def prepare_response(self, request, cached):
"""Verify our vary headers match and construct a real urllib3
HTTPResponse object.
"""
# Special case the '*' Vary value as it means we cannot actually
# determine if the cached response is suitable for this request.
# This case is also handled in the controller code when creating
# a cache entry, but is left here for backwards compatibility.
if "*" in cached.get("vary", {}):
return
# Ensure that the Vary headers for the cached response match our
# request
for header, value in cached.get("vary", {}).items():
if request.headers.get(header, None) != value:
return
body_raw = cached["response"].pop("body")
headers = CaseInsensitiveDict(data=cached["response"]["headers"])
if headers.get("transfer-encoding", "") == "chunked":
headers.pop("transfer-encoding")
cached["response"]["headers"] = headers
try:
body = io.BytesIO(body_raw)
except TypeError:
# This can happen if cachecontrol serialized to v1 format (pickle)
# using Python 2. A Python 2 str(byte string) will be unpickled as
# a Python 3 str (unicode string), which will cause the above to
# fail with:
#
# TypeError: 'str' does not support the buffer interface
body = io.BytesIO(body_raw.encode("utf8"))
return HTTPResponse(body=body,
preload_content=False,
**cached["response"])
def prepare_response(self, request, cached):
"""Verify our vary headers match and construct a real urllib3
HTTPResponse object.
"""
# Special case the '*' Vary value as it means we cannot actually
# determine if the cached response is suitable for this request.
if "*" in cached.get("vary", {}):
return
# Ensure that the Vary headers for the cached response match our
# request
for header, value in cached.get("vary", {}).items():
if request.headers.get(header, None) != value:
return
body_raw = cached["response"].pop("body")
headers = CaseInsensitiveDict(data=cached['response']['headers'])
if headers.get('transfer-encoding', '') == 'chunked':
headers.pop('transfer-encoding')
cached['response']['headers'] = headers
try:
body = io.BytesIO(body_raw)
except TypeError:
# This can happen if cachecontrol serialized to v1 format (pickle)
# using Python 2. A Python 2 str(byte string) will be unpickled as
# a Python 3 str (unicode string), which will cause the above to
# fail with:
#
# TypeError: 'str' does not support the buffer interface
body = io.BytesIO(body_raw.encode('utf8'))
return HTTPResponse(
body=body,
preload_content=False,
**cached["response"]
)
def cache_response(self, request, response, body=None, status_codes=None):
"""
Algorithm for caching requests.
This assumes a requests Response object.
"""
# From httplib2: Don't cache 206's since we aren't going to
# handle byte range requests
cacheable_status_codes = status_codes or self.cacheable_status_codes
if response.status not in cacheable_status_codes:
logger.debug("Status code %s not in %s", response.status,
cacheable_status_codes)
return
response_headers = CaseInsensitiveDict(response.headers)
if "date" in response_headers:
date = calendar.timegm(parsedate_tz(response_headers["date"]))
else:
date = 0
# If we've been given a body, our response has a Content-Length, that
# Content-Length is valid then we can check to see if the body we've
# been given matches the expected size, and if it doesn't we'll just
# skip trying to cache it.
if (body is not None and "content-length" in response_headers
and response_headers["content-length"].isdigit()
and int(response_headers["content-length"]) != len(body)):
return
cc_req = self.parse_cache_control(request.headers)
cc = self.parse_cache_control(response_headers)
cache_url = self.cache_url(request.url)
logger.debug('Updating cache with response from "%s"', cache_url)
# Delete it from the cache if we happen to have it stored there
no_store = False
if "no-store" in cc:
no_store = True
logger.debug('Response header has "no-store"')
if "no-store" in cc_req:
no_store = True
logger.debug('Request header has "no-store"')
if no_store and self.cache.get(cache_url):
logger.debug('Purging existing cache entry to honor "no-store"')
self.cache.delete(cache_url)
if no_store:
return
# https://tools.ietf.org/html/rfc7234#section-4.1:
# A Vary header field-value of "*" always fails to match.
# Storing such a response leads to a deserialization warning
# during cache lookup and is not allowed to ever be served,
# so storing it can be avoided.
if "*" in response_headers.get("vary", ""):
logger.debug('Response header has "Vary: *"')
return
# If we've been given an etag, then keep the response
if self.cache_etags and "etag" in response_headers:
expires_time = 0
if response_headers.get("expires"):
expires = parsedate_tz(response_headers["expires"])
if expires is not None:
expires_time = calendar.timegm(expires) - date
expires_time = max(expires_time, 14 * 86400)
logger.debug(
"etag object cached for {0} seconds".format(expires_time))
logger.debug("Caching due to etag")
self.cache.set(
cache_url,
self.serializer.dumps(request, response, body),
expires=expires_time,
)
# Add to the cache any permanent redirects. We do this before looking
# that the Date headers.
elif int(response.status) in PERMANENT_REDIRECT_STATUSES:
logger.debug("Caching permanent redirect")
self.cache.set(cache_url,
self.serializer.dumps(request, response, b""))
# Add to the cache if the response headers demand it. If there
# is no date header then we can't do anything about expiring
# the cache.
elif "date" in response_headers:
date = calendar.timegm(parsedate_tz(response_headers["date"]))
# cache when there is a max-age > 0
if "max-age" in cc and cc["max-age"] > 0:
logger.debug("Caching b/c date exists and max-age > 0")
expires_time = cc["max-age"]
self.cache.set(
cache_url,
self.serializer.dumps(request, response, body),
expires=expires_time,
)
# If the request can expire, it means we should cache it
# in the meantime.
elif "expires" in response_headers:
if response_headers["expires"]:
expires = parsedate_tz(response_headers["expires"])
if expires is not None:
expires_time = calendar.timegm(expires) - date
else:
expires_time = None
logger.debug(
"Caching b/c of expires header. expires in {0} seconds"
.format(expires_time))
self.cache.set(
cache_url,
self.serializer.dumps(request, response, body=body),
expires=expires_time,
)