Revision

draft

Forecast:tornado is a wrapper around Tornado web server that makes easy to organize your projects "forcing" some package standards.

Forecast:tornado requires the following packages:

• Tornado 2.4 or newer
• Blinker 1.2 or newer

Install from sources:

$ python setup.py install

Or from PyPI (using pip):

$ pip install forecast

You can install the current development release:

$ pip install -e git+https://github.com/osantana/forecast.git#egg=forecast-dev

You can run forecast:tornado tests with the following command:

$ python setup.py test

A forecast:tornado project uses the following terminology:

project

A forecast:tornado project is a directory containing a manage.py script file and a settings package containing the project configuration for various environments.

environment

A project can run in several environments like base, production, test, integration, dev, etc. You can define a set of settings for each environment. When you start a forecast:tornado project it will load the default settings from forecast:tornado package, then these settings will be overriden by settings contained in settings.base module of your project and, at the end, by the settings in settings.{ENVIRONMENT} module.

application

An application is a standard Python package containing an Application class in it __init__.py. You need to specify which applications you will use in environment's settings.

routes

An URL mapping containing routes in format URLs -> tornado.web.RequestHandler

middleware

A forecast:tornado application could install middlewares in a project. The middlewares are executed before and after the RequestHandler.method() associated with the requested ÙRL. They are useful for implementing orthogonal functionality in projects (eg. database connections).

signals

Forecast:tornado uses Blinker library to send and handle signals in a project.

Let's start a project from scratch. Our project will be called rainning and it will return a JSON-serialized object of the weather forecast for 'today' or 'tomorrow'.

Before we start our project lets create a virtualenv:

$ virtualenv rainning
$ cd rainning
$ source bin/activate
(rainning) $

Now we can install the development version of forecast:tornado:

$ pip install -e git+https://github.com/osantana/forecast.git#egg=forecast-dev
... installing forecast and its dependencies ...

Forecast:tornado it creates a script called forecast that we can use to start our project:

$ forecast start --skel=project rainning
$ ls rainning
__init__.py      manage.py        requirements.txt settings

Now we've a basic project to start with:

$ cd rainning
$ python manage.py runserver

You can point your browse to http://localhost:8888/ and you can see a random text about the 'weather'.

Now it's time to start a application where we'll handle the requests:

$ python manage.py start --skel=application willitrain
$ ls willitrain
__init__.py    controllers.py routes.py

We've created an application based on a skel bundled with forecast:tornado. It's time to install this application in our project. Let's add the following line in our settings/base.py file:

SETTINGS = {
    "name": "settings.base",
    "applications": (
        "willitrain",
    ),
    "location": 455827, # We will use this code below. (Sao Paulo, BR - http://woeid.rosselliot.co.nz/lookup)
}

We can install an application adding the package name in applications settings (above) or we can specify a tuple containing the applications' package name and a base_url that will be prepended to every route defined by our application:

SETTINGS = {
    "name": "settings.base",
    "applications": (
        ("willitrain", r"/willitrain"),
    ),
    "location": 455827, # We will use this code below. (Sao Paulo, BR - http://woeid.rosselliot.co.nz/lookup)
}

Let's define a default route for "/willitrain/" URL in willitrain/routes.py:

from forecast.routes import Route
from .controllers import IndexHandler

routes = (
    Route(r"^(?P<when>today|tomorrow)/?$", IndexHandler, name="index"),
)

Now, we will implement a handler that will get the forecast for a location (specified in settings) using Yahoo! Weather API:

from xml.dom import minidom
from datetime import datetime, timedelta
from urllib import urlencode

from tornado.web import HTTPError, asynchronous
from tornado.httpclient import AsyncHTTPClient
from tornado.gen import engine, Task

from forecast.web import RequestHandler


class IndexHandler(RequestHandler):
    @asynchronous
    @engine
    def get(self, when):
        date = datetime.now()
        if when == "tomorrow":
            date += timedelta(1)

        date_string = datetime.strftime(date, "%d %b %Y")
        woeid = urlencode({"w": self.settings["location"]})  # WOEID defined in settings/base.py

        # Use Yahoo! Weather API: http://developer.yahoo.com/weather/
        url = "http://weather.yahooapis.com/forecastrss?" + woeid
        http_client = AsyncHTTPClient()
        response = yield Task(http_client.fetch, url)

        xmldoc = minidom.parseString(response.body)
        elements = xmldoc.getElementsByTagName("yweather:forecast")

        for element in elements:
            if element.getAttribute("date") == date_string:
                self.write({"date": date_string, "text": element.getAttribute("text"), "code": element.getAttribute("code")})
                break

        else:  # date not found
            self.write({"date": date_string, "text": "not available", "code": "3200"})

        self.finish()

We use the Tornado's asynchronous API to avoid block the Web server until we get the response from Yahoo! Weather API.