Contributing

Setup

git clone https://github.com/disoardi/hadoopscope
cd hadoopscope
python3 tests/run_all.py   # verify baseline

No virtual environment needed — the core uses only stdlib.

Coding rules

These are non-negotiable (see CLAUDE.md):

• Python 3.6+ compatible — no walrus operator, no dataclasses, no f"{x:.2f}"
• Zero pip dependencies for core (checks/, alerts/, config.py, bootstrap.py)
• run() never raises — always return CheckResult(status=UNKNOWN) on error
• Use # type: (...) annotations inline or import from typing explicitly
• No | in type hints — use Optional[X], Union[X, Y]

Adding a new check

1. Create or edit a file in checks/
2. Extend CheckBase:

class MyCheck(CheckBase):
    requires = []   # or [["ansible"]] if needed

    def run(self):
        # type: () -> CheckResult
        try:
            # ... your logic ...
            return CheckResult("MyCheck", CheckResult.OK, "all good")
        except Exception as e:
            return CheckResult("MyCheck", CheckResult.UNKNOWN, str(e))

1. Register it in hadoopscope.py → build_check_registry()
2. Add a fixture + test in tests/test_checks.py

Adding a new alert

1. Create alerts/my_alert.py with a dispatch(results, config, env_name) function
2. Import and call it in hadoopscope.py → dispatch_alerts()

Pull requests

• Run python3 tests/run_all.py before submitting
• One feature/fix per PR
• Update docs if adding a new check or config option