def test_backlink_extraction(self):
links = util.extract_links(None)
self.assertEquals(links, [])
text = "[[foo]], [[foo|bar]]"
links = util.extract_links(text)
self.assertEquals(links, ["foo"])
def test_backlink_extraction(self):
links = util.extract_links(None)
self.assertEquals(links, [])
text = "[[foo]], [[foo|bar]]"
links = util.extract_links(text)
self.assertEquals(links, ["foo"])
def put(self):
"""Adds the gaewiki:parent: labels transparently."""
if self.body is not None:
options = util.parse_page(self.body)
self.redirect = options.get('redirect')
self.pread = options.get(
'public') == 'yes' and options.get('private') != 'yes'
self.labels = options.get('labels', [])
if 'date' in options:
try:
self.created = datetime.datetime.strptime(
options['date'], '%Y-%m-%d %H:%M:%S')
except ValueError:
pass
if 'name' in options and options['name'] != self.title:
if self.get_by_title(options['name'],
create_if_none=False) is not None:
raise ValueError('A page named "%s" already exists.' %
options['name'])
self.title = options['name']
self.__update_geopt()
self.links = util.extract_links(self.body)
self.add_implicit_labels()
db.Model.put(self)
settings.check_and_flush(self)
def handle_feed(feed, source):
"""Handles a Superfeedr JSON feed.
Creates :class:`models.BlogPost` entities and adds propagate-blogpost tasks
for new items.
http://documentation.superfeedr.com/schema.html#json
http://documentation.superfeedr.com/subscribers.html#pubsubhubbubnotifications
Args:
feed: unicode string, Superfeedr JSON feed
source: Blogger, Tumblr, or WordPress
"""
logging.info('Source: %s %s', source.label(), source.key.string_id())
logging.info('Raw feed: %s', feed)
if source.status != 'enabled':
logging.info('Dropping because source is %s', source.status)
return
elif 'webmention' not in source.features:
logging.info("Dropping because source doesn't have webmention feature")
return
for item in json.loads(feed).get('items', []):
url = item.get('permalinkUrl') or item.get('id')
if not url:
logging.error('Dropping feed item without permalinkUrl or id!')
continue
# extract links from content, discarding self links.
#
# i don't use get_webmention_target[s]() here because they follows redirects
# and fetch link contents, and this handler should be small and fast and try
# to return a response to superfeedr successfully.
#
# TODO: extract_links currently has a bug that makes it drop trailing
# slashes. ugh. fix that.
content = item.get('content') or item.get('summary', '')
links = [
util.clean_url(util.unwrap_t_umblr_com(l))
for l in util.extract_links(content)
if util.domain_from_link(l) not in source.domains
]
logging.info('Found links: %s', links)
if len(url) > _MAX_KEYPART_BYTES:
logging.warning(
'Blog post URL is too long (over 500 chars)! Giving up.')
bp = models.BlogPost(id=url[:_MAX_KEYPART_BYTES],
source=source.key,
feed_item=item,
failed=links)
else:
bp = models.BlogPost(id=url,
source=source.key,
feed_item=item,
unsent=links)
bp.get_or_save()
def handle_feed(feed, source):
"""Handles a Superfeedr JSON feed.
Creates :class:`models.BlogPost` entities and adds propagate-blogpost tasks
for new items.
http://documentation.superfeedr.com/schema.html#json
http://documentation.superfeedr.com/subscribers.html#pubsubhubbubnotifications
Args:
feed: unicode string, Superfeedr JSON feed
source: Blogger, Tumblr, or WordPress
"""
logging.info('Source: %s %s', source.label(), source.key.string_id())
logging.info('Raw feed: %s', feed)
if source.status != 'enabled':
logging.info('Dropping because source is %s', source.status)
return
elif 'webmention' not in source.features:
logging.info("Dropping because source doesn't have webmention feature")
return
for item in json.loads(feed).get('items', []):
url = item.get('permalinkUrl') or item.get('id')
if not url:
logging.error('Dropping feed item without permalinkUrl or id!')
continue
# extract links from content, discarding self links.
#
# i don't use get_webmention_target[s]() here because they follows redirects
# and fetch link contents, and this handler should be small and fast and try
# to return a response to superfeedr successfully.
#
# TODO: extract_links currently has a bug that makes it drop trailing
# slashes. ugh. fix that.
content = item.get('content') or item.get('summary', '')
links = [util.clean_url(util.unwrap_t_umblr_com(l))
for l in util.extract_links(content)
if util.domain_from_link(l) not in source.domains]
unique = []
for link in util.dedupe_urls(links):
if len(link) <= _MAX_STRING_LENGTH:
unique.append(link)
else:
logging.info('Giving up on link over %s chars! %s', _MAX_STRING_LENGTH, link)
logging.info('Found links: %s', unique)
if len(url) > _MAX_KEYPART_BYTES:
logging.warning('Blog post URL is too long (over 500 chars)! Giving up.')
bp = models.BlogPost(id=url[:_MAX_KEYPART_BYTES], source=source.key,
feed_item=item, failed=unique)
else:
bp = models.BlogPost(id=url, source=source.key, feed_item=item, unsent=unique)
bp.get_or_save()
async def on_message(self, message: Message):
if message.author == self.user:
return
reply = partial(message.reply, mention_author=False)
for match in extract_links(message.content):
cprint(match[0], blue)
try:
info = extract_info(match[0])
except NoVideoException:
cprint("Tweet has no video", yellow)
continue
except Exception as ex: # pylint: disable=broad-except
cprint(f"Extraction error: {ex}", red)
await reply("Failed to download video")
continue
if "url" not in info:
cprint("No url in info dict", yellow)
continue
url = info["url"]
try:
buffer = await download(url, message.guild.filesize_limit)
except FileSizeException as ex:
if ex.filesize:
cprint(
f"Not uploading, file is too large: {ex.filesize} > {ex.limit}",
red)
else:
cprint(
f"Not uploading, file is larger than [{ex.limit}] bytes",
red)
cprint("Falling back to direct URL", yellow)
# If the file is too large, we fall back to posting a direct URL
await reply(url)
continue
except Exception as ex: # pylint: disable=broad-except
cprint(f"Http error: {ex}", red)
await reply("Failed to download video")
continue
await reply(file=DiscordFile(fp=buffer, filename=f"{match[1]}.mp4")
)
cprint("Upload success", green)
def handle_feed(feed, source):
"""Handles a Superfeedr JSON feed.
Creates BlogPost entities and adds propagate-blogpost tasks for new items.
http://documentation.superfeedr.com/schema.html#json
http://documentation.superfeedr.com/subscribers.html#pubsubhubbubnotifications
Args:
feed: string, Superfeedr JSON feed
source: Blogger, Tumblr, or WordPress
"""
logging.info('Source: %s %s', source.label(), source.key.string_id())
logging.info('Raw feed: %s', feed)
if source.status != 'enabled':
logging.warning('Dropping because source is %s', source.status)
return
elif 'webmention' not in source.features:
logging.warning("Dropping because source doesn't have webmention feature")
return
for item in json.loads(feed).get('items', []):
url = item.get('permalinkUrl') or item.get('id')
if not url:
logging.error('Dropping feed item without permalinkUrl or id!')
continue
source.preprocess_superfeedr_item(item)
# extract links from content, discarding self links.
#
# i don't use get_webmention_target[s]() here because they follows redirects
# and fetch link contents, and this handler should be small and fast and try
# to return a response to superfeedr successfully.
#
# TODO: extract_links currently has a bug that makes it drop trailing
# slashes. ugh. fix that.
content = item.get('content') or item.get('summary', '')
links = [l for l in util.extract_links(content)
if util.domain_from_link(l) not in source.domains]
logging.info('Found links: %s', links)
models.BlogPost(id=url,
source=source.key,
feed_item=item,
unsent=links,
).get_or_save()
def test_extract_links(self):
self.assertEquals([], util.extract_links(None))
self.assertEquals([], util.extract_links(''))
self.assertEquals([], util.extract_links('asdf qwert'))
for text in ('http://foo.com',
' http://foo.com ',
' http://foo.com \n http://foo.com ',
'x http://foo.com\ny',
'x\thttp://foo.com.',
'x\rhttp://foo.com! ',
'x http://foo.com? ',
'<a href="http://foo.com">',
"<a href='http://foo.com'>",
'<a href="xyz">http://foo.com</a>',
):
self.assertEquals(['http://foo.com'], util.extract_links(text),
'Failed on %r' % text)
self.assertEquals(
['http://foo.com', 'https://www.bar.com'],
util.extract_links('x http://foo.com y https://www.bar.com z'))
self.assertEquals(
['http://foo.com', 'http://bar.com'],
util.extract_links('asdf http://foo.com qwert <a class="x" href="http://bar.com" >xyz</a> www.baz.com'))
# trailing slash
# TODO: make this work
# self.assertEquals(['http://foo.com/'],
# util.extract_links('x http://foo.com/'))
# query
self.assertEquals(['http://foo.com/bar?baz=baj'],
util.extract_links('http://foo.com/bar?baz=baj y'))
# preserve order
self.assertEquals(['http://%s' % c for c in 'a', 'b', 'c', 'd'],
util.extract_links('http://a http://b http://c http://d'))
def test_extract_links(self):
self.assertEquals([], util.extract_links(None))
self.assertEquals([], util.extract_links(''))
self.assertEquals([], util.extract_links('asdf qwert'))
for text in ('http://foo.com',
' http://foo.com ',
' http://foo.com \n http://foo.com ',
'x http://foo.com\ny',
'x\thttp://foo.com.',
'x\rhttp://foo.com! ',
'x http://foo.com? ',
'<a href="http://foo.com">',
"<a href='http://foo.com'>",
'<a href="xyz">http://foo.com</a>',
):
self.assertEquals(['http://foo.com'], util.extract_links(text),
'Failed on %r' % text)
self.assertEquals(
['http://foo.com', 'https://www.bar.com'],
util.extract_links('x http://foo.com y https://www.bar.com z'))
self.assertEquals(
['http://foo.com', 'http://bar.com'],
util.extract_links('asdf http://foo.com qwert <a class="x" href="http://bar.com" >xyz</a> www.baz.com'))
# trailing slash
# TODO: make this work
# self.assertEquals(['http://foo.com/'],
# util.extract_links('x http://foo.com/'))
# query
self.assertEquals(['http://foo.com/bar?baz=baj'],
util.extract_links('http://foo.com/bar?baz=baj y'))
# preserve order
self.assertEquals(['http://%s' % c for c in 'a', 'b', 'c', 'd'],
util.extract_links('http://a http://b http://c http://d'))
def put(self):
"""Adds the gaewiki:parent: labels transparently."""
if self.body is not None:
options = util.parse_page(self.body)
self.redirect = options.get('redirect')
self.pread = options.get('public') == 'yes' and options.get('private') != 'yes'
self.labels = options.get('labels', [])
if 'date' in options:
try:
self.created = datetime.datetime.strptime(options['date'], '%Y-%m-%d %H:%M:%S')
except ValueError:
pass
if 'name' in options and options['name'] != self.title:
if self.get_by_title(options['name'], create_if_none=False) is not None:
raise ValueError('A page named "%s" already exists.' % options['name'])
self.title = options['name']
self.__update_geopt()
self.links = util.extract_links(self.body)
self.add_implicit_labels()
db.Model.put(self)
settings.check_and_flush(self)
def visit(self, conn, crawl_id, source_id, visit_url):
'''
Not abstract. Visits a url during a crawl.
Inserts all relevant information to the database for a single visit.
Inserts article information if the url matches the article regex.
'''
visit_url_id = visit_url['id']
visit_url_string = visit_url['url']
base_url_string = self.base_url_string
html = util.download_html(visit_url_string)
found_links = util.extract_links(html, base_url_string)
visit_id = queries.insert_visit(conn, crawl_id, visit_url_id)
new_url_ids = queries.insert_urls(conn, found_links)
queries.insert_links(conn, visit_id, new_url_ids)
if self.is_article(visit_url_string):
article = util.extract_article(html, visit_url_string)
article_title = article.title
article_text = article.text
article_date = self.extract_date_from_url(visit_url_string)
queries.insert_article(conn, visit_url_id, article_title, article_text, article_date, source_id)
""" Driver to call the get_batch method in scraper """
scraper.get_batch(pages,
page_hook=parse_page,
ret_hook=(ret_hook, [output, pages[-1]]),
verbose=True)
if __name__ == '__main__':
import sys
scraper = WebScraper(BASE_URI)
try:
output = open(sys.argv[1], 'w')
except IndexError:
print("Usage: {} <file>".format(sys.argv[0]))
exit(1)
output.write('{')
latest_update = "" # todo: something with this
base = scraper.get_html()
latest_update = get_latest_update(base)
pages = extract_links(base, 'td', 'data')
do_scrape(pages, output)
output.write('}')
output.close()
def handle_feed(feed, source):
"""Handles a Superfeedr JSON feed.
Creates :class:`models.BlogPost` entities and adds propagate-blogpost tasks
for new items.
http://documentation.superfeedr.com/schema.html#json
http://documentation.superfeedr.com/subscribers.html#pubsubhubbubnotifications
Args:
feed: unicode string, Superfeedr JSON feed
source: Blogger, Tumblr, or WordPress
"""
logger.info(f'Source: {source.label()} {source.key_id()}')
logger.info(f'Raw feed: {feed}')
if not feed:
return
if source.status != 'enabled':
logger.info(f'Dropping because source is {source.status}')
return
elif 'webmention' not in source.features:
logger.info("Dropping because source doesn't have webmention feature")
return
for item in feed.get('items', []):
url = item.get('permalinkUrl') or item.get('id')
if not url:
logger.error('Dropping feed item without permalinkUrl or id!')
continue
# extract links from content, discarding self links.
#
# i don't use get_webmention_target[s]() here because they follows redirects
# and fetch link contents, and this handler should be small and fast and try
# to return a response to superfeedr successfully.
#
# TODO: extract_links currently has a bug that makes it drop trailing
# slashes. ugh. fix that.
content = item.get('content') or item.get('summary', '')
links = [
util.clean_url(util.unwrap_t_umblr_com(url))
for url in util.extract_links(content)
if util.domain_from_link(url) not in source.domains
]
unique = []
for link in util.dedupe_urls(links):
if len(link) <= _MAX_STRING_LENGTH:
unique.append(link)
else:
logger.info(
f'Giving up on link over {_MAX_STRING_LENGTH} chars! {link}'
)
if len(unique) >= MAX_BLOGPOST_LINKS:
logger.info('Stopping at 10 links! Skipping the rest.')
break
logger.info(f'Found links: {unique}')
if len(url) > _MAX_KEYPART_BYTES:
logger.warning(
'Blog post URL is too long (over 500 chars)! Giving up.')
bp = models.BlogPost(id=url[:_MAX_KEYPART_BYTES],
source=source.key,
feed_item=item,
failed=unique)
else:
bp = models.BlogPost(id=url,
source=source.key,
feed_item=item,
unsent=unique)
bp.get_or_save()
