass a timeout value as the last argument or configure it globally with "testTimeout".

Github Test Reporter

[github-actions]

Test Results (Node.js 22)

test: Run #290

Tests 📝	Passed ✅	Failed ❌	Skipped ⏭️	Pending ⏳	Other ❓	Flaky 🍂	Duration ⏱️
3504	3504	0	0	0	0	0	1m8s

🎉 All tests passed!

Github Test Reporter

[github-actions]

Test Results (Node.js 22)

test: Run #297

Tests 📝	Passed ✅	Failed ❌	Skipped ⏭️	Pending ⏳	Other ❓	Flaky 🍂	Duration ⏱️
3504	3503	1	0	0	0	0	1m56s

❌ Some tests failed!

Name	Failure Message
❌ ExperimentStore CAS contracts > plans chunked-to-chunked growth and chunked-to-inline shrink independently	Test timed out in 5000ms. If this is a long-running test, pass a timeout value as the last argument or configure it globally with "testTimeout".

Github Test Reporter