Daggie The Airflow DAG Quality Auditor

Features

🔍 1. Daggie-Score API (FastAPI Plugin)

A FastAPI application embedded inside Airflow.
It exposes the endpoint:

GET /api/daggie-score/score?dag_id=<your_dag_id>

Returns:

• Overall DAG score (0--100)

• Points earned vs max possible

• Detailed checks with:

  • name

  • weight

  • passed / failed

  • reason

This works by scanning the DAG loaded in Airflow's DagBag.

🧠 2. Scoring Rules (Built-in)

Daggie evaluates:

• Owner configured\

• default_args defined\

• retries & retry_delay\

• schedule / timetable\

• catchup behavior\

• tags present\

• DAG documentation\

• max_active_runs sanity\

• number of tasks\

• snake_case task IDs\

• placeholder check: no import-time side effects

All rules have weights and descriptive explanations.

🧩 3. Custom Scoring Plugins (Extensible)

You can add your own checks by adding Python callables:

Example:

from daggie_score import register_check @register_check def check_bigquery_usage(dag): if "bigquery" in dag.tags: return { "name": "BigQuery DAG Standards", "weight": 8, "passed": False, "reason": "Needs special review." }

Custom checks are merged into the official scoring pipeline.

🎨 4. React UI Plugin (Built for Airflow UI)

The React frontend appears in:

/dags/<dag_id>/plugin/simple-react-app

It shows:

• Score summary\

• Weighted check breakdown\

• Clean Material-UI styled presentation\

• Animated Daggie character whose mood changes:

  • 😀 Happy (score > 80)

  • 😐 Neutral (score 50--80)

  • 😢 Sad (score < 50)

Daggie Animation

Built entirely using CSS + React.
No Lottie/Javascript animations required.
Runs smoothly inside Airflow without external assets.