Flask Weather App

A small Flask application that looks up current weather from OpenWeatherMap by US ZIP code.

Features added:

• Reads OpenWeather API key from the environment variable OPENWEATHER_API_KEY (recommended).
• Optional local .env support for development via python-dotenv.
• WeatherClient using a shared requests Session with retries, timeouts and a small in-memory TTL cache.
• Unit tests (pytest) and a GitHub Actions CI workflow that runs linting and tests on PRs.

Quick start

1. Create and activate a virtual environment (macOS/Linux):

python3 -m venv .venv
source .venv/bin/activate
pip install --upgrade pip
pip install -r requirements.txt

2. Configure your OpenWeather API key (one of):

• Preferred: create a .env file from the provided example and add your key:

cp .env.example .env
# edit .env and set OPENWEATHER_API_KEY=your_real_key

• Or copy the example ini and add your key (legacy):

cp config.example.ini config.ini
# edit config.ini and set the api value

3. Run the app locally (development server):

# For flask CLI
export FLASK_APP=app.py
export FLASK_ENV=development
flask run

# Or directly with python
python app.py

Open http://127.0.0.1:5000 in your browser and submit a US ZIP code.

Tests and linting

Run tests and lint locally:

pytest -q
flake8 .

Unit tests mock external HTTP calls so they don't require a real API key.

Continuous Integration

A GitHub Actions workflow is included at .github/workflows/ci.yml. On pull requests to main it will:

• Install dependencies from requirements.txt
• Run flake8 (linting)
• Run pytest

If any tests require real secrets in CI, add OPENWEATHER_API_KEY as a repository secret under Settings → Secrets & variables → Actions.

Notes and next steps

• The app uses an in-memory TTL cache suitable for development or single-instance deployments. For multi-instance or production use, swap this for Redis-backed Flask-Caching.
• Consider running the app under Gunicorn for production and adding a /health endpoint for healthchecks.
• The WeatherClient currently lives in app.py for simplicity — moving it to its own module improves testability and separation of concerns.

Screenshots