This package communicates with several web-based APIs to deliver weather data for US cities.

A GUI-based client is provided for use in standard desktop environments. To execute from this directory use the command:

python main.py

The source of the GUI client can be found in the gui subdirectory of this directory.

This package includes the WeatherService class in the source file weather/service.py which provides a simple API for running a local weather service on your system. An example of a server implementation is contained in weather-service.py. This implementation will run the service on your localhost at a provided port number; if no port number or an invalid port number is provided, the service will be run on an arbitrarily-chosen non-priveleged port (and the port so chosen will be printed to stderr).

The service accepts and responds with JSON-formatted data. An example client is provided below, and an additional example is provided in client-demo.py.

Here is a UML sequence diagram:

• request-type: The value of this key conveys the type of data being requested. Currently, the only recognized value is current-conditions which returns the most recent weather conditions reported by the weather station.
• city: The name of a US city, town, village, etc. This, along with state, is used to find the nearest weather station.
• state: The name of a US state. This, along with city, is used to find the nearest weather station.

• status: This key is used to indicate if a problem was encountered during retrieval. If the value of this key is ok, then no problems were encountered. If the value of this key is not ok, its value will an error code.
• response-type: The value of this key matches the request-type value of the request which generated this response.
• message: The actual content of the response. In the case of an error, this will contain information about the error.

In the following examples, we make use of the encode function provided by weather/service.py which transforms a Python dict object into a byte array of JSON-formatted data for socket transmission. The decode function does the opposite.

We can make use of the WeatherService class to retrieve information about a local service; namely, the host and port properties retrieve that information from the active service.

import socket
import weather.service as ws

request_dict = {
    'request-type':'current-conditions',
    'city': 'Cleveland',
    'state': 'Ohio'
}

server = ws.WeatherService()
payload = ws.encode(request_dict)
with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s:
    s.connect((server.host, server.port))
    s.sendall(payload)
    response = ws.decode(s.recv(1024))
print(response)

This would print something like:

{'status': 'ok', 'response-type': 'current-conditions', 'message': '47.0 ℉ (8.3 ℃) Fair'}

This is not substantially different from communicating with a local service, however we cannot rely on the WeatherService class to provide host and port information, which we'll need to know by other means.

import socket
import weather.service as ws

request_dict = {
    'request-type':'current-conditions',
    'city': 'Corvallis',
    'state': 'Oregon'
}

HOST = '192.168.5.23'
PORT = 49562
payload = ws.encode(request_dict)
with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s:
    s.connect((HOST, PORT))
    s.sendall(payload)
    response = ws.decode(s.recv(1024))
print(response)

This would print something like:

{'status': 'ok', 'response-type': 'current-conditions', 'message': '44.0 ℉ (6.7 ℃) Fog/Mist'}