chore: establish pre-exhibit project baseline

.dockerignore

@@ -0,0 +1,6 @@
+node_modules
+dist
+data
+.git
+.env
+npm-debug.log

.env.example

@@ -0,0 +1,5 @@
+DATABASE_URL=postgres://osint:osint_secret@localhost:5433/osint_dev
+PORT=8787
+CORS_ORIGIN=http://localhost:5173
+LEVEL_EDITING_ENABLED=true
+MAX_DOCUMENT_BYTES=26214400

.gitignore

@@ -0,0 +1,5 @@
+node_modules/
+dist/
+.env
+.DS_Store
+*.tsbuildinfo

docs/TODO.md

@@ -1,15 +1,49 @@
 # GUPI OSINT Board roadmap
 
-This list tracks domain and product work that follows the accepted exhibit model. It is not a substitute for migrations or implementation issues.
+This is the ordered implementation roadmap following the accepted exhibit model. Work generally proceeds from top to bottom. It is not a substitute for migrations or implementation issues.
 
-## Exhibit-schema cutover
+## Working agreement
+
+- Spend roughly 60% of development time on features and model work, 25% on tests and bug fixing, and 15% on repository and deployment hygiene.
+- Stop feature work for interaction-breaking, data-loss, migration, or deployment bugs. Batch cosmetic defects separately.
+- A new exhibit behavior is complete only when its PostgreSQL representation, API behavior, frontend widget, persistence, and focused tests agree.
+- Mystery content must be authored as template data through supported application operations, never hard-coded into React components.
+
+## Milestone 0: trustworthy baseline
+
+- [x] Track `.gitignore`, `.dockerignore`, and `.env.example`; confirm that secrets and generated artifacts cannot enter Git accidentally.
+- [x] Commit the current working POC and tag the checkpoint `poc-pre-exhibit-model`.
+- [x] Confirm `main` is pushed to the Ramanujan-hosted `origin` repository.
+- [ ] Keep `npm run build` green and make `npm test` run real tests rather than an empty suite.
+
+## Milestone 1: focused safety net
+
+- [ ] Test screen/board coordinate conversion across pan and zoom levels.
+- [ ] Test timeline date-to-pixel projection and recomputation after viewport resizing.
+- [ ] Test the interaction boundary between exhibit dragging, hand-tool panning, and board-only pinch zoom.
+- [ ] Test folder open/close behavior, retained file positions, and containment bands.
+- [ ] Test document upload, metadata persistence, board save/reload, and reset.
+- [ ] Run migrations and API integration tests against disposable PostgreSQL, not SQLite or mocked persistence.
+- [ ] Add one browser smoke test: open a level, drag an exhibit, zoom, open a folder, move a file, reload, and verify persistence.
+
+## Milestone 2: exhibit-schema cutover
 
 - [ ] Add boards, exhibits, exhibit types, subtype tables, immutable template versions, and cloned mutable levels.
 - [ ] Migrate transitional `widgets`, `widget_relations`, and `playthrough_*` data with equivalence checks.
 - [ ] Implement template instantiation and “save level as template” as transactional clone operations.
+- [ ] Introduce a server-side repository/service boundary so SQL and cloning transactions do not live in Express route handlers.
 - [ ] Move the frontend to an exhibit/widget registry backed by the normalized API.
+- [ ] Remove transitional tables only after automated data-equivalence and behavior checks pass.
 
-## Party exhibits
+## Milestone 3: events and parties
+
+### Events and narrative
+
+- [ ] Implement Event exhibits with occurrence time and investigator-authored narrative text.
+- [ ] Implement normalized Event-to-Evidence links and their board visualization.
+- [ ] Present chronologically ordered events as the emerging investigation story.
+
+### Party exhibits
 
 - [ ] Add a Party exhibit supertype representing an investigation participant.
 - [ ] Add a Person subtype with display name, normalized aliases, and extensible structured identity fields.
@@ -19,8 +53,10 @@ This list tracks domain and product work that follows the accepted exhibit model
 - [ ] Build distinct Person and Organization widgets that can open as dossiers and reveal associated evidence without using folder ownership semantics.
 - [ ] Include parties, aliases, evidence associations, and party relationships in template/level cloning.
 
-## Events and narrative
+## Milestone 4: first playable mystery
 
-- [ ] Implement Event exhibits with occurrence time and investigator-authored narrative text.
-- [ ] Implement normalized Event-to-Evidence links and their board visualization.
-- [ ] Present chronologically ordered events as the emerging investigation story.
+- [ ] Design a small mystery that exercises documents, folders, notes, events, people, organizations, connections, and the timeline.
+- [ ] Create it as an immutable level-template version using the same supported operations available to an author.
+- [ ] Instantiate and solve a cloned level without modifying the template or relying on hard-coded case behavior.
+- [ ] Turn the successful solve path into an end-to-end acceptance test.
+- [ ] Perform a manual playability and visual-polish pass before production deployment.