def set_url(): # 通过 set_url() 方法设置 robots.txt 的链接。
rp = RobotFileParser()
rp.set_url('http://www.bilibili.com/robots.txt')
rp.read()
# 使用 can_fetch() 方法判断网页是否可以被抓取。
print(rp.can_fetch('*', 'http://www.bilibili.com/vide/BV15J411T7WQ'))
print(
rp.can_fetch(
'*',
'http://www.bilibili.com/search?q=python&page=1&type=collections'))
def parse(): # 使用 parse() 方法执行读取和分析
rp = RobotFileParser()
rp.parse(
urlopen('http://www.bilibili.com/robots.txt').read().decode(
'utf-8').split('\n'))
print(rp.can_fetch('*', 'http://www.bilibili.com/vide/BV15J411T7WQ'))
print(
rp.can_fetch(
'*',
'http://www.bilibili.com/search?q=python&page=1&type=collections'))
def test_robots_txt(self):
parser = RobotFileParser(self.live_server_url + '/robots.txt')
parser.read()
url = self.live_server_url + '/index.html'
for robot in self.robots:
self.assertTrue(parser.can_fetch(robot, url))
url = self.live_server_url + '/admin/'
for robot in self.robots:
self.assertFalse(parser.can_fetch(robot, url))
def filter_googlebot(entries):
""" Given a bunch of robots.txt entries, figure out if googlebot is allowed
but other random bots are banned. yields tuples of (entry, reppy.Robots) objects
that match this condition"""
for entry in entries:
if entry.status_code != 200:
continue
parser = RobotFileParser(entry.url)
parser.parse(entry.body.split("\n"))
if parser.can_fetch("GoogleBot", "/") and not parser.can_fetch("BensCoolBot", "/"):
yield entry, parser
def filter_googlebot(entries):
""" Given a bunch of robots.txt entries, figure out if googlebot is allowed
but other random bots are banned. yields tuples of (entry, reppy.Robots) objects
that match this condition"""
for entry in entries:
if entry.status_code != 200:
continue
parser = RobotFileParser(entry.url)
parser.parse(entry.body.split("\n"))
if parser.can_fetch("GoogleBot",
"/") and not parser.can_fetch("BensCoolBot", "/"):
yield entry, parser
def allowed_to_crawl(url, host_url, scheme):
'''
url: the full url, string
host_url: the domain url, string (eg. wikipedia.org)
scheme: the communication protocol, string (eg. https)
'''
# if host URL is google, assume we are allowed to crawl
if host_url == 'google.com':
return True
# if it is not a link, return False
if host_url == '' or scheme == '':
return False
try:
# get the robots.txt
rp = RobotFileParser()
rp.set_url(scheme + "://" + host_url + "/robots.txt")
rp.read()
return rp.can_fetch("*", url)
except:
pass
return True
def robotExclusion(link):
# print("Robot exclusion.....")
rp=RobotFileParser();
rp.set_url(urljoin(link, '/robot.txt'))
rp.read()
# print(rp.can_fetch("*", link))
return rp.can_fetch("*", link)
class Website:
def __init__(self, scheme: str, hostname: str):
self.scheme = scheme
self.hostname = hostname
self.last_time = 0
self._urls = set()
self._queue = deque()
# parse robots.txt
self._robot_parser = RobotFileParser()
self._robot_parser.set_url("{}://{}/robots.txt".format(
scheme, hostname))
self._robot_parser.read()
def can_fetch(self, user_agent: str, url: str) -> bool:
return self._robot_parser.can_fetch(user_agent, url)
def add_url(self, url: str, depth: int = 0):
if url not in self._urls:
self._urls.add(url)
self._queue.append((url, depth))
def get_url(self) -> (str, int):
return self._queue.popleft()
def crawl_delay(self, user_agent: str) -> int:
delay = self._robot_parser.crawl_delay(user_agent)
return delay * 300 if delay is not None else None
def is_empty(self) -> bool:
return len(self._queue) == 0
def can_fetch(url, robots_url, useragent):
from urllib.robotparser import RobotFileParser
parser = RobotFileParser()
parser.set_url(robots_url)
parser.read()
return parser.can_fetch(useragent, url)
def http_open(self, request):
url = request.get_full_url()
host = urlsplit(url)[1]
robots_url = urlunsplit(('http', host, '/robots.txt', '', ''))
rp = RobotFileParser(robots_url)
rp.read()
if not rp.can_fetch(self.agentname, url):
raise RuntimeError('Forbidden by robots.txt')
return urllib2.HTTPHandler.http_open(self, request)
class Crawler(Thread):
def __init__(self, scheduler, id):
Thread.__init__(self)
self.scheduler = scheduler
self.robot_parser = RobotFileParser()
self.running = True
self.id = id
def run(self):
global RUNNING
print(self.id,"running!")
while RUNNING and self.running:
url = self.scheduler.next()
while url is None:
sleep(2)
url = self.scheduler.next()
try:
''' Check robots.txt '''
parsed_url = urlparse(url)
robots_url = parsed_url.scheme + "://" + parsed_url.netloc + "/robots.txt"
robots_response = urlopen(robots_url, timeout=2)
robots_file = robots_response.read()
self.robot_parser.parse(robots_file.decode('utf-8').splitlines())
if not self.robot_parser.can_fetch("*", url):
#print(self.id,"is not allowed to fetch",url)
continue
''' Fetch the url '''
print(self.id,"->",url)
response = urlopen(url, timeout=2)
data = response.read().decode('utf-8', errors='ignore')
data = data.split('href="')
del data[0]
data = sorted(data, key=len)
except (HTTPError,TimeoutError,ConnectionResetError,UnicodeDecodeError,BaseException):
data = []
except URLError as e:
print(url,"->",str(e))
data = []
for d in data:
d = d[:d.find('"')]
if d.endswith('/'):
d = d[:-1]
''' Format relative URLs '''
if not d.startswith("http://"):
path = parsed_url.path
url = path[:path.rfind("/")]
while d.startswith("../"):
url = url[:url.rfind("/")]
d = d[3:]
self.scheduler.add(parsed_url.scheme + \
"://" + parsed_url.netloc + \
url + "/" + d)
else:
self.scheduler.add(d)
def get_robot_txt(cls, url):
try:
rp = RobotFileParser()
parsed_url = urlparse(url)
robots_url = "{0.scheme}://{0.netloc}/robot.txt".format(parsed_url)
rp.set_url(robots_url)
rp.read()
return rp.can_fetch("*", url)
except Exception as e:
raise Exception(e.args[0])
def check_robots_txt(self, parsed_link):
"""
Checks the site's robots.txt file to make sure our user agent is allowed to visit that url.
:param parsed_link:
:return: boolean . True if we're allowed to visit (or there's no robots.txt)
"""
rp = RobotFileParser()
rp.set_url(urljoin(parsed_link.geturl(), '/robots.txt'))
rp.read()
return rp.can_fetch(self.user_agent, parsed_link.geturl())
class Driver:
def __init__(self,
root_url: str,
header: str,
access_delay: int = 3,
cookies: dict = None,
logger=None):
self.logger = logger
self.root_url = root_url
self.cookies = cookies
self.header = header
self.access_delay = access_delay
self.now_content = None
self.robots = None
self.load_robots_txt()
def load_robots_txt(self):
self.robots = RobotFileParser()
self.robots.set_url(self.root_url + '/robots.txt')
self.robots.read()
def get(self, path):
try:
sleep(self.access_delay)
url = f'{self.root_url}/{path}'
if self.robots.can_fetch("*", url):
res = requests.get(url,
headers=self.header,
cookies=self.cookies)
if self.logger is not None:
self.logger.debug(f"Access to {url}.")
self.now_content = BeautifulSoup(res.text, 'html.parser')
else:
if self.logger is not None:
self.logger.warning(
f"Access to this url is prohibited by robots.txt.\n<*>[URL={url}]"
)
except Exception as e:
if self.logger is not None:
self.logger.warning(e)
def find_element_by_class_name(self, name):
return self.now_content.select('.' + name)[0]
def find_elements_by_class_name(self, name):
return self.now_content.select('.' + name)
def find_element_by_id(self, name):
return self.now_content.select('#' + name)[0]
def find_elements_by_id(self, name):
return self.now_content.select('#' + name)
def find_element_by_tag(self, name):
return self.now_content.find_all(name)
def check_robot_txt(url):
try:
# proxy_handler = ProxyHandler(
# {
# 'http' : 'http://127.0.0.1:63231',
# 'https' : 'http://127.0.0.1:63231'
# }
# )
#
# opener = build_opener(proxy_handler)
rp = RobotFileParser(url + '/robots.txt')
rp.read()
# rp.parse(urlopen(url.read().decode('utf-8').split('\n')))
print(rp.can_fetch('*', url + '/stock/'))
# print("***************1")
# response = opener.open('https://www.baidu.com')
# print("***************2")
# print(response.read().decode('utf-8'))
# headers = {'User-Agent': 'Mozilla / 5.0(Macintosh;Intel Mac OS X) AppleWebKit / 537.36(KHTML, like Gecko)',
# 'Host': 'httpbin.org'}
# dict = {'name':'Tobi'}
#
# username= '******'
# password= '******'
#
# url = 'http://localhost:5000'
#
# p = urllib.request.HTTPPasswordMgrWithDefaultRealm()
# p.add_password(None,url,username,password)
# auth_handler = urllib.request.HTTPBasicAuthHandler(p)
# opener = urllib.request.build_opener(auth_handler)
#
# response = opener.open(url)
# html = response.read().decode('utf-8')
# print(html)
# request = urllib.request.Request('http://httpbin.org/post', headers=headers, data=bytes(urllib.parse.urlencode(dict),encoding='utf8'), method='POST')
# print("***************")
# response = urllib.request.urlopen(request, context=ssl._create_unverified_context())
# if response.status != 200:
# print('status : ' + response.status)
# print("*******result********")
# print(response.read().decode('utf-8'))
# print(type(response))
# print(response.status)
# print(response.getheaders())
# print(response.getheader('Server'))
except URLError as e:
# if isinstance(e.reason, socket.timeout):
# print('time out!!!')
print(e.reason)
return
def can_fetch(self, url):
parsed_url = urlparse(url)
# Fetching and parsing the robots.txt file can be expensive in it-self.
# Let's cache the RobotFileParser instances, one per host, on the
# scraper itself to reuse them for consecutive queries.
rfp = self.robot_file_parsers.get(parsed_url.hostname)
if rfp is None:
rfp = RobotFileParser(self.get_robot_url(url))
rfp.read()
self.robot_file_parsers[parsed_url.hostname] = rfp
return rfp.can_fetch(self.user_agent, parsed_url.path)
def robots_get(url, *args, **kwargs):
u = urlparse(url)
robot_url = '{scm}://{loc}/robots.txt'.format(scm=u.scheme, loc=u.netloc)
robot = RobotFileParser(robot_url)
robot.read()
ua = kwargs.get('headers', dict()).get('User-Agent', '*')
if not robot.can_fetch(ua, url):
return 'Not Allowed By robots.txt'
delay = robot.crawl_delay(ua)
if delay:
time.sleep(delay)
return requests.get(url, *args, **kwargs)
def check_robots(base_url, ext_url):
'''
Check the robots.txt
Prints note if base_url + ext_url is legal for crawling
'''
bot = RobotFileParser(base_url + '/robots.txt')
bot.read()
if bot.can_fetch('*', base_url + ext_url):
print('robots.txt permits parsing')
else:
print('Do not parse')
return bot
def is_scraping_allowed(self):
"""
Megnezi, hogy a robots.txt nem tiltja-e a scrapelest. Nem igazan teljes
az ellenorzes, mert csak az all job url-t vizsgalja.
:return:
"""
robot_parser = RobotFileParser()
robots_url = urljoin(self.base_url, 'robots.txt')
robot_parser.set_url(robots_url)
robot_parser.read()
return robot_parser.can_fetch('*', urljoin(
self.base_url, self.all_job_url))
def _robot_parser(self, txt, url):
"""Parses robots.txt with user-agent="*".
:param txt: robots.txt to parse
:param url: URL to check
:returns: if url is allowed in robots.txt
:rtype: bool
"""
parser = RobotFileParser()
if txt:
parser.parse(txt.decode("ascii", "replace").splitlines())
return parser.can_fetch("*", url)
else:
return True
class Exclusion(object):
def __init__(self):
self.robot_cache = {}
self.rp = RobotFileParser()
# Broken? disallows anything in the robots.txt.
# Even if it is marked with Allow:
def test_url(self, url):
self.rp = RobotFileParser()
robot_url = uu.domain_name(url) + '/robots.txt'
self.rp.set_url(robot_url)
self.rp.read()
return self.rp.can_fetch('*', url)
def check_delay(user_agent, url):
robots_url = to_robots(url)
# Fetch the delay for all programs, ond this one in particular.
general_delay, have_general = get_delay(GENERAL_AGENT, robots_url)
agent_delay, have_agent = get_delay(user_agent, robots_url)
# The next part is straightforward:
# Check the general rules, and then check the program's rules,
# if there are rules.
robots_parser = RobotFileParser(robots_url)
robots_parser.read()
if (have_general and not robots_parser.can_fetch("*", url)):
return None
if (have_agent and not robots_parser.can_fetch(user_agent, url)):
return None
# Use the larger delay.
if general_delay < agent_delay:
return agent_delay
else:
return general_delay
def can_fetch(self, url):
parse = urllib.parse.urlparse(url)
hostname = parse.hostname
try:
robot = self.robot[hostname]
except Exception:
roboturl = urllib.parse.urlunparse((parse.scheme,parse.netloc,"robots.txt","","",""))
robot = RobotFileParser(roboturl)
try:
robot.read()
except Exception:
robot = _RobotAllowAll()
self.__robots[hostname] = robot
return robot.can_fetch("*", url)
def can_fetch_url(robots_url, site_url, useragent="*"):
"""
Using robots.txt found at robots_url, decides if useragent can fetch
site url
:param robots_url: robots.txt url
:param site_url: to be fetched url
:param useragent: useragent
:return: True, if fetching is allowed
"""
rfp = RobotFileParser()
rfp.set_url(robots_url)
rfp.read()
return rfp.can_fetch(useragent=useragent, url=site_url)
class RobotsTxt:
def __init__(self, base_url):
try:
url = urljoin(base_url, 'robots.txt')
self.rp = RobotFileParser()
self.set_url(url)
self.rp.read()
except:
self.rp = None
def canFetch(self, url):
if self.rp is None:
return True
else:
return self.rp.can_fetch('*', url)
def get_robots_text(self,url):
"""get_robots_txt
url: robots.txt を確認するサイトURL
"""
try:
rp = RobotFileParser()
# robots の url 取得
parsed_url = urlparse(url)
robots_url = "{0.scheme}://{0.netloc}/robots.txt".format(parsed_url)
# robots.txt 取得
rp.set_url(robots_url)
rp.read()
# 取得していいか確認
return rp.can_fetch("*",url)
except:
return False
class PythonRobotParser(RobotParser):
def __init__(self, robotstxt_body, spider):
from urllib.robotparser import RobotFileParser
self.spider = spider
robotstxt_body = decode_robotstxt(robotstxt_body, spider, to_native_str_type=True)
self.rp = RobotFileParser()
self.rp.parse(robotstxt_body.splitlines())
@classmethod
def from_crawler(cls, crawler, robotstxt_body):
spider = None if not crawler else crawler.spider
o = cls(robotstxt_body, spider)
return o
def allowed(self, url, user_agent):
user_agent = to_unicode(user_agent)
url = to_unicode(url)
return self.rp.can_fetch(user_agent, url)
def parse_robotstxt(url):
"""
Parse robots.txt
"""
parsed = urlsplit(url)
if parsed.scheme not in ['http', 'https']:
return False
if parsed.netloc == '':
return False
robot = RobotFileParser()
robot.set_url(parsed.scheme + "://" + parsed.netloc + "/robots.txt")
robot.read()
return dict(
allowed=robot.can_fetch('*', url),
rate=robot.request_rate('*'),
delay=robot.crawl_delay('*'),
)
def can_fetch(self, url):
parsed = urlparse(url)
if parsed.netloc in self.robots_txt_dict:
if self.robots_txt_dict[
parsed.
netloc] == 0: # 0 means that a robots.txt was not able to be gotten from that site so it assumes its all able to be crawled
return True
return self.robots_txt_dict[parsed.netloc].can_fetch('*', url)
try:
rp = RobotFileParser()
rp.set_url('{}://{}/robots.txt'.format(parsed.scheme,
parsed.netloc))
rp.read()
self.robots_txt_dict[parsed.netloc] = rp
return rp.can_fetch('*', url)
except:
print('error getting robots.txt')
self.robots_txt_dict[parsed.netloc] = 0
return True
def scrape_urls(URL, destfile):
''' This function will scrape all of the URLs
from a given site and store them in a file.
This function accepts two arguments, the
target URL and a file name of your choosing
to store the links in, both as strings.'''
links = []
#open the robots.txt file, which contains scraping permissions
robotparser = RobotFileParser()
robotparser.set_url(parse.urljoin(URL_base, 'robots.txt'))
robotparser.read()
#determine whether you're allowed to scraoe
if robotparser.can_fetch(useragent='Python-urllib/3.6', url=URL):
#open the page as a file object and save its contents
with request.urlopen(URL) as response:
content = response.read()
#make a beautifulSoup object from the content, to be read as HTML
soup = BeautifulSoup(content, 'html.parser')
#'a' is a link in HTML
for a in soup.find_all('a'):
#below is necessary to make sure your link has
#a web address associated with it
if a.has_attr('href'):
links.append(a['href'])
#only proceed if the "links" list was populated in previous step
if links:
#save all of the link URLS in an external file
with open(destfile, 'w') as outfile:
for link in links:
#newline character makes the links separate!
outfile.write(link + '\n')
class Crawler():
# Variables
parserobots = False
output = None
report = False
config = None
domain = ""
exclude = []
skipext = []
drop = []
debug = False
tocrawl = set([])
crawled = set([])
excluded = set([])
# TODO also search for window.location={.*?}
linkregex = re.compile(b'<a href=[\'|"](.*?)[\'"].*?>')
rp = None
response_code={}
nb_url=1 # Number of url.
nb_rp=0 # Number of url blocked by the robots.txt
nb_exclude=0 # Number of url excluded by extension or word
output_file = None
target_domain = ""
def __init__(self, parserobots=False, output=None, report=False ,domain="", exclude=[], skipext=[], drop=[], debug=False):
self.parserobots = parserobots
self.output = output
self.report = report
self.domain = domain
self.exclude = exclude
self.skipext = skipext
self.drop = drop
self.debug = debug
if self.debug:
logging.basicConfig(level=logging.DEBUG)
self.tocrawl = set([domain])
try:
self.target_domain = urlparse(domain)[1]
except:
raise ("Invalid domain")
if self.output:
try:
self.output_file = open(self.output, 'w')
except:
logging.debug ("Output file not available.")
exit(255)
def run(self):
print (config.xml_header, file=self.output_file)
logging.debug("Start the crawling process")
self.__crawling()
logging.debug("Crawling as reach the end of all found link")
print (config.xml_footer, file=self.output_file)
def __crawling(self):
crawling = self.tocrawl.pop()
url = urlparse(crawling)
self.crawled.add(crawling)
request = Request(crawling, headers={"User-Agent":config.crawler_user_agent})
try:
response = urlopen(request)
except Exception as e:
if hasattr(e,'code'):
if e.code in self.response_code:
self.response_code[e.code]+=1
else:
self.response_code[e.code]=1
logging.debug ("{1} ==> {0}".format(e, crawling))
return self.__continue_crawling()
# Read the response
try:
msg = response.read()
if response.getcode() in self.response_code:
self.response_code[response.getcode()]+=1
else:
self.response_code[response.getcode()]=1
response.close()
# Get the last modify date
if 'last-modified' in response.headers:
date = response.headers['Last-Modified']
else:
date = response.headers['Date']
date = datetime.strptime(date, '%a, %d %b %Y %H:%M:%S %Z')
except Exception as e:
logging.debug ("{1} ===> {0}".format(e, crawling))
return self.__continue_crawling()
print ("<url><loc>"+url.geturl()+"</loc><lastmod>"+date.strftime('%Y-%m-%dT%H:%M:%S')+"</lastmod></url>", file=self.output_file)
if self.output_file:
self.output_file.flush()
# Found links
links = self.linkregex.findall(msg)
for link in links:
link = link.decode("utf-8")
#logging.debug("Found : {0}".format(link))
if link.startswith('/'):
link = 'http://' + url[1] + link
elif link.startswith('#'):
link = 'http://' + url[1] + url[2] + link
elif not link.startswith('http'):
link = 'http://' + url[1] + '/' + link
# Remove the anchor part if needed
if "#" in link:
link = link[:link.index('#')]
# Drop attributes if needed
for toDrop in self.drop:
link=re.sub(toDrop,'',link)
# Parse the url to get domain and file extension
parsed_link = urlparse(link)
domain_link = parsed_link.netloc
target_extension = os.path.splitext(parsed_link.path)[1][1:]
if (link in self.crawled):
continue
if (link in self.tocrawl):
continue
if (link in self.excluded):
continue
if (domain_link != self.target_domain):
continue
if ("javascript" in link):
continue
# Count one more URL
self.nb_url+=1
# Check if the navigation is allowed by the robots.txt
if (not self.can_fetch(link)):
self.exclude_link(link)
self.nb_rp+=1
continue
# Check if the current file extension is allowed or not.
if (target_extension in self.skipext):
self.exclude_link(link)
self.nb_exclude+=1
continue
# Check if the current url doesn't contain an excluded word
if (not self.exclude_url(link)):
self.exclude_link(link)
self.nb_exclude+=1
continue
self.tocrawl.add(link)
return self.__continue_crawling()
def __continue_crawling(self):
if self.tocrawl:
self.__crawling()
def exclude_link(self,link):
if link not in self.excluded:
self.excluded.add(link)
def checkRobots(self):
if self.domain[len(self.domain)-1] != "/":
self.domain += "/"
request = Request(self.domain+"robots.txt", headers={"User-Agent":config.crawler_user_agent})
self.rp = RobotFileParser()
self.rp.set_url(self.domain+"robots.txt")
self.rp.read()
def can_fetch(self, link):
try:
if self.parserobots:
if self.rp.can_fetch("*", link):
return True
else:
logging.debug ("Crawling of {0} disabled by robots.txt".format(link))
return False
if not self.parserobots:
return True
return True
except:
# On error continue!
logging.debug ("Error during parsing robots.txt")
return True
def exclude_url(self, link):
for ex in self.exclude:
if ex in link:
return False
return True
def make_report(self):
print ("Number of found URL : {0}".format(self.nb_url))
print ("Number of link crawled : {0}".format(len(self.crawled)))
if self.parserobots:
print ("Number of link block by robots.txt : {0}".format(self.nb_rp))
if self.skipext or self.exclude:
print ("Number of link exclude : {0}".format(self.nb_exclude))
for code in self.response_code:
print ("Nb Code HTTP {0} : {1}".format(code, self.response_code[code]))
class SiteMap():
""" This class composes of all the functionalities needed to generate site_map"""
def __init__(self, main_page=None, robotrules=True, threadcount=1):
"""ctor that checks args and decides to enable single or multithreaded
generation of sitemap
"""
logging.info("Consider Robot.txt ? ==> "+str(robotrules))
self.robotrules = robotrules
self.site_map = {}
self.unvisited = set([])
self.start_page = None
self.robot_txt_rules = None
if main_page:
self.unvisited.add(main_page)
try:
self.start_page = urlparse(main_page).netloc
except:
logging.error("Improper URL, Please provide a Valid Url:"+main_page)
exit(0)
if self.robotrules == "True":
try:
logging.info("robot.txt respected")
self.robot_txt_rules = RobotFileParser()
self.robot_txt_rules.set_url(main_page + "/robots.txt")
self.robot_txt_rules.read()
except:
logging.error("Unable to read the robot.txt file")
self.robotrules = False # error reading robot.txt, ignore it forever
self.threadcount = int(threadcount)
def execute(self):
if self.threadcount <= 1: # if single threaded model is chosen, avoid threading
self.generate()
else:
self.start() # fasten by multi threads
def start(self):
"""This creates a pool of chosen limit so as to have the control and
spawns the main function and waits until process and subsequently
spawned process finish.
"""
self.pool = pool.Pool(self.threadcount)
self.pool.spawn(self.generate_parallels)
self.pool.join()
self.generate_reports()
def generate(self):
"""Non multithreaded model method that crawls until all pages are
crawled and assets are extracted. Once its done, it creates the
sitemap and assets json file for the given domain.
"""
while self.unvisited:
self.crawl()
self.generate_reports()
def generate_reports(self):
"""composes the xml tags with the keys in site_map member which are
nothing but the sitemap urls
"""
header = """<?xml version="1.0" encoding="UTF-8"?>
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9"
xmlns:xhtml="http://www.w3.org/1999/xhtml"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://www.sitemaps.org/schemas/sitemap/0.9
http://www.sitemaps.org/schemas/sitemap/0.9/sitemap.xsd">
"""
footer = """\n</urlset>\n"""
entry = "\t<url><loc>%s</loc></url>\n"
xml = header
for url in self.site_map.keys():
xml += entry % (url)
xml += footer
name = self.start_page.replace(".", "_")
self.create_file("%s.xml" % (name), xml)
self.create_file("%s_assets.json" % (name), json.dumps(self.site_map, indent=2, sort_keys=True))
def generate_parallels(self):
"""
This method is similar to recursive in a way that crawls pages and clears
the queue, which is self.unvisited. It stops when there are no urls to crawl
and all threads in pool are empty i.e they are not active anymore due to
finishing of crawling. Since its spawning a new thread and not calling
directly, it is a nice way to go about it for now :)
[Note:] There is a limit for recursion in Python and it can be increased by
sys.setrecursionlimit(1500)
An assumption has been made for this implementation that a website that has
more than 500 nested links needs a bit higher design in terms to store the
assets which might go above a hundred MB. In such cases, this can just be
converted into a loop. More over, there is no extra stack variables.
"""
self.crawl()
while len(self.unvisited) > 0 and not self.pool.full():
self.pool.spawn(self.generate_parallels)
def create_file(self, file, content):
"""writes the given content to the file"""
f = open(file, 'w')
f.write(content)
f.close()
def compose_url_from_href(self, page, href):
"""composes a proper url from domainlink and intralinks with in the page"""
url = urlparse(page)
if href.startswith('/'):
return "http://%s%s"%(url.netloc, href)
elif href.startswith('#'):
return "http://%s%s%s"%(url.netloc, url.path, href)
elif href.startswith('./'):
return "http://%s%s"%(url.netloc, href[1:])
elif not href.startswith('http'):
return "http://" + url.netloc + '/' + href
elif href.endswith('/'):
return href[:-1]
return href
def get_out_going_links(self, page, html_body):
"""extracts all the outgoing links and adds links that belong to
main page domain for further crawling if they are not crawled yet
This avoids:
- links that are .zip files
- links mentioned in href that are javascript methods
- mailto: links
"""
soup = BeautifulSoup(html_body, "html.parser")
valid_links_for_this_page = []
for a in soup.find_all('a', href=True):
href = a['href'].lower()
href = self.compose_url_from_href(page, href)
# clean the href so that it will have legitimate urls instead of #cluttered ones and q=param prints
href = urldefrag(href)[0] # skip intra links [this took time to find out !] ##1
# remove query params as only the path matters
if href.find('?') != -1:
href = href[:href.find('?')] ##2
new_page = urlparse(href)
# add to the queue only it it doesn't cause a cycle
# assumption: if a link ends with domain.com, assuming it can be crawled to make sitemap complete
if not str(new_page.netloc).endswith(self.start_page): # doesn't belong to domain
valid_links_for_this_page.append(href)
continue
if self.robot_allows(href) and \
not href in self.site_map and \
not href in self.unvisited and \
not 'javascript:' in href and \
not 'mailto:' in href:
if not ( href.endswith(".zip") or
href.endswith(".gz") or
href.endswith(".gzip") or
href.endswith(".tar") or
href.endswith(".bz2") or
href.endswith(".jpg") or
href.endswith(".png") or
href.endswith(".exe")
):
self.unvisited.add(href)
valid_links_for_this_page.append(href)
return valid_links_for_this_page
def get_assets(self, page, headers, html_body):
"""A nice feature of response header is that it reports the last-modified
time of the link on the server. If we are doing regular crawling, we can
avoid if the link is not updates since the last time. This method is
useful for indexing the data so as to minimize the crawling effort to
save execution time.
It updates the site_map dictionary with the links, css, images and scripts
"""
if 'last-modified' in headers:
date = headers['Last-Modified']
else:
date = headers['Date']
soup = BeautifulSoup(html_body, "html.parser")
img = soup.findAll("img")
css = soup.findAll("link", {"rel": "stylesheet"})
js = soup.findAll('script')
self.site_map[page] = {
'date': date,
'links': self.get_out_going_links(page, html_body),
'css': [c['href'] for c in css],
'img': [i['src'] for i in img],
'js': [x.get('src', 'inline jscode') for x in js]
}
def crawl(self):
"""This actually opens the url and calls the assets method """
if len(self.unvisited) <= 0:
return
page = self.unvisited.pop()
if page in self.site_map:
return
logging.info("Starting to Crawl Page: " + page)
try:
response = self.access_page(page)
if (response.status_code != 200):
return None
html_body = response.text
self.get_assets(page, response.headers, html_body)
except:
logging.error("Issue while opening url: %s" + page)
return None
logging.debug("Crawled Pages: {}".format(len(self.site_map)))
def access_page(self, url):
"""accesses the url from the server. This method was created
to enable mock tests.
"""
return requests.get(url)
def get_site_map(self):
"""exposes site_map"""
return self.site_map
def set_start_page(self, url):
"""sets the start page for the crawler"""
self.start_page = url
def robot_allows(self, link):
"""method to check if link can be accessed as per robot rules"""
if not self.robotrules: return True
try:
if self.robot_txt_rules.can_fetch("*", link):
return True
return False
except:
return True
# Copyright(C) 2018 刘珅珅
# Environment: python 3.6.4
# Date: 2018.9.8
# robots协议测试
from urllib.robotparser import RobotFileParser
url = 'http://www.jianshu.com/robots.txt'
rp = RobotFileParser(url)
rp.read()
# 在scrapy中，如果希望爬取robots.txt禁止的页面，需要禁用robots
print(rp.can_fetch('*', 'http://www.jianshu.com/p/b67554025d7d')) # False，不允许爬取
class Crawler():
# Variables
parserobots = False
output = None
report = False
config = None
domain = ""
exclude = []
skipext = []
drop = []
debug = False
tocrawl = set([])
crawled = set([])
excluded = set([])
marked = {}
not_parseable_ressources = (".epub", ".mobi", ".docx", ".doc", ".opf", ".7z", ".ibooks", ".cbr", ".avi", ".mkv", ".mp4", ".jpg", ".jpeg", ".png", ".gif" ,".pdf", ".iso", ".rar", ".tar", ".tgz", ".zip", ".dmg", ".exe")
# TODO also search for window.location={.*?}
linkregex = re.compile(b'<a [^>]*href=[\'|"](.*?)[\'"][^>]*?>')
imageregex = re.compile (b'<img [^>]*src=[\'|"](.*?)[\'"].*?>')
rp = None
response_code={}
nb_url=1 # Number of url.
nb_rp=0 # Number of url blocked by the robots.txt
nb_exclude=0 # Number of url excluded by extension or word
output_file = None
target_domain = ""
scheme = ""
def __init__(self, parserobots=False, output=None, report=False ,domain="",
exclude=[], skipext=[], drop=[], debug=False, verbose=False, images=False):
self.parserobots = parserobots
self.output = output
self.report = report
self.domain = domain
self.exclude = exclude
self.skipext = skipext
self.drop = drop
self.debug = debug
self.verbose = verbose
self.images = images
if self.debug:
log_level = logging.DEBUG
elif self.verbose:
log_level = logging.INFO
else:
log_level = logging.ERROR
logging.basicConfig(level=log_level)
self.tocrawl = set([self.clean_link(domain)])
try:
url_parsed = urlparse(domain)
self.target_domain = url_parsed.netloc
self.scheme = url_parsed.scheme
except:
logging.error("Invalide domain")
raise ("Invalid domain")
if self.output:
try:
self.output_file = open(self.output, 'w')
except:
logging.error ("Output file not available.")
exit(255)
def run(self):
print(config.xml_header, file=self.output_file)
if self.parserobots:
self.check_robots()
logging.info("Start the crawling process")
while len(self.tocrawl) != 0:
self.__crawling(first,domainname,all_link)
logging.info("Crawling has reached end of all found links")
print (config.xml_footer, file=self.output_file)
def __crawling(self,frst,dmname,all_links):
crawling = self.tocrawl.pop()
if frst==0:
dmname=obtaindomain(crawling)
frst=1
url = urlparse(crawling)
self.crawled.add(crawling)
logging.info("Crawling #{}: {}".format(len(self.crawled), url.geturl()))
request = Request(crawling, headers={"User-Agent":config.crawler_user_agent})
# Ignore ressources listed in the not_parseable_ressources
# Its avoid dowloading file like pdf… etc
if not url.path.endswith(self.not_parseable_ressources):
try:
response = urlopen(request)
except Exception as e:
if hasattr(e,'code'):
if e.code in self.response_code:
self.response_code[e.code]+=1
else:
self.response_code[e.code]=1
# Gestion des urls marked pour le reporting
if self.report:
if e.code in self.marked:
self.marked[e.code].append(crawling)
else:
self.marked[e.code] = [crawling]
logging.debug ("{1} ==> {0}".format(e, crawling))
return self.__continue_crawling()
else:
logging.debug("Ignore {0} content might be not parseable.".format(crawling))
response = None
# Read the response
if response is not None:
try:
msg = response.read()
if response.getcode() in self.response_code:
self.response_code[response.getcode()]+=1
else:
self.response_code[response.getcode()]=1
response.close()
# Get the last modify date
if 'last-modified' in response.headers:
date = response.headers['Last-Modified']
else:
date = response.headers['Date']
date = datetime.strptime(date, '%a, %d %b %Y %H:%M:%S %Z')
except Exception as e:
logging.debug ("{1} ===> {0}".format(e, crawling))
return None
else:
# Response is None, content not downloaded, just continu and add
# the link to the sitemap
msg = "".encode( )
date = None
# Image sitemap enabled ?
image_list = "";
if self.images:
# Search for images in the current page.
images = self.imageregex.findall(msg)
for image_link in list(set(images)):
image_link = image_link.decode("utf-8", errors="ignore")
# Ignore link starting with data:
if image_link.startswith("data:"):
continue
# If path start with // get the current url scheme
if image_link.startswith("//"):
image_link = url.scheme + ":" + image_link
# Append domain if not present
elif not image_link.startswith(("http", "https")):
if not image_link.startswith("/"):
image_link = "/{0}".format(image_link)
image_link = "{0}{1}".format(self.domain.strip("/"), image_link.replace("./", "/"))
# Ignore image if path is in the exclude_url list
if not self.exclude_url(image_link):
continue
# Ignore other domain images
image_link_parsed = urlparse(image_link)
if image_link_parsed.netloc != self.target_domain:
continue
# Test if images as been already seen and not present in the
# robot file
if self.can_fetch(image_link):
logging.debug("Found image : {0}".format(image_link))
image_list = "{0}<image:image><image:loc>{1}</image:loc></image:image>".format(image_list, self.htmlspecialchars(image_link))
# Last mod fetched ?
lastmod = ""
if date:
lastmod = "<lastmod>"+date.strftime('%Y-%m-%dT%H:%M:%S+00:00')+"</lastmod>"
print ("<url><loc>"+self.htmlspecialchars(url.geturl())+"</loc>" + lastmod + image_list + "</url>", file=self.output_file)
tempdom=obtaindomain(self.htmlspecialchars(url.geturl()))
templink=self.htmlspecialchars(url.geturl())
if tempdom==dmname:
if templink not in all_links:
outputprint(templink)
all_links.append(templink)
if self.output_file:
self.output_file.flush()
# Found links
links = self.linkregex.findall(msg)
for link in links:
link = link.decode("utf-8", errors="ignore")
link = self.clean_link(link)
logging.debug("Found : {0}".format(link))
if link.startswith('/'):
link = url.scheme + '://' + url[1] + link
elif link.startswith('#'):
link = url.scheme + '://' + url[1] + url[2] + link
elif link.startswith(("mailto", "tel")):
continue
elif not link.startswith(('http', "https")):
link = url.scheme + '://' + url[1] + '/' + link
# Remove the anchor part if needed
if "#" in link:
link = link[:link.index('#')]
# Drop attributes if needed
for toDrop in self.drop:
link=re.sub(toDrop,'',link)
# Parse the url to get domain and file extension
parsed_link = urlparse(link)
domain_link = parsed_link.netloc
target_extension = os.path.splitext(parsed_link.path)[1][1:]
if link in self.crawled:
continue
if link in self.tocrawl:
continue
if link in self.excluded:
continue
if domain_link != self.target_domain:
continue
if parsed_link.path in ["", "/"]:
continue
if "javascript" in link:
continue
if self.is_image(parsed_link.path):
continue
if parsed_link.path.startswith("data:"):
continue
# Count one more URL
self.nb_url+=1
# Check if the navigation is allowed by the robots.txt
if not self.can_fetch(link):
self.exclude_link(link)
self.nb_rp+=1
continue
# Check if the current file extension is allowed or not.
if (target_extension in self.skipext):
self.exclude_link(link)
self.nb_exclude+=1
continue
# Check if the current url doesn't contain an excluded word
if (not self.exclude_url(link)):
self.exclude_link(link)
self.nb_exclude+=1
continue
self.tocrawl.add(link)
return None
def clean_link(self, link):
l = urlparse(link)
l_res = list(l)
l_res[2] = l_res[2].replace("./", "/")
l_res[2] = l_res[2].replace("//", "/")
return urlunparse(l_res)
def is_image(self, path):
mt,me = mimetypes.guess_type(path)
return mt is not None and mt.startswith("image/")
def __continue_crawling(self):
if self.tocrawl:
self.__crawling(first,domainname,all_link)
def exclude_link(self,link):
if link not in self.excluded:
self.excluded.add(link)
def check_robots(self):
robots_url = urljoin(self.domain, "robots.txt")
self.rp = RobotFileParser()
self.rp.set_url(robots_url)
self.rp.read()
def can_fetch(self, link):
try:
if self.parserobots:
if self.rp.can_fetch("*", link):
return True
else:
logging.debug ("Crawling of {0} disabled by robots.txt".format(link))
return False
if not self.parserobots:
return True
return True
except:
# On error continue!
logging.debug ("Error during parsing robots.txt")
return True
def exclude_url(self, link):
for ex in self.exclude:
if ex in link:
return False
return True
def htmlspecialchars(self, text):
return text.replace("&", "&").replace('"', """).replace("<", "<").replace(">", ">")
def make_report(self):
print ("Number of found URL : {0}".format(self.nb_url))
print ("Number of link crawled : {0}".format(len(self.crawled)))
if self.parserobots:
print ("Number of link block by robots.txt : {0}".format(self.nb_rp))
if self.skipext or self.exclude:
print ("Number of link exclude : {0}".format(self.nb_exclude))
for code in self.response_code:
print ("Nb Code HTTP {0} : {1}".format(code, self.response_code[code]))
for code in self.marked:
print ("Link with status {0}:".format(code))
for uri in self.marked[code]:
print ("\t- {0}".format(uri))
def robot_can_fetch(robots_txt_content, url):
parser = RobotFileParser()
parser.parse(robots_txt_content.splitlines())
return parser.can_fetch(USER_AGENT, urlparse(url).path)
class _RobotAllowAll:
def can_fetch(self, a,b):
return True
class Robot:
def __init__(self):
self.__robots = {}
def can_fetch(self, url):
parse = urllib.parse.urlparse(url)
hostname = parse.hostname
try:
robot = self.robot[hostname]
except Exception:
roboturl = urllib.parse.urlunparse((parse.scheme,parse.netloc,"robots.txt","","",""))
robot = RobotFileParser(roboturl)
try:
robot.read()
except Exception:
robot = _RobotAllowAll()
self.__robots[hostname] = robot
return robot.can_fetch("*", url)
if __name__ == "__main__":
r = RobotFileParser("http://www.letudiant.fr", {})
r.read()
print(r.can_fetch("*", "http://www.letudiant.fr/"))
class Crawler(object):
def __init__(self, database, fetcher, analyzer, verbose=False):
self.database = database
self.fetcher = fetcher
self.analyzer = analyzer
self.verbose = verbose
self.queue = set()
self.robot_parser = RobotFileParser()
def crawl(self, url):
"""Begin recursively crawling pages starting from the given URL.
:param url: Starting URL
:returns: None
"""
if self.database.is_page_stored(url):
print("Page is already crawled. Use --flush to flush the database file.", file=sys.stderr)
else:
# Because crawling is restricted to pages on the same domain, the
# robots.txt file can be loaded once at the beginning of the crawl
self.load_robots_file(url)
# Add the starting URL to the queue of pages to be crawled, and
# then keep crawling while there are still URLs in the queue
self.queue.add(url)
while len(self.queue) > 0:
self.crawl_one(self.queue.pop())
def crawl_one(self, url):
"""Fetch a single page and analyze it for links. The found triples are
stored in the database, and found links that should be crawled are
added to the queue.
:param url: The page to fetch and analyze
:returns: None
"""
if self.verbose:
print(url, file=sys.stderr)
status, html = self.fetcher.fetch(url)
if status is None:
# The status code will be None if retrieval failed
print("Failed to get {}".format(url), file=sys.stderr)
else:
# Search for links and images in the page, and get them as triples
# of (page URL, link type, link URL)
triples = self.analyzer.analyze(url, html)
self.database.store_triples(triples)
# Any linked URLs that are eligible for crawling are added to the
# pending crawl queue
for page_url, link_type, link_url in triples:
if self.should_crawl(page_url, link_type, link_url):
self.queue.add(link_url)
def should_crawl(self, page_url, link_type, link_url):
"""Determine whether a URL should be crawled.
:param page_url: The page the link came from.
:param link_type: The type of link URL.
:param link_url: The link URL to test.
:returns: True if the link URL should be crawled, otherwise False.
"""
# Only HTML pages should be crawled, not other media
if link_type not in ('page', 'iframe'):
return False
# The link should be on the same domain as the page it's linked from
if not self.have_same_domain(page_url, link_url):
return False
# Fetching the link URL should be permitted by robots.txt
if not self.robot_parser.can_fetch('Cosmo', link_url):
return False
# The linked page should not have been crawled already
if self.database.is_page_stored(link_url):
return False
return True
def have_same_domain(self, url1, url2):
"""Test whether two URLs have the same hostname and port.
:returns: True if they do, otherwise False
"""
return urlparse(url1).netloc == urlparse(url2).netloc
def load_robots_file(self, url):
"""Load the /robots.txt file for the given URL by reusing the scheme
and authority parts.
:param url: The URL from which to take the scheme and authority parts.
:returns: None
"""
# Create a new URL with the same scheme, host and port, but with a
# path of /robots.txt
parsed = urlparse(url)
robots_url = urlunparse((parsed.scheme, parsed.netloc, '/robots.txt', '', '', ''))
# Load the robots.txt file using the requests library, because we need
# to specify the User-Agent header. I noticed on a CloudFlare-fronted
# site that it returns a 403 for /robots.txt if the the user agent is
# Python-urllib, but 200 if it's Cosmo.
status, robots_file = self.fetcher.fetch(robots_url)
if status in (401, 403):
self.robot_parser.disallow_all = True
elif status >= 400:
self.robot_parser.allow_all = True
else:
self.robot_parser.parse(robots_file.splitlines())
