.github/workflows/train_models.yml

name: Train Models

on:
  workflow_run:
    workflows: ["Daily Data Update"]
    types: [completed]
  workflow_dispatch:          # Allow manual trigger
    inputs:
      model:
        description: "Which model to train (all / a / b / c)"
        required: false
        default: "all"
      epochs:
        description: "Max epochs"
        required: false
        default: "80"

jobs:
  train:
    runs-on: ubuntu-latest
    # Only run if the triggering workflow succeeded (or manual)
    if: >
      github.event_name == 'workflow_dispatch' ||
      github.event.workflow_run.conclusion == 'success'
    timeout-minutes: 360      # 6 hours max for heavy training

    steps:
      - name: Checkout repo
        uses: actions/checkout@v4

      - name: Set up Python
        uses: actions/setup-python@v5
        with:
          python-version: "3.11"
          cache: "pip"

      - name: Install dependencies
        run: pip install -r requirements.txt

      - name: Download latest data from HF
        env:
          HF_TOKEN:        ${{ secrets.HF_TOKEN }}
          HF_DATASET_REPO: ${{ secrets.HF_DATASET_REPO }}
        run: |
          python - <<'EOF'
          from huggingface_hub import hf_hub_download
          import os, shutil, config

          files = ["etf_price","etf_ret","etf_vol",
                   "bench_price","bench_ret","bench_vol","macro"]
          os.makedirs("data", exist_ok=True)
          for f in files:
              try:
                  path = hf_hub_download(
                      repo_id=config.HF_DATASET_REPO,
                      filename=f"data/{f}.parquet",
                      repo_type="dataset",
                      token=config.HF_TOKEN,
                  )
                  shutil.copy(path, f"data/{f}.parquet")
                  print(f"  Downloaded {f}.parquet")
              except Exception as e:
                  print(f"  Warning: {f} not found on HF: {e}")
          EOF

      - name: Train all models
        env:
          HF_TOKEN:        ${{ secrets.HF_TOKEN }}
          HF_DATASET_REPO: ${{ secrets.HF_DATASET_REPO }}
          FRED_API_KEY:    ${{ secrets.FRED_API_KEY }}
        run: |
          MODEL="${{ github.event.inputs.model || 'all' }}"
          EPOCHS="${{ github.event.inputs.epochs || '80' }}"
          python train.py --model $MODEL --epochs $EPOCHS

      - name: Run evaluation
        env:
          HF_TOKEN:        ${{ secrets.HF_TOKEN }}
          HF_DATASET_REPO: ${{ secrets.HF_DATASET_REPO }}
          FRED_API_KEY:    ${{ secrets.FRED_API_KEY }}
        run: python evaluate.py

      - name: Run prediction
        env:
          HF_TOKEN:        ${{ secrets.HF_TOKEN }}
          HF_DATASET_REPO: ${{ secrets.HF_DATASET_REPO }}
          FRED_API_KEY:    ${{ secrets.FRED_API_KEY }}
        run: python predict.py

      - name: Push weights + results to Hugging Face
        env:
          HF_TOKEN:        ${{ secrets.HF_TOKEN }}
          HF_DATASET_REPO: ${{ secrets.HF_DATASET_REPO }}
        run: python data_upload_hf.py --weights

      - name: Commit models + results to GitHub repo
        run: |
          git config user.name  "github-actions[bot]"
          git config user.email "github-actions[bot]@users.noreply.github.com"
          git add models/ evaluation_results.json latest_prediction.json || true
          git diff --cached --quiet || \
            git commit -m "[auto] Retrain complete — $(date -u +%Y-%m-%d)"
          git push || true

      - name: Push updated app to HF Space
        env:
          HF_TOKEN: ${{ secrets.HF_TOKEN }}
        run: |
          python - <<'EOF'
          from huggingface_hub import HfApi, CommitOperationAdd
          import glob, config

          api = HfApi(token=config.HF_TOKEN)

          # Files to push to HF Space
          space_files = (
              ["app.py", "config.py", "preprocess.py",
               "model_a.py", "model_b.py", "model_c.py",
               "predict.py", "evaluate.py",
               "requirements.txt",
               "evaluation_results.json",
               "latest_prediction.json"] +
              glob.glob("models/**/*.keras", recursive=True) +
              glob.glob("models/**/*.pkl",   recursive=True) +
              glob.glob("models/**/*.json",  recursive=True)
          )

          operations = []
          import os
          for f in space_files:
              if os.path.exists(f):
                  operations.append(CommitOperationAdd(
                      path_in_repo=f,
                      path_or_fileobj=f,
                  ))

          if operations:
              api.create_commit(
                  repo_id=config.HF_SPACE_REPO,
                  repo_type="space",
                  operations=operations,
                  commit_message="[auto] Deploy updated models to HF Space",
              )
              print(f"  Pushed {len(operations)} files to HF Space.")
          EOF

      - name: Done
        run: echo "✅ Training pipeline complete."