class Page(Item):
current = Attr('div.mt20 > div.pages > a.anext', 'href')
total = Attr('div.mt20 > div.pages > a.alast', 'href')
def clean_total(self, total):
return int(total.rsplit('/', 1)[1])
def clean_current(self, current):
print(current)
return int(current.rsplit('/', 1)[1]) - 20
class Repic(Item):
img = Attr('a.cook-img', 'style')
url = Attr('a.cook-img', 'href')
title = Text('div.cook-info > a.cookname')
major = Text('div.cook-info > p.major')
def clean_img(self, img):
re_img = re.compile(
'background: url[(](.*)[)] no-repeat center center;background-size: cover;position: relative;'
)
return re_img.match(img).groups()[0]
class Page(Item):
next_page = Attr('.morelink', 'href')
def clean_next_page(self, value):
return api.convert_string(
'/' + value, '/news?p={page}',
request.host_url.strip('/') + '/posts?page={page}')
class Cookbook(Item):
title = Text('h2.title')
img = Attr('.recipe-content > div > div > div > a > img', 'src')
browse_count = Text('div.vcnum > span')
collect_count = Text('div.vcnum > span.collectnum')
intro = Text('p.intro')
tip = Text('div.tips > p')
class Page(Item):
next_page = Attr(".morelink", "href")
def clean_next_page(self, value):
return api.convert_string(
"/" + value,
"/news?p={page}",
request.host_url.strip("/") + "/posts?page={page}",
)
import requests
from htmlparsing import Element, HTMLParsing, Text, Attr, Parse
url = 'http://localhost:8082/home/serveList.html'
r = requests.get(url)
article_detail = HTMLParsing(r.text).detail({
'title':
Text('a.storylink'),
'points':
Parse('span.score', '>{} points'),
'link':
Attr('a.storylink', 'href')
})
print(article_detail)
class Post(Item):
url = Attr('.read-more', 'href')
title = Text('h1 > a')
class Post(Item):
url = Attr('.storylink', 'no this attribute')
title = Text('.storylink')
class Post(Item):
url = Attr('.storylink', 'href')
title = Text('.storylink')
class Post(Item):
url = Attr(".storylink", "href")
title = Text(".storylink")
class Post(Item):
url = Attr(".storylink", "no this attribute")
title = Text(".storylink")
class Course(Item):
url = Attr('a', 'href')
title = Text('h4')
class Step(Item):
img = Attr('img', 'src')
step = Text('p')
class Selected(Item):
title = Text('.name')
img = Attr('a > img', 'src')
url = Attr('.name', 'href')