.github/workflows/sync_to_hf_space.yml

name: Sync GitHub to Hugging Face Space

on:
  push:
    branches:
      - main
  workflow_dispatch:

jobs:
  sync-to-hf:
    runs-on: ubuntu-latest
    steps:
      - name: Checkout
        uses: actions/checkout@v4
        with:
          fetch-depth: 0
          lfs: true

      - name: Pull LFS files
        run: git lfs pull

      - name: Push to Hugging Face Space
        env:
          HF_TOKEN: ${{ secrets.HF_TOKEN }}
        run: |
          if [ -z "${HF_TOKEN}" ]; then
            echo "HF_TOKEN secret is not set."
            exit 1
          fi

          DEPLOY_DIR="/tmp/hf-deploy"
          rm -rf "${DEPLOY_DIR}"
          mkdir -p "${DEPLOY_DIR}"

          # Export the current tree (with real LFS file contents, not pointers)
          tar --exclude=.git -cf - . | (cd "${DEPLOY_DIR}" && tar -xf -)
          cd "${DEPLOY_DIR}"

          # Keep GitHub README clean; inject Space front matter only for HF deploy.
          if [ -f README.md ]; then
            awk '
              NR == 1 && $0 == "---" {in_yaml=1; next}
              in_yaml && $0 == "---" {in_yaml=0; next}
              !in_yaml {print}
            ' README.md > README.clean.md

            printf '%s\n' \
              '---' \
              'title: HS Code Classifier Micro' \
              'emoji: ⚡' \
              'colorFrom: pink' \
              'colorTo: blue' \
              'sdk: docker' \
              'app_port: 7860' \
              '---' \
              > README.frontmatter.md

            cat README.frontmatter.md README.clean.md > README.md
            rm -f README.frontmatter.md README.clean.md
          fi

          # HF rejects files >10MB without Git LFS.
          git lfs install
          git init
          git checkout -b main
          git config user.name "github-actions[bot]"
          git config user.email "41898282+github-actions[bot]@users.noreply.github.com"

          # Track large files with LFS for HF (only files still bundled)
          git lfs track "data/*.json" "models/umap_data.json"
          git add .gitattributes

          # Remove files not needed at runtime to stay under HF Space 1GB limit.
          # Large artifacts (sentence model, embeddings, classifier, training data)
          # are hosted on HF Hub at $SENTENCE_MODEL_NAME and downloaded at startup.
          rm -rf scripts/
          rm -rf models/sentence_model/
          rm -f models/embeddings.npy models/knn_classifier.pkl models/label_encoder.pkl models/metadata.json models/umap_data.json
          touch models/.gitkeep
          rm -f data/training_data.csv data/training_data_indexed.csv
          rm -f data/hts/hts_*.csv
          rm -f data/cargo_descriptions.csv
          rm -f data/training_data.json
          rm -f data/hf_real_data.csv

          git add -A
          git commit -m "Sync from GitHub ${GITHUB_SHA}"

          git remote add hf "https://oauth2:${HF_TOKEN}@huggingface.co/spaces/Mead0w1ark/MicroHS"
          git push --force hf main