name: E2E Tests

on:
  pull_request:
    branches: [main, dev]
    types: [opened, synchronize, reopened, ready_for_review]
    paths:
      - 'surfsense_web/**'
      - 'surfsense_backend/**'
      - '.github/workflows/e2e-tests.yml'
  workflow_dispatch:

concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}
  cancel-in-progress: true

jobs:
  e2e:
    name: Playwright E2E
    runs-on: ubuntu-latest
    if: github.event.pull_request.draft == false
    timeout-minutes: 45

    services:
      postgres:
        image: pgvector/pgvector:pg17
        env:
          POSTGRES_USER: postgres
          POSTGRES_PASSWORD: postgres
          POSTGRES_DB: surfsense_e2e
        ports:
          - 5432:5432
        options: >-
          --health-cmd "pg_isready -U postgres -d surfsense_e2e"
          --health-interval 10s
          --health-timeout 5s
          --health-retries 5
      # Required by Celery (broker + result backend) AND by the app's
      # own Redis-backed features (heartbeats, podcast markers, anon
      # quota). The previous workflow omitted this and indexing journeys
      # silently hung.
      redis:
        image: redis:8-alpine
        ports:
          - 6379:6379
        options: >-
          --health-cmd "redis-cli ping"
          --health-interval 10s
          --health-timeout 5s
          --health-retries 5

    env:
      # ---- Backend ------------------------------------------------------
      DATABASE_URL: postgresql+asyncpg://postgres:postgres@localhost:5432/surfsense_e2e
      CELERY_BROKER_URL: redis://localhost:6379/0
      CELERY_RESULT_BACKEND: redis://localhost:6379/0
      REDIS_APP_URL: redis://localhost:6379/0
      SECRET_KEY: ci-test-secret-key-not-for-production
      AUTH_TYPE: LOCAL
      REGISTRATION_ENABLED: "TRUE"
      ETL_SERVICE: DOCLING
      EMBEDDING_MODEL: sentence-transformers/all-MiniLM-L6-v2
      NEXT_FRONTEND_URL: http://localhost:3000

      # ---- Composio sentinel -------------------------------------------
      # Production code does `from composio import Composio` at import
      # time. `tests/e2e/run_backend.py` and `run_celery.py` hijack
      # sys.modules BEFORE that import resolves, so the real SDK is
      # never loaded. This sentinel API key is defense layer 3 from
      # surfsense_backend/tests/e2e/README.md: if the hijack ever
      # silently breaks, any real Composio call will 401 loudly with
      # this token instead of using a stray developer key.
      COMPOSIO_API_KEY: e2e-deny-real-call-sentinel
      COMPOSIO_ENABLED: "TRUE"

      # ---- Frontend (read by `next dev` via playwright.config.ts) -----
      NEXT_PUBLIC_FASTAPI_BACKEND_URL: http://localhost:8000
      NEXT_PUBLIC_FASTAPI_BACKEND_AUTH_TYPE: LOCAL

      # ---- Playwright --------------------------------------------------
      PLAYWRIGHT_TEST_EMAIL: e2e-test@surfsense.net
      PLAYWRIGHT_TEST_PASSWORD: E2eTestPassword123!

    steps:
      - name: Checkout code
        uses: actions/checkout@v6

      # =================================================================
      # Backend: Python + uv + dependencies + migrations
      # =================================================================
      - name: Set up Python
        uses: actions/setup-python@v6
        with:
          python-version: '3.12'

      - name: Install uv
        uses: astral-sh/setup-uv@v8.1.0

      - name: Cache backend dependencies
        uses: actions/cache@v5
        with:
          path: |
            ~/.cache/uv
            surfsense_backend/.venv
          key: python-deps-${{ hashFiles('surfsense_backend/uv.lock') }}
          restore-keys: |
            python-deps-

      - name: Cache HuggingFace models
        uses: actions/cache@v5
        with:
          path: ~/.cache/huggingface
          key: hf-models-${{ env.EMBEDDING_MODEL }}-${{ env.ETL_SERVICE }}

      - name: Install backend dependencies
        working-directory: surfsense_backend
        run: uv sync

      - name: Run database migrations
        working-directory: surfsense_backend
        run: uv run alembic upgrade head

      # =================================================================
      # Boot the E2E backend.
      #
      # CRITICAL: do NOT run `uvicorn main:app` here. Production code
      # binds `from composio import Composio` (and friends) at import
      # time. `tests/e2e/run_backend.py` is the test-only entrypoint
      # that hijacks sys.modules before that import — without it, every
      # connector journey would call the real SDK.
      # See surfsense_backend/tests/e2e/README.md.
      # =================================================================
      - name: Start backend (E2E entrypoint with sys.modules hijack)
        working-directory: surfsense_backend
        run: |
          uv run python tests/e2e/run_backend.py \
            > backend.log 2>&1 &
          echo $! > backend.pid

      # Celery runs in its own interpreter, so the hijack from
      # run_backend.py does NOT carry over. run_celery.py reapplies it
      # before importing celery_app. Without this worker, indexing
      # tasks queue but never execute and journey specs hang.
      - name: Start Celery worker (E2E entrypoint)
        working-directory: surfsense_backend
        run: |
          uv run python tests/e2e/run_celery.py \
            > celery.log 2>&1 &
          echo $! > celery.pid

      - name: Wait for backend readiness
        run: |
          for i in $(seq 1 60); do
            if curl -sf http://localhost:8000/openapi.json > /dev/null; then
              echo "Backend up after ${i} attempts"
              exit 0
            fi
            sleep 2
          done
          echo "::error::Backend failed to start within 120s"
          echo "===== backend.log (tail 200) ====="
          tail -200 surfsense_backend/backend.log || true
          echo "===== celery.log (tail 200) ====="
          tail -200 surfsense_backend/celery.log || true
          exit 1

      - name: Wait for Celery worker readiness
        working-directory: surfsense_backend
        run: |
          for i in $(seq 1 30); do
            if uv run celery -A app.celery_app inspect ping --timeout 2 \
                > /dev/null 2>&1; then
              echo "Celery worker up after ${i} attempts"
              exit 0
            fi
            sleep 2
          done
          echo "::error::Celery worker failed to start within 60s"
          echo "===== celery.log (tail 200) ====="
          tail -200 celery.log || true
          exit 1

      - name: Register E2E test user
        run: |
          # Idempotent: 200/201 = created, 400 = already exists (also OK)
          STATUS=$(curl -s -o /tmp/register.json -w "%{http_code}" \
            -X POST http://localhost:8000/auth/register \
            -H "Content-Type: application/json" \
            -d "{\"email\":\"${PLAYWRIGHT_TEST_EMAIL}\",\"password\":\"${PLAYWRIGHT_TEST_PASSWORD}\"}")
          echo "Register status: ${STATUS}"
          cat /tmp/register.json
          if [ "${STATUS}" != "200" ] && [ "${STATUS}" != "201" ] && [ "${STATUS}" != "400" ]; then
            echo "::error::Failed to register test user (status ${STATUS})"
            exit 1
          fi

      # =================================================================
      # Frontend: Node + pnpm + Playwright
      # =================================================================
      - name: Setup Node.js
        uses: actions/setup-node@v6
        with:
          node-version: '20'

      - name: Install pnpm
        uses: pnpm/action-setup@v6
        with:
          version: 10

      - name: Get pnpm store directory
        id: pnpm-cache
        shell: bash
        run: echo "STORE_PATH=$(pnpm store path --silent)" >> $GITHUB_OUTPUT

      - name: Cache pnpm store
        uses: actions/cache@v5
        with:
          path: ${{ steps.pnpm-cache.outputs.STORE_PATH }}
          key: pnpm-${{ runner.os }}-${{ hashFiles('surfsense_web/pnpm-lock.yaml') }}
          restore-keys: |
            pnpm-${{ runner.os }}-

      - name: Install web dependencies
        working-directory: surfsense_web
        run: pnpm install --frozen-lockfile

      - name: Cache Playwright browsers
        id: playwright-cache
        uses: actions/cache@v5
        with:
          path: ~/.cache/ms-playwright
          key: playwright-${{ runner.os }}-${{ hashFiles('surfsense_web/pnpm-lock.yaml') }}

      - name: Install Playwright browsers
        if: steps.playwright-cache.outputs.cache-hit != 'true'
        working-directory: surfsense_web
        run: pnpm exec playwright install --with-deps chromium

      - name: Install Playwright system deps (cache hit)
        if: steps.playwright-cache.outputs.cache-hit == 'true'
        working-directory: surfsense_web
        run: pnpm exec playwright install-deps chromium

      # playwright.config.ts boots `pnpm exec next dev` automatically
      # via webServer config (skipped when PLAYWRIGHT_NO_WEB_SERVER set).
      - name: Run Playwright tests
        working-directory: surfsense_web
        run: pnpm test:e2e

      # =================================================================
      # Diagnostics
      # =================================================================
      - name: Upload Playwright HTML report
        if: always()
        uses: actions/upload-artifact@v7
        with:
          name: playwright-report
          path: surfsense_web/playwright-report/
          retention-days: 14

      - name: Upload Playwright traces / videos
        if: failure()
        uses: actions/upload-artifact@v7
        with:
          name: playwright-traces
          path: surfsense_web/test-results/
          retention-days: 14

      - name: Upload backend + celery logs
        if: failure()
        uses: actions/upload-artifact@v7
        with:
          name: backend-celery-logs
          path: |
            surfsense_backend/backend.log
            surfsense_backend/celery.log
          retention-days: 7

      - name: Stop backend + Celery worker
        if: always()
        working-directory: surfsense_backend
        run: |
          for f in backend.pid celery.pid; do
            if [ -f "$f" ]; then
              kill "$(cat $f)" 2>/dev/null || true
            fi
          done