weather-exporter/main.py

"""
Present observational weather data to Prometheus
"""

import json
import time
import os

import requests
from flask import Flask, Response

app = Flask(__name__)


def fetch_metoffice_data(location: int, apikey: str) -> dict:
    """
    Fetch current data from the Met Office for the provided postcode
    """

    obs_data = requests.get(
        f'http://datapoint.metoffice.gov.uk/public/data/val/wxobs/all/json/{location}?'
        f'key={apikey}&'
        'res=hourly'
    )

    return json.loads(obs_data.content)


@app.route('/metrics')
def metrics():
    """
    Output Prometheus-style metrics
    """
    metoff_apikey = os.environ.get('METOFF_APIKEY')
    metoff_location = os.environ.get('METOFF_LOCATION')

    hour_data = fetch_metoffice_data(metoff_location, metoff_apikey,)['SiteRep']['DV'][
        'Location'
    ]['Period'][-1]['Rep'][time.gmtime().tm_hour]

    ret_data = [
        {
            'key': 'sensor_weather_outdoor_temperature_celsius',
            'labels': {
                'location': metoff_location,
            },
            'type': 'gauge',
            'value': float(hour_data['T']),
        },
        {
            'key': 'sensor_weather_outdoor_humidity_percent',
            'labels': {
                'location': metoff_location,
            },
            'type': 'gauge',
            'value': float(hour_data['H']),
        },
    ]

    ret_strs = list()
    for item in ret_data:
        ret_strs.append(f'# HELP {item["key"]} Weather metric')
        ret_strs.append(f'# TYPE {item["key"]} {item["type"]}')
        ret_strs.append(
            item["key"]
            + '{'
            + " ".join([f'{key}="{val}"' for key, val in item["labels"].items()])
            + '} '
            + str(item["value"])
        )

    resp = Response('\n'.join(ret_strs))
    resp.headers['Content-type'] = 'text/plain'

    return resp


if __name__ == '__main__':
    app.run(host='0.0.0.0')