🧪 Weather Client Project — Summary

This project is a simple Python weather client that fetches current weather data from the WeatherStack API and models the response using Pydantic. In this project we also used UV.

🗂️ Project Structure

myweather/
├── main.py                  # App entry point
├── .env                     # Contains WEATHERSTACK_TOKEN (not committed)
├── pyproject.toml           # Dependency and package config
├── uv.lock                  # Locked dependencies
├── myweather/
│   ├── __init__.py
│   ├── client.py            # WeatherClient class with logging + error handling
│   └── models.py            # Pydantic models for typed weather response
└── tests/
    └── test_client.py       # Unit tests with pytest

🔐 Secret Management

• The API token is stored in .env and loaded using python-dotenv
• .env is added to .gitignore
• Loaded in main.py using os.getenv("WEATHERSTACK_TOKEN")

🔧 WeatherClient Class

Defined in myweather/client.py:

• Calls http://api.weatherstack.com/current
• Handles:
  • HTTPError
  • ConnectionError
  • Timeout
  • RequestException
  • Pydantic ValidationError
• Returns a WeatherResponse Pydantic model on success, or None on failure
• Uses logging instead of print() for all diagnostics

🧱 Pydantic Models

Defined in myweather/models.py:

from pydantic import BaseModel
from typing import List

class Location(BaseModel):
    name: str
    country: str

class CurrentWeather(BaseModel):
    temperature: int
    weather_descriptions: List[str]

class WeatherResponse(BaseModel):
    location: Location
    current: CurrentWeather

🧪 Unit Testing with pytest

Unit tests are located in the tests/ folder and use pytest for simple testing of the WeatherClient.

• monkeypatch is used to mock the requests.get call
• The test checks both:
  • Successful weather response
  • Handling of HTTP errors

✅ Example test (tests/test_client.py)

def fake_get_error(*args, **kwargs):
    class FakeResponse:
        def raise_for_status(self):
            raise requests.exceptions.HTTPError("401 Client Error: Unauthorized")
    return FakeResponse()

def test_get_weather_http_error_handled(monkeypatch):
    import requests
    monkeypatch.setattr(requests, "get", fake_get_error)

    client = WeatherClient(token="invalid")
    result = client.get_weather("Oslo")

    assert result is None

✅ Future Ideas

• Expand Pydantic models
  Add support for more fields from the API, such as:

  • Humidity
  • "Feels like" temperature
  • Moon phase, visibility, etc.

• replace argparse with something fancier (like typer, which is based on Pydantic too).

🚀 Usage

We added a CLI interface (with argparse) to allow passing the city as a command-line argument. To run the weather client, use:

uv run main.py --city Oslo

• Make sure your .env file contains a valid WEATHERSTACK_TOKEN.
• If your city is a composed name, use double quotes, for example: "Los Angeles".
• If no city name is passed, the default city is "Tananger".