def test_request_retries(self):
retries = 5
e = error.HTTPError('some url', 503, 'some msg', '', BytesIO())
self.mock_open = MagicMock(side_effect=e)
request.make_request('http://fauxurl.com', max_retries=retries, open_func=self.mock_open)
self.assertEqual(self.mock_open.call_count, retries)
def save_from_url(url, filename):
"""
Saves a remote file to S3 and returns
its S3 URL.
"""
try:
res = make_request(url)
except error.HTTPError as e:
# Wikimedia 404 errors are very common, since images may go
# out of date.
# So common that for now these exceptions are just ignored.
if e.code == 404 and 'wikimedia' in url:
logger.warning('Error requesting {0} : {1}'.format(url, e))
# Other exceptions are more remarkable and should be brought up.
else:
logger.exception('Error requesting {0} : {1}'.format(url, e))
return None
except (ConnectionResetError, BadStatusLine, ValueError) as e:
logger.exception('Error requesting {0} : {1}'.format(url, e))
return None
data = BytesIO(res.read())
return save_from_file(data, filename)
def save_from_url(url, filename):
"""
Saves a remote file to S3 and returns
its S3 URL.
"""
try:
res = make_request(url)
except error.HTTPError as e:
# Wikimedia 404 errors are very common, since images may go
# out of date.
# So common that for now these exceptions are just ignored.
if e.code == 404 and 'wikimedia' in url:
logger.warning('Error requesting {0} : {1}'.format(url, e))
# Other exceptions are more remarkable and should be brought up.
else:
logger.exception('Error requesting {0} : {1}'.format(url, e))
return None
except (ConnectionResetError, BadStatusLine, ValueError) as e:
logger.exception('Error requesting {0} : {1}'.format(url, e))
return None
data = BytesIO(res.read())
return save_from_file(data, filename)
def _request(endpoint, url, format='json'):
complete_url = '{0}{1}'.format(endpoint, url)
res = make_request(complete_url)
content = res.read()
if format == 'json':
return json.loads(content.decode('utf-8'))
elif format == 'xml':
return xmltodict.parse(content)
return None
def _request(endpoint, url, format="json"):
complete_url = "{0}{1}".format(endpoint, url)
res = make_request(complete_url)
content = res.read()
if format == "json":
return json.loads(content.decode("utf-8"))
elif format == "xml":
return xmltodict.parse(content)
return None
def _get_html(url):
# Some sites, such as NYTimes, track which
# articles have been viewed with cookies.
# Without cookies, you get thrown into an infinite loop.
cookies = CookieJar()
opener = request.build_opener(request.HTTPCookieProcessor(cookies))
# Get the raw html.
# Spoof a user agent.
# This can help get around 403 (forbidden) errors.
try:
html = make_request(url, open_func=opener.open, headers={'User-Agent': 'Chrome'}).read()
except IncompleteRead as e:
html = e.partial
return html
def _get_html(url):
# Some sites, such as NYTimes, track which
# articles have been viewed with cookies.
# Without cookies, you get thrown into an infinite loop.
cookies = CookieJar()
opener = request.build_opener(request.HTTPCookieProcessor(cookies))
# Get the raw html.
# Spoof a user agent.
# This can help get around 403 (forbidden) errors.
try:
html = make_request(url,
open_func=opener.open,
headers={
'User-Agent': 'Chrome'
}).read()
except IncompleteRead as e:
html = e.partial
return html
