def acquire_chapter(self, url, story=None):
"""Download and scrape a single chapter from a story.
@param url: The URL of the chapter to download.
@param story: The Story object provided by a previous run.
@type url: str
@type story: L{Story}
@return: A tuple consisting of the newly-created L{Chapter} object and
the L{Story} object which is either newly created or was provided
as an argument.
@rtype: (L{Chapter}, L{Story})
"""
# Verify the URL's syntactical validity before wasting bandwidth.
if not self.story_url_re.match(url):
prnt("Not a %s story URL: %s" % (self.site_name, url))
return None
# Retrieve the raw chapter (don't keep the un-parsed HTML wasting memory)
# .parse(handle) for proper encoding detection.
# .urlopen for customizing the User-Agent header.
dom = self.http.get_dom(url)
html.make_links_absolute(dom, copy=False)
chapter_select = dom.find(self.chapter_select_xpath)
chapter_content = dom.find(self.chapter_content_xpath)
if not story:
author = ''
for elem in dom.iterfind('.//a[@href]'):
if self.author_url_fragment in elem.get('href'):
author = elem.text
break
story = Story(self.get_story_title(dom), author)
story.site_name = self.site_name
story.category = self.get_story_category(dom)
if chapter_select is not None:
options = chapter_select.findall(".//option")
if options[0].text.strip().lower() in self.not_chapters:
options = options[1:]
story.chapter_urls = [self.resolve_chapter_url(x.get('value'), url, dom) for x in options]
else:
story.chapter_urls = [url]
cleaned = self.custom_content_cleaning(chapter_content)
if cleaned is not None:
chapter_content = cleaned
# Extract metadata from the chapter selector (or recognize its absence)
if chapter_select is not None:
chapter_title_str = chapter_select.find(".//option[@selected]").text
chapter_title_obj = self.chapter_title_re.match(chapter_title_str)
chapter_title = chapter_title_obj.group('name')
chapter_number = int(chapter_title_obj.group('num'))
chapter = Chapter(chapter_number, chapter_title, chapter_content)
else:
chapter = Chapter(1, '', chapter_content)
return chapter, story
def grab_urls(content, url):
urls = {}
domain = urlparse(url).netloc
html = document_fromstring(content)
html.make_links_absolute(url, resolve_base_href=True)
for element, attribute, link, pos in html.iterlinks():
if attribute != "href":
continue
# skip if not on our domain
if urlparse(link).netloc != domain and urlparse(
link).netloc != "www." + domain:
continue
# skip if self referential
if (url.split("//")[1] + "#") in link:
continue
text = element.text_content() if len(
element) == 0 else element[0].text_content()
text = text.lstrip() if text is not None else ""
# compute relevancy here
relevance[link] = relevancy(link, text, url)
urls[link] = 1
if text != "":
print text
print link
print
return urls.keys()
def get_realtime_title(pages=5):
"""
Get ALL Category Realtime news from libertytimes
realtime url may change or invaild when it is not **realtime**
get_realtime_title(pages=5, encoding="UTF-8")
*pages*: get page 1 to pages, default is 5 pages
return: dict{time, title, url}
"""
result_list = []
for page in xrange(1, pages + 1):
response, content = h.request("%s&ipage=%d" % (news_list_url, page))
html = lxml.html.fromstring(content.decode("utf-8", "ignore"))
html.make_links_absolute(base_url)
# Get news-list section
div = html.findall("*div")[0]
# Get all title-info to list
tr = list(div.iterdescendants("tr"))[1:-1]
for title_info in tr:
news_url = list(title_info.iterlinks())[1][2]
info_list = map(lambda x: x.text_content(), list(title_info))
try:
info_dict = {"title": info_list[0].strip("\r\n "), "time": info_list[1], "url": news_url}
except IndexError, error_infomation:
pass
result_list.append(info_dict)
def scrape(html: str, base_url: str) -> List[dict]:
"""
extracting a book information with the standard expression from the HTML which is given by the parameter HTML
the parameter base_url specifies the URK which is to be the standard when converted to the absolute URL
returns: the lost of book (dict)
"""
books = []
html = lxml.html.fromstring(html)
html.make_links_absolute(base_url) #convered all hrefs in the a element
#by cssselect() method, getting all a elemetns applied to the selctet and prcoessing each a element
#the meaning of selector: the a element with "itemprop=url" passed down from the li element which is also passed forn from "id-listBook"
for a in html.cssselect('#listBook > li > a[itemprop="url"]'):
# getting the URL from the href of a element
url = a.get('href')
# getting the title of the book from the itemprop="title" in the p element
p = a.cssselect('p[itemprop="name"]')[0]
title = p.text_content(
) #since it contains a wbr element, use text_content() not text_content
books.append({'url': url, 'title': title})
return books
def make_html_obj_from_link(link, mode="desktop"):
import urlparse
import requests
from lxml.html import make_links_absolute
if mode == "mobile":
try:
response = requests.get(
link,
headers={"User-Agent": "blahblah iPhone", "Accept-Language": "ru-RU,ru;q=0.8,en-US;q=0.5,en;q=0.3"},
)
except requests.exceptions.ConnectionError:
print "link " + link + " unreachable"
return
domain = "{uri.scheme}://{uri.netloc}/".format(uri=urlparse.urlparse(response.url))[:-1]
body = make_links_absolute(response.content, domain)
else:
try:
response = requests.get(link, headers={"Accept-Language": "ru-RU,ru;q=0.8,en-US;q=0.5,en;q=0.3"})
except requests.exceptions.ConnectionError:
print "link " + link + " unreachable"
return
domain = "{uri.scheme}://{uri.netloc}/".format(uri=urlparse.urlparse(response.url))[:-1]
body = make_links_absolute(response.content.decode("cp1251"), domain)
from lxml import html
parsed_body = html.fromstring(body)
return parsed_body
def html_malware_scan(url, topic):
"""
Crawls a page depth 1, returns the total count for each of
the keywords in topic for each of the pages in the crawl.
"""
response = requests.get(url, timeout=10.0)
html = lxml.html.fromstring(
response.content,
base_url=response.url
)
html.make_links_absolute(resolve_base_href=True)
childs_topic_cnt = 0
main_page_topic_cnt = response.text.lower().count(topic)
for url in Bar().iter({link for element, attribute, link, pos in html.iterlinks()}):
#for url in {link for element, attribute, link, pos in html.iterlinks()}:
childs_topic_cnt += check_content(url, topic)
if (float(main_page_topic_cnt)/(float(childs_topic_cnt)+1.0) >= 1.0 or
float(main_page_topic_cnt)/(float(childs_topic_cnt)+1.0) == 0):
return True
else:
return False
def scrape_restaurant_data(self, example):
# get this from yelp
html = obtain_html(example["url"])
html.make_links_absolute(example["url"])
title = html.cssselect("h1.biz-page-title")[0].text.strip()
review_highlights = html.cssselect("ul.review-highlights-list")
if len(review_highlights) > 0:
description = tree_to_str(clean_up_highlights(review_highlights[0]))
else:
description = create_description_highlights(html)
images = html.cssselect("img.photo-box-img")
image_url = None
if len(images) > 0:
image_url = images[0].attrib["src"]
return {
"title": title,
"description": description,
"categories": example["categories"],
"image_url" : image_url,
"rating": rating_to_string(example["rating"]),
"price": example["price"]
}
def _read_url(url):
log.info('Reading url %s' % url)
u = urllib.urlopen(url)
data = u.read()
html = lxml.html.fromstring(data)
html.make_links_absolute(url)
return data, html
def html_malware_scan(url, topic):
"""
Crawls a page depth 1, returns the total count for each of
the keywords in topic for each of the pages in the crawl.
"""
response = requests.get(url, timeout=10.0)
html = lxml.html.fromstring(response.content, base_url=response.url)
html.make_links_absolute(resolve_base_href=True)
childs_topic_cnt = 0
main_page_topic_cnt = response.text.lower().count(topic)
for url in Bar().iter(
{link
for element, attribute, link, pos in html.iterlinks()}):
#for url in {link for element, attribute, link, pos in html.iterlinks()}:
childs_topic_cnt += check_content(url, topic)
if (float(main_page_topic_cnt) / (float(childs_topic_cnt) + 1.0) >= 1.0
or float(main_page_topic_cnt) /
(float(childs_topic_cnt) + 1.0) == 0):
return True
else:
return False
def scrape_sub(self, chamber, term, district, sub_url):
"Scrape basic info for a legislator's substitute."
with self.urlopen(sub_url) as page:
html = lxml.html.fromstring(page)
html.make_links_absolute(sub_url)
# substitute info div#MAINS35
div = html.xpath('//div[contains(@id, "MAINS")]')[0]
leg = {}
leg['img_url'] = div[0][0].get('src')
subfor = div[1][0].text.replace(u'\xa0', ' ').replace(': ', '')
full_name = div[1][2].text.replace(u'\xa0', ' ')
party = _PARTY[div[1][2].tail.strip()]
leg['contact_form'] = div[1][3].xpath('string(a/@href)')
leg = Legislator(term, chamber, district.strip(), full_name, party,
**leg)
leg['roles'][0] = {
'chamber': chamber,
'state': self.state,
'term': term,
'role': 'substitute',
'legislator': subfor[subfor.rindex('for'):],
'district': district.replace('District', '').strip(),
'party': party,
'start_date': None,
'end_date': None
}
leg.add_source(sub_url)
self.save_legislator(leg)
def get_realtime_title():
"""Get ALL Category and Source Realtime news from chinatimes
realtime url may change or invaild when it is not *realtime*
return: dict{category, source, time, title, url}
"""
response, content = h.request(news_list_url)
html = lxml.html.fromstring(content.decode('big5', 'ignore'))
html.make_links_absolute(base_url)
# Get news-list section
div = html.findall("*div")[1]
# Get all title-info to list
tr = list(div.iterdescendants("tr"))[1:]
result_list = []
for title_info in tr:
news_url = list(title_info.iterlinks())[0][2]
info_list = map(lambda x: x.text_content(), list(title_info))
info_dict = {"title": info_list[0].strip("\r\n "), "time": info_list[1],
"category": info_list[2], "source": info_list[3],
"url": news_url}
result_list.append(info_dict)
return result_list
def scrape_list_page(response: requests.Response) -> Iterator[str]:
html = lxml.html.fromstring(response.text)
html.make_links_absolute(response.url)
for a in html.cssselect('#listBook > li > a[itemprop="url"]'):
url = a.get('href')
yield url
def scrape_list_page(response: requests.Response) -> Iterator[str]:
html = lxml.html.fromstring(response.text)
html.make_links_absolute(response.url)
for a in html.cssselect('table.cassetteitem_other > tbody > tr > td:nth-child(9) > a'):
url = a.get('href')
yield url
def get_oracle(url):
source = util.get_source(url)
html = lxml.html.document_fromstring(source)
html.make_links_absolute(url, resolve_base_href=True)
util.save_file(lxml.html.tostring(html), 'oracle.html')
util.screenshot('oracle.html', 'oracle.png')
return html
def scrape(self, chamber, term):
"""
Scrapes legislators for the current term only
"""
self.validate_term(term, latest_only=True)
url = _BASE_URL % _CHAMBERS[chamber].lower()
with self.urlopen(url) as index:
html = lxml.html.fromstring(index)
html.make_links_absolute(url)
base_table = html.xpath('body/table/tr/td[2]/table[2]')
district = None # keep track of district for substitutes
for row in base_table[0].xpath('tr'):
img_url = row.xpath('string(.//img/@src)')
contact_form, additional_info_url = row.xpath('.//a/@href')
if "Substitute" in row.text_content():
# it seems like the sub always follows the person who he/she
# is filling in for.
# most sub info is provided at the additional info url
self.scrape_sub(chamber, term, district, additional_info_url)
continue
else:
full_name = " ".join(row[1][0].text_content().replace(u'\xa0', ' ').split())
party = _PARTY[row[1][0].tail.strip()]
pieces = [ x.strip() for x in row.itertext() if x ][6:]
# the first index will either be a role or the district
role = None
if 'District' in pieces[0]:
district = pieces.pop(0)
else:
role = pieces.pop(0)
district = pieces.pop(0)
leg = Legislator(term, chamber,
district.replace('District', '').strip(),
full_name,
party=party)
leg.add_source(url)
leg['photo_url'] = img_url
leg['contact_form'] = contact_form
leg['url'] = additional_info_url
leg['address'] = pieces.pop(0)
# at this point 'pieces' still contains phone numbers and profession
# and committee membership
# skip committee membership, pick 'em up in IDCommitteeScraper
end = -1
if 'Committees:' in pieces:
end = pieces.index('Committees:')
for prop in pieces[:end]:
# phone numbers
if prop.lower()[0:3] in _PHONE_NUMBERS:
leg[ _PHONE_NUMBERS[ prop.lower()[0:3] ] ] = prop
# profession
else:
leg['profession'] = prop
self.save_legislator(leg)
def resp2html(resp, encoding='utf-8'):
"""将request返回的response转换为经lxml解析后的document"""
text = get_resp_text(resp, encoding=encoding)
html = lxml.html.fromstring(text)
html.make_links_absolute(resp.url, resolve_base_href=True)
# html = cleaner.clean_html(html)
# lxml.html.open_in_browser(html, encoding=encoding) # for develop and debug
return html
def scraperwiki():
for page_number in itertools.count(1):
url = 'https://classic.scraperwiki.com/profiles/tlevine/?page=%d' % page_number
response = get(url)
html = lxml.html.fromstring(response.text)
html.make_links_absolute(url)
for href in html.xpath('//li[@class="code_object_line"]/descendant::h3/a[position()=2]/@href'):
yield re.sub(r'index.html$', '', str(href))
def scrape_joint_committees(self):
page = self.get(_JOINT_URL).text
html = lxml.html.fromstring(page)
html.make_links_absolute(_JOINT_URL)
joint_li = html.xpath('//div[contains(h2, "Joint")]/ul/li')
for li in joint_li:
name, url = li[0].text, li[0].get('href')
self.get_joint_committees_data(name, url)
def get_oracle(url): source = util.get_source(url) parser = lxml.etree.HTMLParser() etree = lxml.etree.parse(StringIO(source), parser) html = lxml.html.document_fromstring(source) html.make_links_absolute(url, resolve_base_href=True) html.doctype = etree.docinfo.doctype return html
def get_oracle(url):
source = util.get_source(url)
parser = lxml.etree.HTMLParser()
etree = lxml.etree.parse(StringIO(source), parser)
html = lxml.html.document_fromstring(source)
html.make_links_absolute(url, resolve_base_href=True)
html.doctype = etree.docinfo.doctype
return html
def _load_response(response):
'''
In: python-requests Response
Out: lxml HTML tree
'''
html = lxml.html.fromstring(response.text)
html.make_links_absolute(response.url)
return html
def europeana():
url = request.args['url']
hl = request.args.get('hl', '')
words = re.findall(r'\w+', hl.lower())
html = lxml.html.parse(url).getroot()
html.make_links_absolute()
highlight_tree(html, words)
return lxml.html.tostring(html)
def get_page(url):
global fetched_count
fetched_count += 1
print 'Fetching... %s %s' % (fetched_count, url)
doc = lxml.html.parse(urllib.urlopen(url))
html = doc.getroot()
html.make_links_absolute(url)
return html
def get_links(html, base_url, tags = []): links = [] tags = tags html = lxml.html.document_fromstring(html) html.make_links_absolute(base_url) links_html = html.iterlinks() links = [ x[2] for x in links_html if x[0].tag in tags ] return links
def scrape_joint_committees(self):
page = self.get(_JOINT_URL).text
html = lxml.html.fromstring(page)
html.make_links_absolute(_JOINT_URL)
joint_li = html.xpath('//div[contains(h2, "Joint")]/ul/li')
for li in joint_li:
name, url = li[0].text, li[0].get('href')
self.get_joint_committees_data(name, url)
def get_links(html, base_url, tags=[]):
links = []
tags = tags
html = lxml.html.document_fromstring(html)
html.make_links_absolute(base_url)
links_html = html.iterlinks()
links = [x[2] for x in links_html if x[0].tag in tags]
return links
def make_tree(url, data, encoding=None):
"""Build lxml tree."""
s = StringIO(data)
parser = lxml.html.HTMLParser(encoding=encoding)
doc = lxml.html.parse(s, parser=parser)
html = doc.getroot()
html.make_links_absolute(url)
return html
def scrape_chamber(self, chamber):
"""
Scrapes legislators for the current term only
"""
# self.validate_term(term, latest_only=True)
url = BASE_URL % CHAMBERS[chamber].lower()
index = self.get(url).text
html = lxml.html.fromstring(index)
html.make_links_absolute(url)
rows = html.xpath('//div[contains(@class, "row-equal-height")]')
for row in rows:
img_url = row.xpath('.//img/@src')[0]
inner = row.xpath('.//div[@class="vc-column-innner-wrapper"]')[1]
inner_text = inner.text_content()
if 'Resigned' in inner_text or 'Substitute' in inner_text:
continue
name = inner.xpath('p/strong')[0].text.replace(u'\xa0', ' ').strip()
name = re.sub(r'\s+', ' ', name)
party = PARTY[inner.xpath('p/strong')[0].tail.strip()]
email = inner.xpath('p/strong/a')[0].text
district = inner.xpath('p/a')[0].text.replace('District ', '')
person_url = inner.xpath('p/a/@href')[0]
# skip roles for now
role = ''
# for com in inner.xpath('p/a[contains(@href, "committees")]'):
# role = com.tail.strip()
person = Person(name=name, district=district,
party=party, primary_org=chamber,
image=img_url, role=role)
phones = get_phones(inner)
phone = phones.get('home') or phones.get('business')
office_phone = phones.get('office')
address = get_address(inner)
fax = get_fax(inner)
if address:
person.add_contact_detail(type='address', value=address,
note='District Office')
if phone:
person.add_contact_detail(type='voice', value=phone,
note='District Office')
if fax:
person.add_contact_detail(type='fax', value=fax,
note='District Office')
if email:
person.add_contact_detail(type='email', value=email,
note='District Office')
if office_phone:
person.add_contact_detail(type='voice', value=office_phone,
note='Capitol Office')
person.add_source(url)
person.add_link(person_url)
yield person
def scrape_chamber(self, chamber):
"""
Scrapes legislators for the current term only
"""
# self.validate_term(term, latest_only=True)
url = BASE_URL % CHAMBERS[chamber].lower()
index = self.get(url, verify=False).text
html = lxml.html.fromstring(index)
html.make_links_absolute(url)
rows = html.xpath('//div[contains(@class, "row-equal-height")]')
for row in rows:
img_url = row.xpath('.//img/@src')[0]
inner = row.xpath('.//div[@class="vc-column-innner-wrapper"]')[1]
inner_text = inner.text_content()
if 'Resigned' in inner_text or 'Substitute' in inner_text:
continue
name = inner.xpath('p/strong')[0].text.replace(u'\xa0', ' ').strip()
name = re.sub('\s+', ' ', name)
party = PARTY[inner.xpath('p/strong')[0].tail.strip()]
email = inner.xpath('p/strong/a')[0].text
district = inner.xpath('p/a')[0].text.replace('District ', '')
person_url = inner.xpath('p/a/@href')[0]
# skip roles for now
role = ''
# for com in inner.xpath('p/a[contains(@href, "committees")]'):
# role = com.tail.strip()
person = Person(name=name, district=district,
party=party, primary_org=chamber,
image=img_url, role=role)
phones = get_phones(inner)
phone = phones.get('home') or phones.get('business')
office_phone = phones.get('office')
address = get_address(inner)
fax = get_fax(inner)
if address:
person.add_contact_detail(type='address', value=address,
note='District Office')
if phone:
person.add_contact_detail(type='voice', value=phone,
note='District Office')
if fax:
person.add_contact_detail(type='fax', value=fax,
note='District Office')
if email:
person.add_contact_detail(type='email', value=email,
note='District Office')
if office_phone:
person.add_contact_detail(type='voice', value=office_phone,
note='Capitol Office')
person.add_source(url)
person.add_link(person_url)
yield person
def create_html_parser(row):
content = get_content(row.value)
parser = lxml.html.HTMLParser(encoding='utf-8')
html = lxml.html.document_fromstring(content, parser=parser)
if 'url' in row.value:
html.make_links_absolute(row.value['url'])
elif row.key:
html.make_links_absolute(row.key)
return html
def parse_html(self, url=None, response=None, *a, **ka):
if lxml is None:
raise ImportError('No module named \'lxml\'')
if url:
response = self.get(url, *a, **ka)
elif response is None:
raise ValueError('either url or response must be provided')
html = lxml.html.fromstring(response.text)
html.make_links_absolute(response.url)
return html
def thomaslevine():
url = 'http://thomaslevine.com/!/'
try:
response = get(url)
except:
return []
else:
html = lxml.html.fromstring(response.text)
html.make_links_absolute(url)
return ((unicode(a.attrib['href']), a.text_content()) for a in html.xpath('//a') if a.attrib['href'].startswith(url))
def get_pages():
response_text = requests.get(BASE_URL).content
html = lxml.html.fromstring(response_text)
html.make_links_absolute(BASE_URL)
links = []
i = 0
for link in html.iterlinks():
links.append(link[2])
return links
def get_html(url):
"""使用get方法访问指定网页并返回经lxml解析后的document"""
resp = request_get(url)
text = get_resp_text(resp, encoding='utf-8')
html = lxml.html.fromstring(text)
html.make_links_absolute(url, resolve_base_href=True)
# 清理功能仅应在需要的时候用来调试网页（如prestige），否则可能反过来影响调试（如JavBus）
# html = cleaner.clean_html(html)
# lxml.html.open_in_browser(html, encoding=encoding) # for develop and debug
return html
def scrape_list_page(response: requests.Response):
"""
一覧ページのResponseから詳細ページのURLを抜き出すジェネレーター関数
"""
html = lxml.html.fromstring(response.text)
html.make_links_absolute(response.url)
for a in html.cssselect('#listBook > li > a[itemprop="url"]'):
url = a.get('href')
yield url
def scrape(html: str, base_url: str) -> List[dict]:
books = []
html = lxml.html.fromstring(html)
html.make_links_absolute(base_url)
for a in html.cssselect('#listBook > li > a[itemprop="url"]'):
url = a.get('href')
p = a.cssselect('p[itemprop="name"]')[0]
title = p.text_content()
books.append({'url': url, 'title': title})
return books
def addImagePaths(self, body, baseurl):
"""Turn relative paths into absolute paths in section body"""
# This is a convenience method for use in referencemanual_macros
# section_collation macro.
# TODO: when we not longer need 2.1 compatibility, this belongs in
# a view.
html = lxml.html.fromstring(safe_unicode(body))
html.make_links_absolute(baseurl, resolve_base_href=False)
return lxml.html.tostring(html, encoding='utf-8')
def addImagePaths(self, body, baseurl):
"""Turn relative paths into absolute paths in section body"""
# This is a convenience method for use in referencemanual_macros
# section_collation macro.
# TODO: when we not longer need 2.1 compatibility, this belongs in
# a view.
html = lxml.html.fromstring(safe_unicode(body))
html.make_links_absolute(baseurl, resolve_base_href=False)
return lxml.html.tostring(html, encoding='utf-8')
def gethtml(url):
raw = get(url).content.decode('utf-8')
if raw == '':
return None
try:
html = lxml.html.fromstring(raw)
except Exception as e:
print(url)
raise e
html.make_links_absolute(url)
return html
def thomaslevine():
url = 'http://thomaslevine.com/!/'
try:
response = get(url)
except:
return []
else:
html = lxml.html.fromstring(response.text)
html.make_links_absolute(url)
return ((unicode(a.attrib['href']), a.text_content())
for a in html.xpath('//a') if a.attrib['href'].startswith(url))
def useWebsiteSearch (self, url, target):
try:
page = urllib2.urlopen (url)
except (urllib2.URLError, ValueError):
return None
s = page.read()
html = lxml.html.fromstring (s)
html.make_links_absolute (baseURL)
html.forms[0].inputs['query'].value = target
new_page = lxml.html.submit_form(html.forms[0])
new_s = new_page.read()
return new_s
def useWebsiteSearch(self, url, target):
try:
page = urllib2.urlopen(url)
except (urllib2.URLError, ValueError):
return None
s = page.read()
html = lxml.html.fromstring(s)
html.make_links_absolute(baseURL)
html.forms[0].inputs['query'].value = target
new_page = lxml.html.submit_form(html.forms[0])
new_s = new_page.read()
return new_s
def parse_foi(response):
html = lxml.html.fromstring(response.text)
html.make_links_absolute(response.url)
messages = ['\n\n'.join(div.xpath('strong/text()') + div.xpath('p/text()')) \
for div in html.xpath('id("tabs-request")/div')]
downloads = list(map(str, html.xpath('id("tabs")//a[text()="Download"]/@href')))
return OrderedDict([
('messages', messages),
('downloads', downloads),
])
def scrape_list_page():
"""
一覧ページのResponseから詳細ページのURLを抜き出すジェネレーター関数
"""
html = lxml.html.fromstring(response.text)
# 絶対URLに変換する
html.make_links_absolute(response.url)
for a in html.cssselect('#listBook > li > a[itemprop="url"]'):
url = a.get('href')
# yield文でジェネレーターイテレーターの要素を返す
yield url
def _extract_links(self, response_text, response_url):
html = lxml.html.fromstring(response_text)
html.make_links_absolute(response_url)
for e, a, l, p in html.iterlinks():
if self.scan_tag(e.tag):
if self.scan_attr(a):
link = Link(self.process_attr(l), text=e.text)
self.links.append(link)
links = unique_list(self.links, key=lambda link: link.url) if self.unique else self.links
return links
def download_councillors():
with open(WEBPAGESTXT, 'r') as txtfile:
urls = txtfile.readlines()
urls = [url.strip() for url in urls]
session = http.client.HTTPSConnection('www.berlin.de', timeout=10)
councillors = {}
for url in urls:
if councillors:
time.sleep(2)
bezirk = bezirk_from_url(url)
headers = {'Accept-Encoding': 'gzip', 'Connection': 'keep-alive'}
session.request('GET', url, headers=headers)
response = session.getresponse()
response = response.read()
response = zlib.decompress(response, 47)
try:
response = response.decode('latin-1', 'strict')
except UnicodeDecodeError:
response = response.decode('windows-1252', 'strict')
html = lxml.html.fromstring(response)
html.make_links_absolute(url)
tablerows = html.cssselect('.zl12')
tablerows += html.cssselect('.zl11')
number = html.cssselect('table.tk1:nth-child(8)')[0]
number = number.text_content()
_, number = number.split(':')
number = number.strip()
if number.isdigit():
number = int(number)
if not number == len(tablerows):
print('%s:' % bezirk,
'%s councillors were found.' % len(tablerows),
'Should be %s councillors.' % number)
for row in tablerows:
councillor = extract_councillor(row)
councillor['BEZIRK'] = bezirk
identifier = normalized_name(councillor['ANZEIGENAME'])
try:
councillors[bezirk][identifier] = councillor
except KeyError:
councillors[bezirk] = {identifier: councillor}
session.close()
return councillors
def npm():
url = 'https://www.npmjs.org/~tlevine'
try:
response = get(url)
except:
sys.stderr.write('Error loading NPM packages\n')
else:
html = lxml.html.fromstring(response.text)
html.make_links_absolute(url)
for li in html.xpath('id("profile")/ul/li'):
href = li.xpath('a/@href')[0]
text = ''.join(li.xpath('text()')).strip()
yield unicode(href), unicode(text)
def getCandidats (self, s):
html = lxml.html.fromstring (s)
html.make_links_absolute (baseURL)
#lxml.html.open_in_browser(html)
names = list()
links = list()
for name in html.xpath('//@alt'):
names.append(name)
for a in html.xpath('//a'):
if a.attrib.get('title') in names:
links.append (a.attrib.get('href'))
return (names, links)
def npm():
url = 'https://www.npmjs.org/~tlevine'
try:
response = get(url)
except:
sys.stderr.write('Error loading NPM packages\n')
else:
html = lxml.html.fromstring(response.text)
html.make_links_absolute(url)
for li in html.xpath('id("profile")/ul/li'):
href = li.xpath('a/@href')[0]
text = ''.join(li.xpath('text()')).strip()
yield unicode(href), unicode(text)
def extract_links_from_html(base, body):
try:
html = lxml.html.fromstring(body)
html.make_links_absolute(base)
for element, attribute, link, pos in html.iterlinks():
if isinstance(link, str):
link = link.encode('utf-8', 'ignore')
yield link
except Exception:
logging.warning("(lxml) html parse error")
import traceback; traceback.print_exc()
def bootstrap_data():
data = urllib.urlopen(DATA_URL).read()
#open("localcache", 'wb').write(data)
print '='*20
#data = open("localcache").read()
html = lxml.html.fromstring(data)
html.make_links_absolute(DATA_URL)
if len(html.xpath('//td[@class="maintext"]/ul/li/a')) != 15:
logging.error('Something changed/new? %s elements on first page.' % len(html.xpath('//td[@class="maintext"]/ul/li/a')))
for a in html.xpath('//td[@class="maintext"]/ul/li/a'):
if handlers.get(a.text):
handlers.get(a.text)(a.attrib['href'])
else:
print [a.attrib['href'], a.text]
def _extract_links(self, response_text, response_url):
html = lxml.html.fromstring(response_text)
html.make_links_absolute(response_url)
sel = pyquery.PyQuery(html)
evt_links = sel('.news > li:not(.more) > a')
ann_links = sel('.announcement > li:not(.more) > a')
all_links = [
Link(elem.attrib['href'], text=elem.text)
for elem in itertools.chain(evt_links, ann_links)
]
return unique(all_links, key=lambda link: link.url)
def scraperwiki(url = 'https://classic.scraperwiki.com/profiles/tlevine/index.html'):
try:
response = get(url)
except:
pass
else:
html = lxml.html.fromstring(response.text)
html.make_links_absolute(url)
for href in html.xpath('//li[@class="code_object_line"]/descendant::h3/a[position()=2]/@href'):
yield re.sub(r'index.html$', '', unicode(href)), 'A web scraper'
nexts = html.xpath(u'//a[text()="Next »"]/@href')
if nexts != []:
for scraper in scraperwiki(nexts[0]):
yield scraper
def get_uri(self, url, html):
if url is not None and html is not None:
print(url)
parsed_uri = urlparse(url)
domain = '{uri.scheme}://{uri.netloc}/'.format(uri=parsed_uri)
html.make_links_absolute(url)
for l in html.iterlinks():
parsed_uri = urlparse(l[2])
curr_domain = '{uri.scheme}://{uri.netloc}/'.format(uri=parsed_uri)
if curr_domain == domain:
if l[2] not in self.urls:
self.pool.put(l[2])
self.urls.add(l[2])
