.github/workflows/hf_sync.yml

name: Sync to HF Space

on:
  push:
    branches:
      - beta/sanitized-minimal
  workflow_dispatch:
    inputs:
      ref:
        description: "Branch or SHA to deploy (default: beta/sanitized-minimal)"
        required: false
        default: "beta/sanitized-minimal"

concurrency:
  group: hf-space-sync
  cancel-in-progress: true

jobs:
  hf-sync:
    runs-on: ubuntu-latest
    timeout-minutes: 15

    steps:
      - name: Checkout deployment ref
        uses: actions/checkout@v4
        with:
          fetch-depth: 0
          lfs: true
          ref: ${{ github.event_name == 'workflow_dispatch' && github.event.inputs.ref || github.ref }}

      - name: Resolve checked out ref
        id: ref
        run: |
          BRANCH_NAME=$(git rev-parse --abbrev-ref HEAD)
          if [ "$BRANCH_NAME" = "HEAD" ]; then
            BRANCH_NAME=$(git name-rev --name-only --exclude=tags/* HEAD | sed 's#^remotes/origin/##')
          fi
          echo "branch_name=$BRANCH_NAME" >> "$GITHUB_OUTPUT"
          echo "Deploying ref: $BRANCH_NAME @ $(git rev-parse --short HEAD)"

      - name: Enforce sanitized deployment policy
        run: |
          set -euo pipefail

          echo "Checking required runtime files..."
          required_files=(
            "README.md"
            "Dockerfile"
            "requirements-hf.txt"
            "gui/streamlit_app.py"
            "engine/pipeline_orchestrator.py"
            "scripts/docker_entrypoint.sh"
          )
          for f in "${required_files[@]}"; do
            [ -f "$f" ] || { echo "::error::Missing required file: $f"; exit 1; }
          done

          echo "Checking blocked paths/files..."
          blocked_paths=(
            "Docs"
            "Docs/audit"
            "Vault"
            "logs"
            "runs"
            ".env"
            "TODO.md"
            "docs/progress.md"
          )
          violations=0
          for p in "${blocked_paths[@]}"; do
            if [ -e "$p" ]; then
              echo "::error::Blocked path present in deployment ref: $p"
              violations=1
            fi
          done
          [ "$violations" -eq 0 ]

          echo "Checking for obvious secret patterns..."
          if command -v rg >/dev/null 2>&1; then
            rg -n --hidden \
              -g '!.git/**' \
              -g '!**/*.md' \
              -e 'AKIA[0-9A-Z]{16}' \
              -e 'ASIA[0-9A-Z]{16}' \
              -e 'ghp_[A-Za-z0-9]{36}' \
              -e 'hf_[A-Za-z0-9]{30,}' \
              -e 'sk_live_[0-9A-Za-z]{20,}' \
              -e '-----BEGIN (RSA|EC|OPENSSH|DSA) PRIVATE KEY-----' \
              -e 'xox[baprs]-[A-Za-z0-9-]{10,}' \
              . && { echo "::error::Potential secret detected"; exit 1; } || true
          fi

      - name: Check large files (>10 MB)
        run: |
          set -euo pipefail
          large_files=$(find . -not -path './.git/*' -type f -size +10M 2>/dev/null || true)
          if [ -n "$large_files" ]; then
            echo "::warning::Large files detected (>10MB):"
            echo "$large_files"
            echo "Consider Git LFS for required large assets."
          else
            echo "No large files >10MB found."
          fi

      - name: Push to HF Space
        env:
          HF_TOKEN: ${{ secrets.HF_TOKEN }}
        run: |
          set -euo pipefail

          if [ -z "${HF_TOKEN:-}" ]; then
            echo "::error::HF_TOKEN secret is not set. Configure it in GitHub Actions secrets."
            exit 1
          fi

          git config user.email "ci@github.com"
          git config user.name "CI Bot"

          git remote remove hf 2>/dev/null || true
          git remote add hf "https://moddux:${HF_TOKEN}@huggingface.co/spaces/moddux/mod-osint"

          echo "Pushing $(git rev-parse --short HEAD) from '${{ steps.ref.outputs.branch_name }}' -> HF main"
          git push hf HEAD:main --force-with-lease

          echo "HF deploy push complete."
          echo "Logs: https://huggingface.co/spaces/moddux/mod-osint/logs"