ci.yml

# CI — validate that the Airflow DAGs import cleanly.
#
# This runs ENTIRELY on GitHub's runner with NO secrets and NO access to your
# Airflow — it just installs Airflow 2.10.5 (same version as your Dockerfile),
# loads the dags folder into a DagBag, and fails if any DAG has a syntax/import
# error. It's the standard "did I break a DAG?" gate and gives you a green check
# on every push/PR. Start here to see GitHub Actions actually run.

name: CI — validate DAGs

on:
  push:
    branches: [main]
    paths:
      - "airflow/dags/**"
      - "airflow/requirements.txt"
      - ".github/workflows/ci.yml"
  pull_request:
    paths:
      - "airflow/dags/**"

jobs:
  validate-dags:
    runs-on: ubuntu-latest
    env:
      AIRFLOW_VERSION: "3.2.2"
      PYTHON_VERSION: "3.12"
      AIRFLOW_HOME: ${{ github.workspace }}/.airflow_home
      AIRFLOW__CORE__LOAD_EXAMPLES: "false"
      # config.py calls Variable.get(...) at import time. Airflow resolves
      # AIRFLOW_VAR_* from the environment before hitting the DB, so this lets the
      # DAGs import without a populated metadata DB.
      AIRFLOW_VAR_STREAMING_S3_BUCKET: "ci-placeholder-bucket"
      # The Cosmos DAG parses the dbt project at this path (the checkout) via
      # LoadMode.CUSTOM — no dbt binary or warehouse needed at parse time.
      DBT_PROJECT_DIR: ${{ github.workspace }}/dbt/streaming
      DBT_EXECUTABLE_PATH: /usr/bin/true
    steps:
      - uses: actions/checkout@v4

      - name: Set up Python
        uses: actions/setup-python@v5
        with:
          python-version: "3.12"

      - name: Install Airflow + providers (version-constrained)
        run: |
          CONSTRAINTS="https://raw.githubusercontent.com/apache/airflow/constraints-${AIRFLOW_VERSION}/constraints-${PYTHON_VERSION}.txt"
          pip install "apache-airflow==${AIRFLOW_VERSION}" --constraint "$CONSTRAINTS"
          pip install -r airflow/requirements.txt --constraint "$CONSTRAINTS"

      - name: Initialize metadata DB (sqlite, throwaway)
        run: airflow db migrate

      - name: Check DAGs import without errors
        run: |
          python - <<'PY'
          import os, sys
          # Airflow puts the dags folder on sys.path at runtime; replicate that so
          # `from common import config` resolves to airflow/dags/common (not the
          # top-level common/ which only has validation.py).
          sys.path.insert(0, os.path.abspath("airflow/dags"))
          from airflow.models import DagBag
          bag = DagBag("airflow/dags", include_examples=False)
          if bag.import_errors:
              for path, err in bag.import_errors.items():
                  print(f"::error file={path}::{err}")
              raise SystemExit(f"{len(bag.import_errors)} DAG import error(s)")
          print(f"OK — {len(bag.dags)} DAG(s) imported cleanly: {sorted(bag.dags)}")
          PY