Skip to content

Met.no API Client

The metno_api module provides a typed Python client for the met.no locationforecast v2.0 API, making it easy to fetch short-range weather forecasts for any global coordinate. It integrates tightly with the rest of weather_tools, returning Pydantic models and ready-to-use pandas DataFrames.

Highlights

  • ✅ Pydantic-powered queries via MetNoQuery for validated coordinates and formats.
  • ✅ Automatic retries, caching, and backoff to keep requests resilient and efficient.
  • ✅ Convenience helpers to aggregate hourly forecasts into daily summaries.
  • ✅ DataFrame conversion for both hourly and daily views.
  • ✅ Seamless hand-off to merge_weather_data.merge_historical_and_forecast for blending with SILO archives.

Requirements

  • met.no requires a descriptive User-Agent header containing a contact address. The client generates a sensible default (weather-tools/<version> (Python <major>.<minor>)), but you should provide your own identifier in production.
  • Forecast requests do not require authentication, but met.no enforces rate limits—respect their usage policy and cache responses where possible.

Quick Start

from weather_tools.metno_api import MetNoAPI
from weather_tools.metno_models import MetNoFormat, MetNoQuery
from weather_tools.silo_models import AustralianCoordinates

# Configure the client (always include your contact details!)
api = MetNoAPI(user_agent="my-app/1.0 (contact: you@example.com)")

# Build a query for a location near Sydney Airport
query = MetNoQuery(
    coordinates=AustralianCoordinates(latitude=-33.94, longitude=151.18),
    format=MetNoFormat.COMPACT,
)

# Execute and inspect the response metadata
response = api.query_forecast(query)
print(response.get_meta())

Aggregating to Daily Data

The API returns hourly forecasts in GeoJSON format. Use MetNoAPI.to_dataframe to convert the payload into either daily summaries or the raw hourly table:

daily_df = api.to_dataframe(response, aggregate_to_daily=True)
hourly_df = api.to_dataframe(response, aggregate_to_daily=False)
  • Daily data mirrors the DailyWeatherSummary model, providing min/max temperature, total precipitation, wind statistics, humidity, pressure, cloud cover, and the dominant weather symbol.
  • Hourly data preserves the original timestamps alongside instantaneous variables and precipitation totals for the next 1/6/12 hours.

Convenience Helper: get_daily_forecast

For common use cases you can skip manual query construction:

daily = api.get_daily_forecast(latitude=-33.94, longitude=151.18, days=7)

This method returns a list of DailyWeatherSummary models and automatically truncates the forecast horizon (met.no serves up to 9 days).

Configuration Options

Instantiate MetNoAPI with custom behaviour when needed:

api = MetNoAPI(
    user_agent="my-app/1.0 (contact: you@example.com)",
    timeout=45,
    max_retries=4,
    retry_delay=1.5,
    enable_cache=True,
    cache_expiry_hours=2,
    log_level="DEBUG",
)
Parameter Type Default Description
user_agent str auto Identifies your application to met.no
timeout int 30 HTTP timeout in seconds
max_retries int 3 Attempts for transient failures
retry_delay float 1.0 Base delay for exponential backoff
enable_cache bool True In-memory response cache toggle
cache_expiry_hours int 1 Lifetime for cached responses
log_level str/int INFO Logging level for diagnostics

Use clear_cache() and get_cache_size() to manage cached responses explicitly.

Error Handling

The client surfaces clear exceptions:

  • MetNoUserAgentError — missing/invalid User-Agent (met.no returns HTTP 403).
  • MetNoRateLimitError — too many requests (HTTP 429).
  • MetNoAPIError — other HTTP failures or JSON parsing issues.

Wrap network calls accordingly:

try:
    response = api.query_forecast(query)
except MetNoRateLimitError:
    logger.warning("Slow down—met.no rate limit hit")
except MetNoUserAgentError as e:
    raise RuntimeError("Update your User-Agent") from e

Integration with SILO Data

Daily summaries pair naturally with SILO history:

from weather_tools.merge_weather_data import merge_historical_and_forecast

merged = merge_historical_and_forecast(
    silo_data=silo_history_dataframe,
    metno_data=daily_df,
    fill_missing=True,
    overlap_strategy="prefer_silo",
)

See Merge Weather Data for a deeper dive into the blending workflow.

Additional Resources