zester 0.0.3

Zester是一个库，它使开发没有API的网站Python客户端变得更容易。

没有lxml，没有XPath，只有javascript。

让我们通过将以下代码保存到名为hnclient.py的文件中来为Hacker News创建一个客户端库

from zester import MultipleClient, Attribute

class HNClient(MultipleClient):
    url = "http://news.ycombinator.com/"
    title = Attribute(selector="$('.title a')", modifier="$(el).html()")
    link = Attribute(selector="$('.title a')"), modifier="$(el).attr('href')")
    points = Attribute(selector="$('.subtext span')", modifier="$(el).html().replace(' points', '')")

现在，让我们使用我们刚刚创建的客户端。打开Python外壳

>>> from hnclient import HNClient
>>> client = HNClient()
>>> stories = client.process()
>>> stories[0]
HNClientResponse(points=u'200', link=u'http://daltoncaldwell.com/what-twitter-could-have-been', title=u'What Twitter could have been')
>>> print stories[0].title
What Twitter could have been
>>> print stories[0].link
http://daltoncaldwell.com/what-twitter-could-have-been
>>> print stories[0].points
56

我们那里继承自MultipleClient，因为我们计划返回多个结果。如果我们想为像Weather.gov这样的返回单个结果的项目创建客户端，我们可以这样做

from zester import SingleClient, Attribute

class WeatherClient(SingleClient):
    url = "http://forecast.weather.gov/MapClick.php?lat={lat}&lon={lng}"
    temperature = Attribute(selector="$('.myforecast-current-lrg').html()")
    humidity = Attribute(selector="$('.current-conditions-detail li').contents()[1]")
    heat_index = Attribute(selector="$('.current-conditions-detail li').contents()[11]")

    def __init__(self, lat, lng, *args, **kwargs):
        super(WeatherClient, self).__init__(*args, **kwargs)
        self.url = self.url.format(lat=lat, lng=lng)

这也展示了您如何允许接受参数

>>> from weather_client import WeatherClient
>>> client = WeatherClient(lat=40.7143528, lng=-74.0059731)
>>> curr_weather = client.process()
>>> curr_weather
WeatherClientResponse(heat_index=u'82\xb0F (28\xb0C)', temperature=u'80\xb0F', humidity=u'58%')
>>> print curr_weather.temperature
80°F
>>> print curr_weather.humidity
58%
>>> print curr_weather.heat_index
82°F (28°C)

$ pip install zester