# Parallel Work Handoff

Current status: Phase 1 is mostly complete. `app.py` is now a `create_app()`
factory, `/api/v1` lives in `blueprints/api.py`, and legacy Jinja route groups
are split under `routes/`. Runtime verification is still pending because the
local environment does not have the full Flask/Postgres stack available.

## Good Parallel Workstreams

### Agent A: Phase 1 Verification + Cleanup
- Owns: `app.py`, `run_app.py`, `start_server.py`, `Dockerfile`, README/DEPLOYMENT.
- Verify `flask --app app:create_app routes` and startup in a deps-installed environment.
- Check Auth0 registration under repeated `create_app()` calls.
- Confirm `url_for(...)` endpoints still exist for templates.
- Remove stale compatibility notes from `progress.md` once verified.
- Avoid changing filter pipeline internals.

### Agent B: Phase 2 Filter Registry
- Owns: `filter_pipeline/`, `filter_config.json`, `filtersets.json`.
- Add stage/plugin registry and auto-discovery.
- Replace hardcoded stage dicts in `filter_pipeline/engine.py`.
- Bridge `BaseFilterPlugin` to live `FilterResult` handling.
- Avoid editing route modules except for minimal API integration points.

### Agent C: Comment Filtering Consolidation
- Owns: `comment_lib.py`, `filter_lib.py`, comment-related pipeline stages.
- Port comment tree modes into `filter_pipeline/stages/comment_filter.py`.
- Wire filtered comments into `/api/v1/posts/<uuid>` and `/api/v1/comments/<uuid>`.
- Add small sample-data tests if a test harness is added.
- Coordinate with Agent B on registry names and stage contracts.

### Agent D: Phase 3 Data Model Prep
- Owns: `models.py`, new migration scripts, DB query services.
- Design `Post` and `Comment` SQLAlchemy models matching current JSON schema.
- Draft `migrate_content_to_db.py` backfill from `data/posts` and `data/comments`.
- Do not switch live reads/writes until Phase 2 route/filter behavior is stable.

### Agent E: Test Harness
- Owns: `tests/`, pytest config, lightweight fixtures.
- Add app-factory tests that assert route registration and endpoint names.
- Add `/api/v1` contract tests with monkeypatched `post_service` and `get_filter_engine`.
- Keep tests independent of a live Postgres where possible.

## Serialization Points

- Do Phase 1 runtime verification before deleting any more legacy UI behavior.
- Phase 2 registry work and comment filtering can happen together, but merge the registry contract first.
- Phase 3 database cutover should wait until `/api/v1/posts` and comment filtering behavior is stable.
- Phase 4 SPA work can scaffold independently, but feature parity work depends on stable `/api/v1` contracts.

## Shared Cautions

- Preserve endpoint names used by templates until the SPA replaces them.
- Do not delete Jinja templates in this pass.
- Do not remove JSON file reads until Postgres backfill and DB query paths are verified.
- Run at least `python -m py_compile` on touched Python files.
- Update `progress.md` after each substantial change.