name: Nightly Architectural Sweep

on:
  schedule:
    # 4:00 AM CST daily, inspired by the Aarete DocyAI architectural sweep cadence.
    # Gitea schedules use UTC. CST (UTC-6): 10:00 UTC.
    # During CDT this runs at 5:00 AM America/Chicago unless the cron is adjusted.
    - cron: '0 10 * * *'
  workflow_dispatch:

jobs:
  create-architectural-sweep-issue:
    name: Create Architectural Sweep Issue
    runs-on: ubuntu-latest

    steps:
      - name: Create architectural sweep issue
        run: |
          python3 - <<'PYEOF'
          import json
          import urllib.request
          from datetime import datetime, timezone

          gitea_api = "http://100.79.253.19:3000/api/v1"
          repo = "${{ gitea.repository }}"
          token = "${{ gitea.token }}"
          run_url = "http://100.79.253.19:3000/${{ gitea.repository }}/actions/runs/${{ gitea.run_id }}"
          today = datetime.now(timezone.utc).strftime("%Y-%m-%d")

          body = f"""## Architectural Sweep - {today}

          Daily architecture review for **Throne of the Bone King**.

          This follows the Aarete DocyAI sweep pattern: each domain should be reviewed independently, findings should be posted as separate comments, and critical issues should become focused follow-up tickets before moving on.

          **Action run:** {run_url}

          ---

          > Agent instructions:
          >
          > Post a separate comment for each domain as soon as that domain is complete.
          > Title comments with the domain and date, for example `Architecture Findings - {today}`.
          > Link follow-up issues from this sweep and include file paths or commands where relevant.

          ---

          ### 1. Shared App Architecture

          - [ ] Review `App.tsx`, `src/navigation/AppNavigator.tsx`, and `src/components/GameViewport.tsx`.
          - [ ] Confirm shared React Native code is still platform-safe and free of DOM-only APIs.
          - [ ] Check whether navigation, transitions, safe area, and fixed landscape viewport boundaries remain cleanly separated.
          - [ ] Identify any component that is taking on too many responsibilities.

          ### 2. State and Persistence

          - [ ] Review Zustand stores in `src/state/`.
          - [ ] Confirm persisted route and settings state can recover from stale or legacy values.
          - [ ] Check that platform storage remains isolated to `persistStorage.ts` and `persistStorage.web.ts`.
          - [ ] Look for race conditions between route transitions, socket connection updates, and Toolkit content refreshes.

          ### 3. Toolkit and Socket Sync

          - [ ] Review `src/screens/ToolkitScreen.tsx`, `src/socket/toolkitSocketClient.ts`, and `socket/server.mjs`.
          - [ ] Confirm Toolkit access is still Browser-resolution and socket-gated.
          - [ ] Verify content read/save/change/error paths remain typed and tested.
          - [ ] Identify the next volatile gameplay content area that should use the socket-backed JSON pattern.

          ### 4. Data, Content, and Gameplay Loops

          - [ ] Review `src/data/`, `wiki/`, and Toolkit route coverage.
          - [ ] Check that petitioners, rulings, seed logic, resources, factions, flags, story arcs, lore, audio, endings, and references still map to future gameplay loops.
          - [ ] Identify mismatches between wiki expectations and app/toolkit scaffolding.
          - [ ] Recommend one schema or fixture that should be formalized next.

          ### 5. UI, Layout, and Assets

          - [ ] Inspect Home, Settings, Policies, Credits, and Toolkit layouts against fixed landscape resolutions.
          - [ ] Check for accidental scrolling, clipped text, oversized controls, or insufficient contrast.
          - [ ] Review `src/assets/` usage, font loading, favicon, menu art, and background music.
          - [ ] Flag any asset size or bundling risk that affects web/Electron builds.

          ### 6. Test and Build Health

          - [ ] Run or review `npm run lint`.
          - [ ] Run or review `npm run test -- --runInBand`.
          - [ ] Run or review `npm run electron:build`.
          - [ ] Identify missing tests for any recent architecture change.
          - [ ] Scan for TODO/FIXME/HACK/TEMP, empty catches, broad casts, dead exports, or stale docs.

          ### 7. Technical Debt and Next Actions

          - [ ] List concrete architecture risks with severity and owner recommendation.
          - [ ] Estimate effort for each fix as S, M, or L.
          - [ ] Create follow-up issues for any item that should not wait for the next sweep.
          - [ ] Pick the single most valuable architecture improvement for the next work session.

          ---

          _Auto-generated nightly by Gitea Actions._"""

          payload = {
              "title": f"Architectural Sweep - {today}",
              "body": body,
          }

          req = urllib.request.Request(
              f"{gitea_api}/repos/{repo}/issues",
              data=json.dumps(payload).encode("utf-8"),
              headers={
                  "Authorization": f"token {token}",
                  "Content-Type": "application/json",
              },
              method="POST",
          )
          with urllib.request.urlopen(req, timeout=20) as response:
              result = json.loads(response.read())
          print(f"Created issue #{result['number']}: {result['html_url']}")
          PYEOF