From 34866e9c48579de697dbe99efd8bc6bf72707f2e Mon Sep 17 00:00:00 2001 From: Kenneth Reitz Date: Sat, 24 Feb 2018 16:08:50 -0500 Subject: [PATCH] looking nice Signed-off-by: Kenneth Reitz --- requests_html.py | 17 +++++++++++++++-- 1 file changed, 15 insertions(+), 2 deletions(-) diff --git a/requests_html.py b/requests_html.py index 89cfb57..af377cc 100644 --- a/requests_html.py +++ b/requests_html.py @@ -1,6 +1,8 @@ import requests from pyquery import PyQuery +from lxml.etree import tostring + class Element: """docstring for Element""" def __init__(self, element): @@ -27,6 +29,18 @@ class Element: def text(self): return self.pq.text() + @property + def html(self): + return tostring(self.element).decode('utf-8').strip() + + def find(self, selector): + def gen(): + for found in self.pq(selector): + yield Element(found) + + return [g for g in gen()] + + class HTML(object): """docstring for HTML""" @@ -87,7 +101,6 @@ class HTML(object): return PyQuery(self.html) - def handle_response(response, **kwargs): response.html = HTML(response) return response @@ -97,4 +110,4 @@ session = requests.Session() session.hooks = {'response': handle_response} r = session.get('https://kennethreitz.org/') -print(r.html.absolute_links) +print(r.html.find('a')[0].html)