misc/web-scraper
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

rename all instances of base_url to rooturl, add more documentation

Browse Source
This commit is contained in:
Simon Weald
2018-09-18 18:24:15 +01:00
parent 32d7f1e54b
commit 679b1b7b53
3 changed files with 65 additions and 45 deletions
Download Patch File Download Diff File Expand all files Collapse all files

73
utils/helpers.py
View File

@@ -14,11 +14,22 @@ import urllib.robotparser
class AsyncCrawler(object):
'''
docstring
A concurrent recursive web crawler.
A recursive web crawler which finds all URLs local to the domains specified
in the `rooturl` argument.
Arguments:
rooturl: Root domain to begin crawling.
robots: RobotsTxt object for the rooturl.
concurrency: number of concurrent pages to crawl.
Returns:
All discovered pages in a set.
'''
def __init__(self, baseurl=None, robots=None, concurrency=None):
self.baseurl = baseurl
def __init__(self, rooturl=None, robots=None, concurrency=None):
self.rooturl = rooturl
self.robots = robots
self.crawled = set()
self.headers = {'Accept-Encoding': 'gzip, deflate',
@@ -28,7 +39,7 @@ class AsyncCrawler(object):
async def crawl_url(self, url=None):
'''
Crawls the given URL and finds all new URLs in the initial page.
Crawls the given URL and finds all new URLs in the given page.
'''
urls = []
source = await self.get_source(url)
@@ -43,7 +54,7 @@ class AsyncCrawler(object):
allows it.
'''
# ensure the URL is in a sane format
url = standardise_url(url=url, base_url=self.baseurl)
url = standardise_url(url=url, rooturl=self.rooturl)
if url and self.robots.check(url=url):
return url
@@ -53,14 +64,15 @@ class AsyncCrawler(object):
async def get_source(self, url=None):
'''
Obtains the URL's source, provided it is HTML. Usage of semaphores
ensures only a certain number of coroutines can run at once.
ensures only a certain number of coroutines can run at any given
time.
'''
async with self.semaphore:
async with self.client_session.head(url, timeout=5) as head:
try:
_ = await head.read()
except Exception as e:
print(e)
except Exception:
pass
if 'text/html' in head.headers['Content-Type']:
async with self.client_session.get(url, timeout=5) as resp:
try:
@@ -68,6 +80,8 @@ class AsyncCrawler(object):
return source
except Exception:
return None
else:
return None
def find_all_urls(self, source=None):
'''
@@ -99,21 +113,22 @@ class AsyncCrawler(object):
for url in urls:
# mark the URL as seen.
self.crawled.add(url)
# create an coroutine to crawl the URL.
# create a task to crawl the URL.
tasks.append(self.crawl_url(url))
# wait for all tasks to complete.
for task in asyncio.as_completed(tasks):
urls = None
try:
# try getting all tasks as completed.
urls = await task
except Exception as e:
print(e)
except Exception:
# skip until all tasks have completed.
pass
# add the URLs to a set to be returned.
if urls:
for url in urls:
print('Found: {0}'.format(url))
all_urls.add(url)
return all_urls
@@ -125,15 +140,17 @@ class AsyncCrawler(object):
'''
self.client_session = aiohttp.ClientSession(headers=self.headers)
to_crawl = []
to_crawl.append(self.baseurl)
# add the root URL to initialise the crawler.
to_crawl.append(self.rooturl)
print('Crawling: {0}'.format(self.baseurl))
print('Crawling: {0}'.format(self.rooturl))
while len(to_crawl) > 0:
discovered_urls = await self.run(urls=to_crawl)
# empty to_crawl list and then add all newly discovered URLs for
# the next iteration.
to_crawl.clear()
to_crawl.extend(discovered_urls)
print('{0} URLs crawled'.format(len(self.crawled)))
# close the ssions once all URLs have been crawled.
await self.client_session.close()
@@ -143,19 +160,23 @@ class AsyncCrawler(object):
class RobotsTxt(object):
'''
needs a docstring
Retrieve and query robots.txt for a given domain.
Retrieves and parses robots.txt for the given domain. Calling the check()
method returns True or False depending on whether crawling of that given
URL is allowed.
'''
def __init__(self, base_url=None):
def __init__(self, rooturl=None):
'''
Manually retrieve robots.txt to allow us to set the user-agent (works
around sites which disallow access to robots.txt without a sane
user-agent).
'''
self.base_url = base_url
self.rooturl = rooturl
self.headers = {'User-Agent': 'Mozilla/5.0 (X11; Linux x86_64; rv:61.0) Gecko/20100101 Firefox/61.0'}
robots_url = urljoin(self.base_url, 'robots.txt')
robots_url = urljoin(self.rooturl, 'robots.txt')
request = urllib.request.Request(robots_url, headers=self.headers)
robots = urllib.robotparser.RobotFileParser()
@@ -180,16 +201,16 @@ class RobotsTxt(object):
return self.robots.can_fetch("*", url)
def standardise_url(url=None, base_url=None):
def standardise_url(url=None, rooturl=None):
'''
If `base_url` is None then we attempt to standarise the URL to ensure it
If `rooturl` is None then we attempt to standarise the URL to ensure it
can be prepended to relative URLs. If no scheme has been provided then we
default to http as any sane https-only site should 301 redirect http to
https.
If `base_url` is set, we standardise URLs to strip queries and fragments
If `rooturl` is set, we standardise URLs to strip queries and fragments
(we don't want to scrape in-page anchors etc). Any relative URLs will be
appended to the base url.
appended to the root url.
Returns a standardised URL as a string.
'''
@@ -200,7 +221,7 @@ def standardise_url(url=None, base_url=None):
split_url = urlsplit(url)
if not base_url:
if not rooturl:
# This will sanitise the initial url provided by the user.
if split_url.scheme and split_url.scheme.startswith('http'):
return "".join([split_url.scheme, delim, split_url.netloc])
@@ -209,10 +230,10 @@ def standardise_url(url=None, base_url=None):
else:
# if url.endswith(file_extensions):
# Sanitise discovered URLs. We already expect them in the format
# protocol://base_url/path
# protocol://rooturl/path
if url.startswith('/'):
return urljoin(base_url, split_url.path)
elif url.startswith(base_url):
return urljoin(rooturl, split_url.path)
elif url.startswith(rooturl):
return "".join([split_url.scheme, delim, split_url.netloc,
split_url.path])